For decades, industrial companies have been optimizing how they sell products. Today, the competitive advantage is increasingly shifting toward post-sale monetization.

In industries such as automotive, construction, and mechanical engineering, the installed base often exceeds annual new investments by a factor of many. Every piece of equipment in the field generates continuous demand for service, spare parts, and data-driven services.

However, this opportunity often goes untapped: In many places, aftermarket revenue is still based on reactive service calls, manual processes, and fragmented offerings—with limited scalability.

The key question, therefore, is: How can the aftermarket be systematically transformed into a scalable revenue engine?

This article shows how leading companies are shaping this transformation—and turning their installed base into predictable revenue streams.

The revenue engine industrial companies already operate

Margins on equipment sales are tightening across industrial sectors. Growth is becoming harder to sustain with hardware alone.

At the same time, a different source of value keeps expanding quietly: the installed base.

Machines, tools, vehicles, or industrial systems deployed at customer sites generate continuous demand for:

• maintenance and inspections
• spare parts and replacements
• upgrades and retrofits
• performance optimization

This demand is predictable and recurring. It reflects how assets are actually used in day-to-day operations.

According to McKinsey, services and aftermarket activities already account for 30–50% of revenues in industrial sectors, often delivering higher margins than new equipment sales (2024). The economic potential is well established.

In many cases, however, revenue still scales more slowly than the installed base itself. The gap becomes visible over time — in underutilized service capacity, inconsistent pricing, and missed opportunities to translate usage into predictable income.

Installed base without structure: where value gets diluted

From an operational standpoint, aftermarket activity is intense. Service teams are busy, spare parts move continuously, and customer interactions are frequent.

The issue rarely lies in the scope of these activities. It emerges in how they are structured — or, in many cases, how structure evolves over time.

In many companies, local adaptations accumulate without a shared commercial backbone:

• service offerings evolve across regions and differ in structure and scope, limiting their ability to scale as repeatable revenue models
• contracts are tailored to individual customers, making them difficult to standardize and compare
• pricing reflects local decisions rather than a common logic tied to value or usage
• asset usage is not directly reflected in billing, leaving a significant share of value unmonetized
• data about the installed base is fragmented across systems, preventing a consistent translation of service activity into revenue

Individually, these decisions often make sense. They reflect customer needs, market conditions, and operational realities.

Taken together, however, they create a fragmented model — one where activity grows, but the ability to structure, replicate, and scale revenue does not keep pace.

The consequences are visible over time:

• revenue streams remain difficult to forecast
• scaling service operations requires proportional cost increases
• pricing does not consistently reflect delivered value
• commercial decisions rely on experience rather than system logic

At this point, the question is no longer whether aftermarket matters.
It is how systematically it is managed as a revenue domain.

Why aftermarket becomes the economic center of gravity

Several structural shifts are reinforcing the role of aftermarket across industrial sectors.

First, customers increasingly evaluate suppliers based on outcomes:

• uptime
• availability of parts and service
• production throughput
• energy efficiency
• mean time to repair (MTTR)

Second, digital connectivity (IoT, telemetry, monitoring systems) provides real-time visibility into asset usage and performance, enabling:

• predictive maintenance
• usage-based billing
• faster response to issues
• a clear connection between operations, service, and revenue

Third, financing models continue to shift from upfront investment toward usage-oriented structures, including:

• pay-per-use (based on hours, output, or volume)
• subscription-based service contracts
• availability-based agreements (linked to uptime)
• equipment-as-a-service models

Together, these changes move the commercial focus toward the operational phase of assets. Revenue aligns more closely with how equipment performs over time, not only with how it is sold.

Organizations that adapt to this shift start treating aftermarket as a continuous revenue system, rather than a sequence of service events.

The critical step: translating service into scalable revenue models

The transition to scalable aftermarket models does not happen automatically.
It requires a shift from operational activity to structured commercial logic.

In many organizations, service, pricing, contracts, and billing evolve separately.
As a result, even advanced capabilities — such as connected assets or digital services — remain difficult to monetize consistently.

What differentiates leading companies is not the scope of their services, but how systematically they translate them into repeatable, scalable revenue models.

Three monetization models shaping modern aftermarket

The transition becomes tangible in how companies structure their commercial models.

In practice, this means translating service activity, asset usage, and partner interactions into clearly defined revenue mechanisms.

Across industrial sectors, three monetization models consistently emerge as the foundation for scalable aftermarket growth:

1. Contract-based service

Maintenance and service are packaged into standardized, repeatable offerings:

• preventive maintenance programs
• defined service levels with response times
• bundled on-site and remote services

Contracts introduce consistency. They also create a foundation for forecasting and capacity planning. In practice, this shifts revenue from irregular service interventions to multi-year agreements with clearly defined scope, pricing, and delivery models.

Example:

DMG MORI has developed a structured portfolio of lifecycle service contracts. Maintenance, spare parts, and technical support are bundled into standardized contract packages—including clearly defined service levels and response times.

Impact:

• greater predictability of service revenue
• transparent cost structure for customers
• scalable service delivery across regions

2. Usage-based monetization

Revenue increasingly reflects how intensively assets are used.

Typical billing dimensions include:

• machine operating time
• production output, e.g., the number of parts produced
• processed volume

This model aligns commercial outcomes with operational reality.

Example:

TRUMPF applies pay-per-part models in its laser cutting machine business. Machines are installed at customer sites, and customers are charged based on the number of parts produced rather than ownership.

Machines are connected via IoT and telemetry, allowing TRUMPF to monitor production volume, machine status, and maintenance needs in real time. Contract logic and billing are directly linked to this data, converting every unit produced into a measurable revenue event.

As a result, revenue grows with actual machine usage, customers avoid large upfront investment, and both parties benefit from higher utilization, more predictable service, and more efficient production.

3. Platform and ecosystem-based models

At the most advanced level, aftermarket evolves into a connected ecosystem where value is created across networks of partners rather than in direct company–customer interactions.

Such a digital platform connects:

• manufacturers / suppliers
• service partners
• wholesalers and distributors
• end customers (whether a consumer or a business)

A typical process spans multiple participants: a service partner identifies a required part during maintenance, checks availability through the platform, places an order via a distributor, and receives technical documentation and installation guidance in the same workflow. Each step is supported by shared data and coordinated through a single interface.

These interactions often extend across multi-tier relationships, where manufacturers reach end users indirectly through partners — whether in B2B2C or B2B2B scenarios.

This enables transaction-based revenue streams, faster service execution, and broader market reach.

Example

ZF Aftermarket, a supplier of spare parts, repair solutions, and digital services for vehicles, operates a connected digital ecosystem that links workshops, distributors, and service partners in a unified aftermarket platform. The platform provides real-time access to parts catalogues, online ordering, technical documentation, and service support tools within one interface used by all participants.

The system integrates data on parts availability, orders, and service processes with partner workflows, enabling workshops and distributors to execute service and parts transactions with consistent, up-to-date information and fewer manual steps.

As a result, ZF Aftermarket expands revenue beyond direct parts sales by enabling digital services and partner transactions, strengthens engagement and collaboration across its aftermarket network, and increases operational efficiency and transparency in parts ordering and servicing.

What leading companies operationalize differently

Leading companies do not differ in what they offer, but in how consistently they operationalize monetization across the installed base.

They:

• define service offerings as structured products
• manage contracts as data rather than documents
• connect asset usage directly to commercial logic
• automate billing processes end-to-end

In practice, this leads to tangible shifts.

At TRUMPF, production data flows directly into pricing logic, allowing revenue to scale with output rather than installed capacity.

At DMG MORI, lifecycle service contracts bundle maintenance, monitoring, and support into standardized agreements that can be replicated across customers and regions.

Companies such as Jungheinrich combine fleet management with service packages, creating recurring revenue streams tied to the availability and performance of entire equipment fleets rather than individual service events.

At ZF Aftermarket, platform-based models extend monetization beyond direct sales by enabling transactions across a broader ecosystem of partners.

The common denominator is clear:
Every interaction with the installed base is treated as a potential revenue event — and is captured as such within a structured system.

Revenue is increasingly triggered by events in the lifecycle of the asset.

Conclusion: Rethinking aftermarket as a system

Industrial companies already operate extensive fleets of revenue-generating assets. The opportunity lies in how systematically these assets are integrated into a coherent commercial model.

A few questions can help assess the current state:

• How consistently is asset usage translated into revenue?
• To what extent are service offerings standardized and scalable?
• How tightly are contracts, pricing, and billing connected?
• How visible is the installed base as a structured data layer?

The answers often reveal a gap between operational activity and revenue capture.

Closing this gap does not require reinventing the business. It requires structuring what already exists — and connecting it through a system that can scale.

That is where aftermarket begins to function as a true revenue engine.

You might also like:

• E-Book: Digital Self-service & automation in B2B e-commerce » Learn more
• Search in B2B e-commerce can be a strategic differentiator » Learn moren
• E-Book: Build, Buy, or Orchestrate? Rethinking Commerce » Learn more
• The Tech Backbone of Subscription-Based E-Commerce » Learn more
• E-Book: Modern B2B Commerce — Beyond Transactions » Learn more

As automation expands, visibility into process execution becomes increasingly limited – leaving both users and operators without a clear view of what is actually happening.

Digital transformation has delivered real progress. Processes are faster. Costs are lower. Self-service channels have scaled.

Yet as enterprise systems become more distributed, a new challenge is emerging: maintaining end-to-end visibility into automated processes.

In many organizations, processes run smoothly most of the time. But when something unexpected happens, even simple questions become difficult to answer:

Where exactly is the process right now?
What is blocking it?
What happens next?

This visibility gap is becoming one of the hidden operational challenges of modern automation.

Why automation makes visibility harder

Over the past decade, enterprise architectures have changed dramatically.

Processes that once lived inside a single system are now distributed across multiple platforms and services:

• ERP systems
• CRM platforms
• SaaS applications
• internal microservices
• partner APIs

Execution has become asynchronous, event-driven, and cross-system.

From a technical perspective, everything may still be functioning correctly. But from a business perspective, reconstructing the state of a specific process instance becomes far more complex.

This is one of the reasons why Gartner estimates that up to 80% of business processes in large organizations are not fully visible end-to-end, which has fueled the rapid growth of process mining platforms such as Celonis and SAP’s Signavio suite.

The problem is not that systems lack monitoring. Most organizations have extensive technical telemetry.

The problem is that technical observability does not automatically translate into business visibility.

When process visibility breaks down

In practice, limited process visibility typically appears in a few recurring situations.

For example:

• a customer onboarding process stalls somewhere between identity verification and core banking systems
• an enterprise order moves through logistics and billing platforms but fails in a partner integration
• a telecom service activation request falls out of the provisioning workflow

In each case, the automation itself may work correctly in most scenarios.
But when something deviates from the expected path, the exact state of the process becomes unclear.

Organizations then face questions like:

• Which system is blocking the process?
• Which step failed?
• Who should resolve the issue?

Answering these questions often requires manual investigation across multiple systems.

Enterprise examples

Banking onboarding

Customer onboarding in modern banks typically involves:

• identity verification (KYC)
• anti-money laundering checks
• document validation
• credit scoring systems
• core banking platforms
• external data providers

Each step may run in a different system.

If the process stalls, the customer service interface may simply show a generic status such as “verification in progress.”

Internally, however, determining the actual cause can require tracing events across multiple systems.

This is one of the reasons why banks are among the largest adopters of process mining and operational intelligence tools.

Telecom “Order Fallout”

Telecommunications operators have a well-known operational problem called order fallout.

A service order may pass through multiple automated systems:

• customer management
• network provisioning
• billing platforms
• external partner integrations

If one step fails or produces inconsistent data, the order may drop out of the automated workflow.

Operators often maintain dedicated teams responsible for identifying and recovering these orders manually.

Automation handles the majority of cases efficiently—but exceptions require visibility and intervention.

Supply chain order-to-cash

In global supply chains, the order-to-cash process can involve:

• ERP systems
• warehouse management systems
• logistics platforms
• partner integrations

A single order may generate events across dozens of systems.

When disruptions occur, organizations often struggle to identify where exactly the process is delayed.

This challenge has led to the rise of supply chain “control towers”—systems designed specifically to provide end-to-end operational visibility.

Solutions from companies like Kinaxis or SAP aim to address precisely this issue.

Why traditional monitoring isn’t enough

Most enterprise environments already generate vast amounts of telemetry:

• logs
• metrics
• traces
• infrastructure monitoring

However, these tools are designed primarily for system health, not process understanding.

A log entry may reveal that an API call failed. But it does not answer a business question such as: Which customer onboarding case is currently blocked?

This gap between system observability and business understanding has given rise to a new category of tools focused on business observability and process intelligence.

Where the operational impact appears

Limited visibility does not necessarily mean that processes fail frequently. But when issues occur, the operational consequences can be significant.

Typical effects include:

• Longer resolution times
  → Teams must manually trace process execution across systems.
• Higher operational overhead
  → Support teams escalate issues to technical specialists.
• Difficulty improving processes
  → Without clear visibility into where delays occur, optimization becomes guesswork.

According to McKinsey research on automation, 30–50% of automated workflows still require human intervention, often because of exceptions or edge cases.

Understanding where and why those exceptions occur is therefore critical.

Designing automation for transparency

Leading organizations increasingly treat process visibility as a core capability rather than an afterthought.

Several architectural principles are becoming common.

1. Explicit process state models

Instead of relying solely on system logs, processes expose clear business states such as:

• application received
• verification completed
• exception detected

2. Event-driven process tracking

Key process steps emit events that can be aggregated into a real-time process timeline.

3. Unified process dashboards

Operational teams can track the progress of individual cases without investigating multiple systems.

4. Exception management

Edge cases are detected automatically and routed to responsible teams.

Automation’s next maturity stage

The first wave of digital transformation focused on digitizing and automating processes. The next phase is about making those processes understandable in real time.

As enterprise architectures grow more distributed and event-driven, maintaining this visibility becomes increasingly important.

Organizations that invest in process intelligence gain:

• faster incident resolution
• better operational decision-making
• improved customer communication
• stronger control over complex digital operations

You might also like:

• E-Book: Digital self-service & automation in B2B e-commerce » Learn more
• The Tech Backbone of Subscription-Based E-Commerce » Learn more
• Agentic Commerce, Part 1: Where Conversations Become Transactions » Learn more
• Strategic Patterns of IT System Integration » Learn more
• Low-Code Falls Short Without Full-Code » Learn more

Subscription in healthcare is not simply a new business model — it requires a fundamental redesign of data architectures, system integration, and regulatory workflows.

Subscription models have transformed many digital industries.

Software moved to SaaS.
Media adopted streaming subscriptions.
E-commerce relies on recurring delivery models.

Healthcare is now exploring a similar shift.

Instead of episodic treatment, providers increasingly experiment with continuous care models built on digital platforms, connected devices, and long-term patient engagement.

At first glance, the logic seems obvious. Chronic diseases require ongoing monitoring. Preventive care benefits from continuous feedback. Telehealth enables persistent patient interaction.

Yet scaling subscription healthcare is proving far harder than expected.

The reason is simple: healthcare systems were never designed for continuous services. They were designed for episodic treatment.

This structural mismatch explains why subscription healthcare is not primarily a product innovation. It is an infrastructure transformation.

What is subscription healthcare?

Subscription healthcare describes care models in which patients receive continuous services instead of episodic treatment.

These models combine:

• digital health platforms
• connected medical devices
• telehealth interactions
• longitudinal patient data

Instead of paying per procedure, patients or healthcare organizations pay for ongoing access, monitoring, and engagement.

Typical examples include:

• chronic disease management programs
• digital therapeutics
• preventive health subscriptions
• remote patient monitoring services

In these models, value is delivered through continuous patient interaction rather than isolated medical events.

The system architecture behind subscription healthcare

Continuous care models depend on several interconnected system layers.

Subscription healthcare platforms typically rely on four architectural components:

1. Patient interaction layer
   telehealth platforms, mobile health apps, patient portals
2. Device and monitoring layer
   wearables, medical sensors, connected devices
3. Integration and interoperability layer
   APIs, FHIR interfaces, integration platforms
4. Data governance layer
   identity management, consent tracking, secure analytics

Without these layers working together, continuous healthcare services cannot operate reliably.

The structural tension behind subscription healthcare

Subscription businesses normally evolve through rapid iteration. Digital platforms continuously optimize their services based on user behavior and engagement data.

Healthcare operates under very different conditions.

Three structural barriers shape the evolution of subscription healthcare:

1. regulation slows digital iteration
2. healthcare systems remain fragmented
3. continuous care requires real-time data architectures

Understanding these barriers explains why scaling digital health platforms requires more than innovative applications.

Regulation slows down digital iteration

Healthcare is one of the most tightly regulated digital environments.

In Europe, patient data falls under the General Data Protection Regulation (GDPR). Health data requires explicit consent, purpose-bound processing, and full traceability.

At the same time, digital health applications may fall under the EU Medical Device Regulation (MDR). When software supports diagnosis or therapy, even small changes can trigger validation and documentation requirements.

As a result, healthcare platforms cannot evolve as quickly as typical digital services.
Innovation must happen within controlled regulatory frameworks.

Fragmented healthcare systems limit continuous care

Even if regulatory challenges are addressed, subscription healthcare faces another structural barrier: integration.

European healthcare ecosystems remain highly fragmented.

Patient data is typically distributed across multiple systems, including:

• hospital information systems
• laboratory platforms
• insurance databases
• digital health applications

These systems rarely share consistent data models or integration standards.

Standards such as FHIR (Fast Healthcare Interoperability Resources) improve data exchange between healthcare systems. However, adoption remains uneven across organizations and countries.

Healthcare providers therefore often introduce additional integration layers, including:

• API gateways over legacy systems
• interoperability middleware
• integration platforms (iPaaS)

Without interoperability infrastructure, continuous care services remain isolated digital products rather than integrated healthcare platforms.

Continuous care requires real-time architectures

Many subscription healthcare services rely on timely patient information.

Remote monitoring systems collect continuous data streams from connected devices such as glucose sensors or cardiac monitors. These signals must often be processed in near real time.

Traditional healthcare IT systems were designed for batch processing. Continuous care requires different architectures.

Modern digital health platforms increasingly rely on:

• event-driven architectures
• streaming data pipelines
• low-latency processing environments

These technologies allow providers to detect risks earlier and respond to patient signals more quickly.

Data governance becomes the foundation of continuous care

Continuous care depends on maintaining a consistent patient context across time and systems.

Healthcare platforms must integrate data from sources such as electronic health records, laboratory results, wearable devices, and digital health applications. These datasets differ in structure, update frequency, and sensitivity.

To manage this complexity, organizations need strong governance frameworks that handle identity resolution, consent management, and traceability of clinical data.

Master Data Management (MDM) often plays a central role by linking patient identities across systems and preventing data inconsistencies.

Because healthcare data cannot be freely reused for analytics, organizations rely on pseudonymization, encryption, and strict access controls.

The result is not a single centralized database but a governed healthcare data ecosystem that supports continuous care.

Subscription healthcare is an infrastructure transformation

Subscription healthcare is an infrastructure transformation

The promise of subscription healthcare is compelling. Continuous monitoring and personalized interventions could significantly improve patient outcomes.

But delivering these models requires more than new digital services.

Healthcare organizations must redesign how their systems handle data integration, interoperability, and regulatory compliance.

Subscription healthcare is therefore less a product innovation than an infrastructure transformation.

Organizations that treat subscription services as a simple digital add-on will struggle to scale them. Those that rethink their data architectures and integration layers will define the next generation of digital healthcare platforms.

Key takeaway

Subscription healthcare works when three elements align:

• interoperable healthcare systems
• governed patient data architectures
• regulatory-aware digital platforms

Without this foundation, continuous care remains difficult to scale — no matter how promising the digital service may be.

You might also like:

• The Tech Backbone of Subscription-Based E-Commerce » Learn more
• Real-World Applications of Data Anonymization Across Industries » Learn more
• From Ownership to Use: Why Data Sharing Is the New Strategic Imperative » Learn more
• Key Drivers of Cloud Adoption in Pharma and Life Sciences » Learn more
• Pharma’s AI Readiness Index: Who Leads the Race? » Learn more

E-invoicing is often framed as a compliance issue. In reality, it is a structural change in how digital business transactions are executed, validated, and exchanged across Europe.

The European Union is moving decisively toward mandatory electronic invoicing across its member states. What began as a public-sector initiative is now expanding into business-to-business transactions, with clear timelines and legal obligations.

For many organizations, this raises a familiar question:

Is e-invoicing just another regulatory requirement to be handled by finance and accounting?

The answer is no.

For B2B companies operating complex sales models, mandatory e-invoicing is less about invoices — and more about data, systems, and integration readiness.

It fundamentally reshapes how transaction data flows between e-commerce platforms, ERP systems, payment and billing solutions, and government-designated networks, forcing a rethinking of automation and interoperability across the entire order-to-cash process.

A structural shift in digital business transactions

E-invoicing in the EU is not about replacing PDFs with another file format. It is about replacing documents with structured, machine-readable data that can be automatically validated, transmitted, and processed by systems on both sides of a transaction.

At its core, e-invoicing introduces:

• standardized data models,
• predefined validation rules,
• mandatory electronic exchange via approved platforms or networks in jurisdictions implementing clearance or continuous transaction control (CTC) models.

This turns invoicing into a system-driven event, tightly coupled with digital sales, payments, and tax reporting, and represents a fundamental architectural change for B2B companies operating e-commerce and cross-border sales models.

Data foundations of e-invoicing

The backbone of EU e-invoicing is the EN 16931 standard, which defines a common semantic data model for electronic invoices across member states. It specifies what data an electronic invoice must contain and how that data should be structured to ensure interoperability across borders.

In practice, this European model is implemented at the national level through CIUS (Core Invoice Usage Specifications). CIUS define country-specific constraints, mandatory fields, and validation rules that build on top of EN 16931, ensuring alignment with local tax law and reporting requirements while preserving cross-border compatibility.

Key standards and formats include:

• EN 16931 — the European invoice data model
• UBL XML and Peppol BIS Billing 3.0 — widely used structured formats

• Factur-X / ZUGFeRD and XRechnung — national implementations aligned with EN 16931 and CIUS requirements

These formats require far more than header-level invoice data. They rely on high-quality, structured master data, including:

• product identifiers (SKU, GTIN, internal IDs),
• pricing and discount logic,
• VAT rates and tax classifications,
• units of measure,
• customer and supplier master data.

This is where PIM / PCM systems and Master Data Management (MDM) become critical. Inconsistent product data, misaligned tax codes, or fragmented customer records do not merely create operational friction — they can lead to invoice rejection by national e-invoicing platforms.

In practice, e-invoicing acts as a forcing function for better data governance across the organization.

System integration: connecting internal systems with e-invoicing networks

Mandatory e-invoicing does not introduce a single new system. It introduces an ecosystem of integrations.

Internal Systems

• Invoice-related data is generated and enriched in: ERP systems (such as SAP, Microsoft Dynamics, or Oracle)
• finance and billing systems
• B2B e-commerce platforms
• Order Management Systems (OMS)

External networks and platforms

Invoices must be transmitted through:

• the Peppol network and certified Access Points, or
• national e-invoicing platforms, such as Poland’s KSeF

This applies to both:

• outbound invoices (issued to customers)
• inbound invoices (received from suppliers)

The technical challenge lies not in file generation, but in:

• data transformation and mapping to EN-compliant formats
• API-based integration or middleware orchestration
• handling validation responses, rejections, and status updates

Clearance vs reporting models: why integration requirements differ across the EU

Mandatory e-invoicing in the EU is implemented through different control models, which directly impact system architecture and integration complexity.

In clearance-based models, invoices must be validated or approved by a central platform before they are legally issued. This approach is used, for example, in Italy (Sistema di Interscambio) and Poland (KSeF).

In reporting-based models, invoices are exchanged directly between business partners and reported to tax authorities after issuance, often within defined time windows. This model is present, for example, in Germany and Belgium.

From a system integration perspective, the difference is fundamental:

• clearance models require synchronous or near-real-time integrations,
• reporting models allow asynchronous data exchange.

For businesses operating across multiple EU markets, this means that e-invoicing is not a single integration problem, but a set of country-specific orchestration patterns that must be handled consistently across ERP, billing, and integration layers.

Billing and payment platforms in the e-invoicing integration landscape

In modern B2B environments, these differences are particularly relevant for businesses using global payment, billing, and invoicing platforms, which must be integrated into national e-invoicing and clearance ecosystems.

Such platforms — for example Stripe — are widely used to manage pricing logic, tax calculation, and the generation of commercial or billing invoices in digital and subscription-based business models.

With mandatory e-invoicing, the role of these platforms changes.

While billing platforms play a critical role in structuring transaction data, their invoice output does not automatically meet regulatory e-invoicing requirements without additional connectors or extensions. In particular, it may not comply with:

• EN 16931 semantic data models,
• country-specific CIUS rules,
• submission to Peppol networks or national clearance platforms.

As a result, invoice data originating from billing platforms must be treated as upstream input to the e-invoicing process.

In practice, this requires:

• extracting invoice-relevant data from billing and payment platforms
• mapping and enriching this data to EN 16931-compliant structures
• applying country-specific validation and reporting rules
• routing invoices through approved access points or national platforms
• handling validation feedback and processing statuses alongside other invoice flows

In practice, this transformation is often handled by specialized third-party connectors or applications that bridge global billing platforms with country-specific e-invoicing systems, such as Peppol Access Points, tax compliance and e-invoicing middleware platforms, or local connectors designed for national clearance systems, for example, Stripe to KSeF (S2K) in Poland.

From an enterprise architecture perspective, billing platforms become one of several data sources feeding the e-invoicing integration layer, rather than standalone compliance solutions.

This shift places data consistency, master data governance, and cross-system orchestration at the center of e-invoicing readiness.

How mandatory e-invoicing changes B2B digital transaction flows

In traditional B2B environments, invoicing often sits at the end of a long, manual chain. Mandatory e-invoicing shortens and automates that chain.

Key changes include:

• invoice generation triggered directly by system events
• tighter coupling between order, delivery, and invoice data
• near-real-time submission to external platforms
• reduced separation between front-office and back-office systems

For B2B e-commerce platforms and marketplaces, this means invoicing can no longer be treated as an afterthought. It becomes an integral part of the digital transaction lifecycle, with direct dependencies on product data, pricing logic, and tax determination at the moment of purchase.

Regulatory context and timeline

The legal foundation for e-invoicing in the EU is well established.

• Directive 2014/55/EU introduced mandatory acceptance of structured e-invoices in public procurement (B2G).
• The VAT in the Digital Age (ViDA) initiative extends this logic to the private sector, enabling mandatory B2B e-invoicing and digital VAT reporting across member states.
• The EU roadmap points toward EU-wide B2B e-invoicing by 2030, with national rollouts already underway.

Countries such as Germany, Poland, and Belgium are implementing phased mandates, often combining e-invoicing with near-real-time VAT reporting — particularly relevant for cross-border transactions.

For international B2B organizations, this means compliance must be managed across multiple jurisdictions, not as a single local project.

How to prepare: a practical IT and data roadmap

For executives, the most important question is not whether to act, but how.

A pragmatic preparation roadmap typically includes:

1. Assess data quality and completeness
   Review product, pricing, tax, and partner data against EN 16931 requirements.
2. Map the system landscape
   Identify where invoice-relevant data is created, enriched, and stored.
3. Align data models
   Ensure consistency across PIM, e-commerce, ERP, and finance systems.
4. Design the integration architecture
   Decide between direct integrations and middleware-based orchestration.
5. Pilot selected markets
   Start with one or two countries to validate formats, flows, and error handling.
6. Scale gradually
   Extend the approach across countries and transaction volumes.
7. Establish cross-functional ownership
   Involve IT, e-commerce, finance, and operations from the outset.

Conclusion: e-invoicing as a catalyst for digital maturity

Mandatory e-invoicing is not simply another compliance checkbox. It is a catalyst for better data discipline, deeper automation, and more resilient digital architectures.

Companies that approach e-invoicing strategically—by strengthening data foundations and integration capabilities—will not only meet regulatory requirements, but also gain:

• faster and more reliable order-to-cash processes,
• improved transparency across borders,
• a scalable foundation for future digital initiatives.

In the EU’s evolving regulatory landscape, e-invoicing rewards organizations that treat data and integration not as technical details, but as strategic assets.

You might also like:

• Data Anonymization, Part 1: Turning Sensitive Information into Strategic Value » Learn more
• Agentic Commerce, Part 1: Where Conversations Become Transactions » Learn more
• How buzzwords could put your IT at risk » Learn more
• Mastering software delivery: Overcoming development hurdles » Learn more
• How to choose the right IT service provider » Learn more

Headless commerce is often positioned as the modern standard for digital commerce — promising greater flexibility, faster innovation cycles, and lower long-term cost. In practice, these benefits are real, but conditional.

At its core, headless commerce separates the customer-facing experience from core commerce logic and backend systems. Frontends consume commerce capabilities via APIs rather than being tightly coupled to a single platform, enabling independent evolution of user experiences and backend services. In this sense, headless commerce is frequently a foundational element of broader composable commerce and MACH architectures, where modular, API-first services are combined to support scalability, resilience, and incremental platform evolution.

Based on real-world projects across B2B, regulated industries, and complex enterprise landscapes, this article examines where headless commerce delivers tangible strategic value — and where it introduces operational complexity that is often underestimated.

Rather than advocating headless as a default, the goal is to provide decision-makers with a pragmatic framework to assess when headless commerce is the right architectural choice — and when alternative or hybrid approaches may be more effective.

Headless Commerce: A Reality Check on 7 Core Promises

1. Decoupled frontend & backend

The promise:

By decoupling the presentation layer from the backend, companies are said to gain creative freedom. Frontend teams can redesign storefronts, test new experiences, and launch new digital touchpoints — all without disrupting core commerce logic or ERP stability. This independence supposedly leads to faster UX innovation and a better customer experience.

The reality:

Decoupling brings autonomy but also increases the number of moving parts. Multiple codebases, API contracts and versioning, compatibility across frontends, regression testing across endpoints, cross-team coordination (frontend vs backend vs integration) all come into play.

Many organizations often discover that managing two or more separately deployed systems means more coordination, not less. As one practitioner put it on a discussion forum: “You don’t just build a website anymore — you build a frontend app, an API layer, and a backend integration ecosystem.”

Recommendations for executives:

• Enforce API contracts and semantic versioning before decoupling.
• Establish clear ownership between frontend and backend teams.
• Budget for integration testing and regression cycles — they’ll grow.
• Start small: one frontend, one channel, one proof of concept.

2. Faster time-to-market

The promise:

With independent frontends and standardized APIs, organizations can run parallel workstreams. Frontend teams can launch new features, campaigns, or localized experiences without waiting for backend releases. Time-to-market should shrink dramatically.

The reality:

In practice, the opposite often happens. When multiple teams, vendors, and environments are involved, coordination overhead grows. Each sprint depends on stable APIs, synchronized data, and integration testing — and if one piece lags, everything slows down.

Real-world case studies often show that the “faster” model only works in organizations with mature CI/CD pipelines, shared governance, and clear product ownership. Otherwise, the supposed independence simply moves bottlenecks elsewhere.

Recommendations for executives:

• Pilot with one product line or region before scaling.
• Align release cycles between frontend, backend, and integration teams.
• Invest in automation for testing, deployment, and rollback.
• Track time-to-market empirically — not through assumptions.

3. API-managed complexity (pricing, catalogs, contracts)

The promise:

Headless architecture is touted as ideal for B2B commerce because complex business rules — like contract-based pricing, custom assortments, or buyer-specific catalogues — can be surfaced via APIs. The backend systems (ERP, PIM, CPQ) expose tailored data; front-ends simply render relevant subsets. This promises consistency across all channels.

The reality:

APIs do not magically simplify business complexity. In fact, they shift work elsewhere. You still need strong data modelling, domain ownership, mapping of backend data to front-end consumption, governance of API contracts, and reconciliation of legacy data frameworks. Without governance, APIs can multiply inconsistencies instead of eliminating them.

Many B2B practitioners report that implementing contract-based pricing via APIs often leads to duplicated logic between systems, fragile transformations, and heavy reconciliation work.

Recommendations for executives:

• Define domain ownership — which system is the source of truth for each entity.
• Implement robust API governance: documentation, monitoring, and testing.
• Don’t underestimate data mapping and transformation overhead.
• Budget for long-term API maintenance, not just initial delivery.

4. Omnichannel & multi-brand enablement

The promise:

One backend, many frontends — rather than building a separate back-end/storefront per brand, channel or device, one shared commerce back-end is used with multiple front-ends plugging in via APIs. This promises brand-consistent experiences, lower duplication, and faster rollout into new markets or channels.

The reality:

This works beautifully on whiteboards — less so in production. A shared backend becomes a potential bottleneck if not scaled or managed properly. Front-ends may compete for the same services — integration load increases. Naming, versioning, payload sizes, network latencies and monitoring across multiple front-ends become challenges. Additionally, legacy systems may not have been built for many front-ends, so retrofit complexity appears.

Recommendations for executives:

• Map out all channels and their performance requirements before designing APIs.
• Create a versioning strategy that allows gradual rollout per channel.
• Avoid one-size-fits-all APIs — tailor endpoints where business logic diverges.
• Monitor cross-channel latency and concurrency from day one.

5. Improved site performance

The promise:

Modern frontend frameworks — Angular, React, and Vue — combined with headless APIs are said to deliver lightning-fast websites. Decoupling allows lean, optimized user interfaces that load faster, especially on mobile. Better performance means lower abandonments, higher conversion rates, and happier customers.

The reality:

Headless can improve performance — but it can just as easily degrade it. Multiple API calls, client-side rendering, or misconfigured caching can slow pages down instead of speeding them up. SEO performance can also suffer without proper server-side rendering (SSR) or static site generation (SSG).

Practitioners warn that without caching layers, CDN tuning, and backend aggregation, the modern frontend can end up being slower than a traditional monolith.

Recommendations for executives:

• Design for API aggregation — reduce round trips.
• Implement CDN and edge caching with clear invalidation logic.
• Use SSR or SSG to preserve SEO and first-load performance.
• Continuously monitor Core Web Vitals post-launch.

6. Better integration & data consistency

The promise:

Because everything communicates via APIs, data should stay synchronized across ERP, PIM, OMS, CRM, and analytics. Pricing, availability, and customer data should remain consistent — a “single source of truth” accessible to all systems.

The reality:

This ideal rarely materializes automatically. Each system has its own data model, update cycles, and business logic. Without explicit mapping, transformation, and reconciliation, discrepancies emerge — and once they do, they’re hard to eliminate.

For instance, bookkeeping of stock across channels, contract customizations, local regulatory/market differences still require reconciliation. Achieving “single source of truth” remains hard.

Many practitioners call “seamless integration” the biggest myth of headless commerce. It’s achievable, but only with mature data governance and clearly defined ownership of each domain.

Recommendations for executives:

• Define data ownership and synchronization rules early.
• Use middleware or integration platforms to manage transformations.
• Implement continuous reconciliation and anomaly detection.
• Treat data governance as part of the architecture, not a post-launch fix.

7. Lower long-term total cost of ownership (TCO)

The promise:

A modular architecture allows gradual modernization — no more expensive monolithic re-platforming. You can replace or extend components step by step, reducing long-term costs while improving flexibility.

The reality:

While modularity sounds compelling, the initial investment (architecture design, API gateways, more teams, versioning, DevOps) is substantial. Over time, the “hidden costs” of maintaining many services, vendor management, API version drift, glue-code, monitoring, and orchestrating multiple releases may exceed that of a simpler tightly-coupled stack.

Recommendations for executives:

• Calculate TCO including all integration, DevOps, and vendor costs.
• Reassess ROI after 12–18 months of operation — not just at launch.
• Consider hybrid approaches: migrate critical parts first, not the entire stack.
• Build financial models that account for both CAPEX and rising OPEX.

Conclusion: pragmatism over hype

The benefits of headless commerce are real, but conditional. They require organizational maturity, technical discipline, and a clear understanding of business goals.

At Striped Giraffe, we’ve seen headless commerce succeed — but only when companies approach it as a strategic investment, not a checkbox on a digital roadmap.

When headless commerce makes sense

Headless commerce delivers strategic value when organizations have:

• Multiple channels, brands, or regions with diverging experience or business requirements
• Mature DevOps and CI/CD capabilities, including automated testing and deployment
• Clear domain ownership across ERP, PIM, pricing, customer data, and fulfillment
• Strong API governance, including versioning, monitoring, and lifecycle management
• A business model that actively leverages flexibility (not just plans for it)
• Sufficient scale where modularity reduces long-term constraints rather than adding overhead

In these scenarios, headless becomes an enabler of speed, differentiation, and controlled evolution.

When headless commerce often doesn’t (yet)

Headless commerce tends to underperform when:

• The organization operates one or two stable channels with limited variation
• Legacy backend systems already struggle with data consistency and performance
• Team capacity is limited and stretched across operations and delivery
• Governance, API discipline, and data ownership are not clearly established
• Flexibility is adopted as a precautionary measure, not a concrete business need

In such cases, a well-designed integrated or hybrid commerce architecture often delivers lower risk, lower cost, and faster ROI.

You might also like:

• Agentic Commerce, Part 1: Where Conversations Become Transactions » Learn more
• E-Book: Digital Self-Service & Automation in B2B E-Commerce » Learn more
• The Tech Backbone of Subscription-Based E-Commerce » Learn more
• Strategic Patterns of IT System Integration » Learn more
• Low-Code Falls Short Without Full-Code » Learn more

Few technologies manage to reconcile two opposing imperatives — the hunger for insight and the duty of discretion. Data anonymization does.

Data anonymization has quietly become one of the most powerful instruments of digital cooperation. It allows data to move — between teams, companies, and even competitors — without crossing the lines of privacy. What once was seen as a defensive measure is now emerging as a catalyst of progress.

When applied with precision and context awareness, anonymization transforms data that cannot be shared into knowledge that can be acted upon — as we showed in the first part of this series, where we explored the underlying principles, methods, and trade-offs behind effective anonymization.

The growing adoption of anonymization across industries marks a quiet shift in how organizations perceive data privacy. It’s no longer just about compliance or ethical responsibility. It’s about creating a trusted space where innovation can happen — responsibly, collaboratively, and at scale.

🚘 Automotive and Mobility

Modern vehicles are moving data centers. Every mile generates images, GPS coordinates, sensor readings, and behavioral traces — all of which may expose identifiable details such as faces, license plates, home addresses, or driving patterns.

As automakers collaborate with AI startups, mapping providers, and regulators, data anonymization becomes the passport for cross-organizational innovation without breaching privacy laws such as the GDPR.

Typical data anonymized

• Video streams (faces, license plates, pedestrians)
• GPS traces and location metadata
• Vehicle identifiers linked to telematics logs
• Sensor fusion data when combined with route or time markers

Potential use cases

• Anonymization of vehicle sensor and camera data for internal testing and validation of autonomous driving and advanced driver-assistance systems (ADAS).
• Processing connected vehicle telematics to monitor fleet performance and predictive maintenance while masking individual driver identities.
• Exchanging anonymized sensor and environmental data between automotive manufacturers for collaborative development of autonomous driving algorithms and safety testing.
• Sharing anonymized telematics and mobility data with urban planners to optimize traffic flow, parking infrastructure, and public transport systems.

Real-world initiatives

• Audi A2D2 (Audi Autonomous Driving Dataset) applied automated blurring to faces and license plates before releasing its perception dataset publicly. This allowed external researchers to work with real-world driving data without violating EU privacy standards.
• Companies like Brighter AI and Celantur specialize in privacy-preserving video anonymization for connected vehicles — using AI-based pixel replacement or synthetic overlays instead of simple blur to prevent re-identification. These technologies enable OEMs to share test data with engineering partners safely.
• European mobility projects such as SimRa (Safety in Bicycle Traffic) — led by Technische Universität Berlin — use anonymized route and event data from cyclists to support urban planning and traffic safety research, showing how anonymization fuels collaboration between industry, academia, and cities.

Risks

Location data can still enable re-identification when combined with external sources (e.g., home addresses, parking habits). Privacy-by-design frameworks such as ISO/TS 29003 and EDPB guidance recommend context-specific aggregation and perturbation to mitigate this.

🏛️ Banking and Finance

In financial services, transaction histories, behavioral analytics, and credit records are rich in personal insights — but also highly sensitive. Banks must balance innovation in fraud detection and AI-driven credit scoring with regulatory obligations like PSD2 and GDPR.

Typical data anonymized

• Payment transactions (names, account numbers replaced or hashed)
• Credit histories and spending patterns
• Customer demographics tied to location or income

Potential use cases

• Anonymization of customer transaction and credit data for internal fraud detection, risk assessment, and predictive modeling.
• Masking sensitive information in customer behavior datasets to enable internal analytics and product optimization.
• Providing anonymized transactional datasets to fintech companies or regulators for fraud pattern detection, financial risk modeling, and compliance benchmarking.
• Sharing anonymized credit behavior data between banks to support collaborative credit scoring models or market trend analyses.

Real-world initiatives

• The Bank of England’s anonymized transaction data sandbox allows researchers to analyze spending behavior without access to identifiable customer records.
• Several major European banks collaborate with fintech startups through shared anonymized datasets to improve fraud models. Data is masked or tokenized before leaving internal systems, allowing third parties to detect suspicious patterns without seeing account identifiers.
• The European Central Bank (ECB) and the European Banking Authority (EBA) have both endorsed the use of anonymized financial data in regulatory sandboxes to support responsible innovation.

Risks

Even de-identified financial data can be vulnerable to linkage attacks if patterns are unique (e.g., rare spending combinations). Re-identification studies have shown that as few as three transactions can identify a customer with >80% probability if combined with external data. Continuous monitoring of k-anonymity thresholds is essential.

🏗️ Construction and Smart Infrastructure

As construction becomes digitalized, projects rely on IoT devices, smart sensors, and wearables — all of which collect information about workers, building occupancy, and environmental conditions. These datasets are crucial for safety, efficiency, and sustainability, but they also contain traces of personal information.

Typical data anonymized

 IoT sensor readings (linked to time, room, or device ID)

• Worker wearables (biometric or motion data)
• Building usage and location-based metrics

Potential use cases

• Anonymization of sensor data from smart buildings, including environmental and occupancy measurements, for internal monitoring of energy usage, occupancy, and site efficiency.
• Processing wearable IoT data from construction workers in an anonymized manner to monitor safety, ergonomics, and health without compromising personal identities.
• Sharing anonymized construction site or building performance data with research institutions, urban planners, or regulatory bodies for collaborative studies, safety benchmarking, and sustainable development initiatives.

Real-world initiatives

• Skanska and Autodesk have experimented with anonymized wearable data to track safety compliance and ergonomics on construction sites. Personal identifiers are stripped before data aggregation to focus only on movement patterns and hazard events.
• Smart building projects in Europe, such as EU Horizon’s SmartBuilt4EU, use anonymized sensor data (temperature, CO₂, occupancy) to optimize energy management while preserving occupant privacy.

Risks

Combining location, time, and movement data may still allow re-identification of individuals (e.g., unique shift patterns). Aggregation and temporal generalization help reduce these risks.

🛒 E-commerce

Every click, view, and purchase in e-commerce reveals behavioral insights. Yet linking this data to identifiable customers exposes privacy risks — especially when collaborating with marketing partners or AI vendors. Anonymization enables analytics that respect identity boundaries.

Typical data anonymized

• Browsing histories and session logs
• Transactional and payment records
• Device IDs and behavioral sequences

Potential use cases

• Anonymizing customer browsing, purchasing, and interaction data to enable internal analytics without exposing personally identifiable information.
• Use of anonymized transaction data to detect and prevent fraudulent activities while safeguarding customer privacy.
• Sharing anonymized sales and transaction data across e-commerce platforms to improve fraud detection models and consumer behavior research.

Real-world initiatives

• Amazon’s differential privacy research and Google’s open-source differential privacy frameworks have influenced retail analytics, where anonymization supports trend analysis without revealing user-level behavior.
• European e-commerce consortia have experimented with sharing anonymized fraud data to train collective detection models — using hashed transaction IDs and synthetic augmentation to prevent exposure of actual users.
• Shopify applies pseudonymization and anonymization layers to aggregated merchant and consumer analytics to comply with GDPR while still enabling product optimization.

Risks

Even anonymized behavioral data can allow inference of identities if cross-referenced with known purchase patterns. Combining anonymization with synthetic data generation mitigates linkage risks.

💊 Healthcare and Pharmaceuticals

No domain handles more sensitive data than healthcare. Patient records, genetic profiles, and clinical trial data are lifelines of research — but also legally protected under GDPR, HIPAA, and EMA guidelines.

Typical data anonymized

• Electronic health records (EHRs) and lab results
• Imaging data (CT, MRI, pathology slides)
• Genomic sequences and drug usage logs

Potential use cases

• Anonymization of patient records and clinical trial data for internal research, AI model training, and service improvement.
• Processing anonymized telemedicine and wearable health data to monitor outcomes and improve care delivery while protecting privacy.
• Sharing anonymized health and drug usage datasets between hospitals, research institutions, and regulatory authorities to advance epidemiological studies, pharmacovigilance, and collaborative research.
• Collaborative development of AI models for diagnostics or treatment prediction using anonymized multi-institutional health datasets.

Real-world initiatives

• ClinicalTrials.gov, operated by the U.S. National Institutes of Health (NIH), requires anonymization of all published clinical trial records to ensure research transparency without compromising participant privacy.
• The YODA Project (Yale Open Data Access), run by Yale University, provides a framework for sharing anonymized clinical trial data between pharmaceutical companies and independent researchers under clear ethical governance.
• European Medicines Agency (EMA) enforces anonymization of patient-level data in submissions made public under its data transparency policy.
• UK NHS Digital’s Data Access Environment (DAE) provides researchers with de-identified datasets for epidemiological studies and AI development under strict governance.

Risks

Re-identification risks persist when indirect identifiers — such as rare diseases, treatment timelines, or genetic variants — can be cross-matched with public data. EMA recommends quantitative risk assessment and expert review before release.

Conclusion

Anonymization doesn’t strip data of its meaning. It refines it — distilling only what’s needed to drive understanding while removing what should never leave the source.

Across all discussed sectors, anonymization not only supports regulatory compliance but also enables organizations to unlock the value of sensitive data safely, facilitating innovation, research, and operational efficiency.

As industries grow more interconnected, the ability to separate identity from insight will decide who can innovate safely — and who will be left behind, constrained by their own data silos.

In this sense, anonymization is not an act of concealment. It is an act of liberation — allowing information to flow, connect, and create value without crossing ethical or legal boundaries.

Data anonymization empowers organizations to unlock insights while preserving privacy, enabling safe collaboration across teams and companies to enhance analytics, security, research or product development.

In a world where every transaction, click, and interaction generates data, the concept of anonymity has become paradoxical. Data is the backbone of innovation, yet it is also one of the greatest sources of corporate risk. As organizations strive to comply with tightening privacy regulations and public expectations, anonymization is no longer a technical afterthought — it’s a strategic necessity.

But anonymization is not a single technique or tool. It’s a complex, multi-layered process that must reconcile three often conflicting goals: privacy, usability, and compliance. This article explores anonymization from a technological and operational standpoint — from the foundational methods and emerging approaches to the real-world challenges of keeping systems functional while data loses its identity.

1. Definition and Scope of Data Anonymization

Anonymization goes far beyond replacing names or emails with random characters. It refers to the irreversible transformation of personal data so that individuals can no longer be identified — directly or indirectly.

What distinguishes true anonymization is irreversibility. Once data is anonymized, no technical or organizational measures should make re-identification possible. This is what separates it from pseudonymization, where identifiers are replaced but can later be restored with a key. Encryption, too, is not anonymization — it protects data in transit or storage but does not eliminate personal attributes.

Key aspects of anonymization’s scope include:

• Granularity control: deciding how much detail can be safely retained without risking re-identification.
• Context awareness: ensuring anonymized datasets remain non-identifiable even when combined with other available information.
• Purpose limitation: balancing privacy with the need to preserve analytical or testing value.

Anonymization is thus not a one-time process but an engineering discipline — one that intersects data architecture, governance, and regulatory strategy.

2. Regulatory and Legal Context

Modern privacy frameworks have elevated anonymization from a compliance checkbox to a legal demarcation line between personal and non-personal data.

• GDPR: Anonymized data falls outside the scope of the regulation, but only if re-identification is “reasonably unlikely.”
• CCPA: Follows a similar principle, focusing on the impossibility of re-linking data to an individual.
• HIPAA, ISO/IEC 20889: Define specific criteria and methodologies.

In practice, regulators assess risk and reasonableness, not theoretical perfection. Organizations that treat anonymization as a dynamic process — continuously reviewed and tested — stand the best chance of staying compliant.

3. Core Techniques of Anonymization

At the heart of anonymization lie a set of mathematical and statistical techniques designed to disrupt identifiable patterns while retaining utility.

Common categories include:

• Data masking (static or dynamic): replaces real values with fictional but plausible data.
• Generalization and suppression: techniques like k-anonymity, l-diversity, and t-closeness reduce data granularity to make individuals indistinguishable within a group.
• Randomization: introduces controlled noise or swaps attributes between records to obscure identity.
• Aggregation: merges individual-level data into grouped insights.

While each method has merits, their effectiveness depends on context and purpose. For instance, k-anonymity works well for tabular data but can fail under correlated attributes. Masking is fast and simple but may break dependencies critical for analytics or testing.

True anonymization usually combines multiple methods, guided by:

• Risk assessment — probability of re-identification.
• Data type — structured, semi-structured, or unstructured.
• Intended use — reporting, model training, or test automation.

4. Advanced Approaches

Emerging technologies are redefining what’s possible in anonymization, moving beyond static transformations toward privacy-by-design architectures.

Key innovations include:

• Differential privacy: ensures that statistical outputs do not reveal information about any individual, even indirectly. Used by organizations like Apple and the U.S. Census Bureau, it quantifies privacy loss mathematically.
• Synthetic data generation: uses artificial intelligence (AI) or machine learning (ML) to produce data that mimics real datasets without exposing actual records.
• Federated learning: allows AI models to be trained locally across devices or systems without sharing raw data.

These approaches shift the focus from altering existing data to redesigning how data is collected, shared, and analyzed. However, they come with caveats — synthetic data can still embed latent biases, and differential privacy requires expert calibration to avoid degrading analytical quality.

In mature organizations, these methods often complement, not replace, traditional anonymization.

5. Re-identification Risks and Attacks

No anonymization is immune to re-identification. As computing power, data availability, and correlation techniques advance, even well-designed transformations face new risks.

Typical threats include:

• Linkage attacks: combining anonymized datasets with external sources (e.g., social media, public records) to infer identities.
• Inference attacks: deducing personal attributes through statistical correlations.
• Background knowledge attacks: leveraging domain-specific information to narrow down candidates.

Real-world failures — such as the AOL search data leak (2006) or the Netflix Prize dataset (2008) — showed how easily “anonymous” records can be traced back to individuals. In 2021, a particularly explosive incident occurred at the French software manufacturer Dedalus Biologie (France): software migration errors resulted in the intimate medical data of almost 500,000 people being made publicly accessible. This led to a significant loss of reputation and legal consequences.

Organizations must therefore adopt risk-based anonymization, recognizing that:

• Re-identification potential grows as datasets become richer and more interconnected.
• Periodic testing and adversarial evaluation are crucial.
• Security through obscurity (hiding methods) is ineffective; transparency and validation matter more.

In short, anonymization is never absolute — it is a continuum that must evolve alongside threats.

6. Balancing Privacy and Data Utility

Anonymization is a trade-off — every layer of protection reduces the richness and precision of data.

Key dimensions of this balance:

• Analytical utility: How much insight remains after anonymization?
• Process continuity: Will downstream applications still function correctly?
• Regulatory defensibility: Can you prove compliance under audit?

Organizations often misjudge this balance by applying excessive transformation that renders data useless, or by underestimating re-identification risk.

The solution lies in contextual modeling:

• Define explicit use cases (e.g., A/B testing, AI training, performance monitoring).
• Choose anonymization levels appropriate for each.
• Continuously measure utility loss through predefined metrics.

When done right, anonymization doesn’t destroy data — it liberates it from legal and ethical constraints, enabling innovation without compromising trust.

7. Technological Implementations

Effective anonymization requires seamless integration across the data ecosystem — from ingestion to analytics.

Modern implementations typically involve:

• ETL integration: embedding anonymization in Extract-Transform-Load pipelines ensures consistent transformation before data enters downstream systems.
• Automation frameworks: scheduling anonymization jobs and managing dependencies.
• Cloud-native anonymization: leveraging managed services for scalable processing.
• Metadata-driven configuration: defining transformation logic via data catalogs or governance tools.

Leading platforms — such as Informatica TDM, Delphix, or SAP TDMS — provide structured workflows, but each demands customization to fit enterprise data models. Open-source solutions (e.g., ARX, sdcMicro) offer flexibility for organizations with strong in-house expertise.

Ultimately, success depends less on the toolset and more on design discipline — aligning anonymization logic with architectural principles, security policies, and operational workflows.

8. Data Anonymization in Test Environments

One of the most overlooked challenges in anonymization arises in test environments — where data realism meets privacy risk.

Most enterprises rely on production-like datasets to test new features, simulate user journeys, or reproduce bugs. These datasets often contain real personal data, copied from production systems. Yet anonymizing them is far from trivial:

• Every interconnected system in the test landscape must receive consistent anonymized data, or business processes will fail.
• Referential integrity must be preserved — customer IDs, order links, and transactional chains cannot simply be randomized independently.
• There is no universal tool to handle this automatically; every organization’s architecture is unique.

The process demands a comprehensive dependency analysis across all systems — identifying how entities interact, where personal identifiers appear, and how transformations cascade.

A robust strategy includes:

• Building repeatable anonymization pipelines that maintain consistency across environments.
• Using tokenization or deterministic mapping where relationships must persist.
• Testing anonymized data with full business logic to confirm functionality.

This complexity makes test data anonymization one of the most time-consuming and expertise-heavy stages — yet also one of the most critical for compliance and security.

9. Future Trends and Emerging Technologies

The future of anonymization will be shaped by AI-driven analytics, regulatory evolution, and quantum computing capabilities. Static methods will no longer suffice; dynamic, hybrid approaches will dominate.

Emerging directions include:

Differential privacy at scale

Pioneered by tech leaders such as Apple, Google, and Microsoft, differential privacy mathematically limits the risk of exposing any individual’s data.

This approach relies on a parameter called the “privacy budget” (ε), which controls how much information about a single person can influence the output. Choosing the right value for ε is critical: a smaller value increases privacy but reduces data accuracy, while a larger value improves accuracy but weakens privacy protection.

Synthetic data generation

Using GANs (Generative Adversarial Networks) and variational autoencoders, synthetic datasets are becoming critical in sectors like healthcare, automotive testing, and AI model training. The challenge remains ensuring statistical fidelity without leaking identifiable attributes.

Federated learning

Enables machine learning on decentralized data — training models locally and aggregating only insights. This reduces reliance on central anonymization, though metadata leakage remains a risk.

Privacy-preserving computation

Privacy-preserving computation encompasses methods that allow organizations to analyze and process data without exposing sensitive information. By enabling computations on anonymized or encrypted datasets, these techniques ensure that individual-level data remains protected while insights can still be extracted.

Key approaches include:

• homomorphic encryption, which allows calculations on encrypted data
• secure multiparty computation, enabling joint data processing among multiple parties without revealing raw inputs
• trusted execution environments (TEE), isolated areas in processors where data can be securely processed.

These methods collectively expand the analytical capabilities of organizations while preserving privacy.

Post-quantum considerations

Advances in quantum computing pose a potential threat to current cryptographic and anonymization techniques. Algorithms that are considered secure today may become vulnerable, allowing future attackers to re-identify or decrypt previously protected data.

Organizations must anticipate this shift by evaluating the resilience of their anonymization and encryption methods against quantum-enabled attacks. This includes assessing the potential longevity of privacy protections in long-term data storage and processing scenarios.

Emerging post-quantum cryptography offers solutions designed to resist quantum attacks. Implementing these methods proactively ensures that sensitive data remains secure over time, maintaining compliance and trust even as computational capabilities evolve.

Conclusion

Anonymization is not merely a compliance checkbox. It is a strategic enabler of responsible data innovation.

Key takeaways for executives:

• Anonymization reduces, not eliminates, risk. Residual re-identification threats must be continuously assessed.
• Governance and accountability are crucial. Without auditability and oversight, even strong techniques can fail.
• Context matters. Industry, dataset type, and processing goals determine the right anonymization method.
• Technology is evolving fast. AI, synthetic data, and privacy-preserving computation redefine the landscape.
• Trust is strategic. Ethical, transparent anonymization strengthens brand credibility and customer loyalty.

The ultimate question for leadership is not “Have we anonymized our data?” — but “Are we governing anonymization as a living, adaptive process?”

Universal Commerce Protocol (UCP) was designed to reshape online retail, but its broader significance lies in what it reveals about the future of B2B buying.

UCP was introduced by Google as a foundation for agent-driven commerce in online retail. Its goal is clear: standardize how products are discovered, compared and purchased by customers assisted by AI agents such as Gemini.

For B2B, the challenge is not agent-driven buying itself, but the gap between what UCP supports today and the complexity B2B commerce requires.

Yet focusing only on what UCP cannot do today misses the bigger picture.

UCP is not a ready-made solution for B2B commerce — but it is a signal of where buying orchestration may be heading as agents become first-class participants in commercial transactions.

Why UCP does not fit B2B commerce today

UCP was designed to support agent-driven sales in online retail, where buying is fast, transactional and centered on standardized product offers rather than negotiated relationships.

In its current form, UCP does not support key B2B requirements such as:

• corporate identity and account hierarchies
• contract-based assortments and customer-specific pricing
• governed purchasing workflows (approvals, policies, PunchOut, EDI)
• settlement models based on POs, invoicing, payment terms and credit limits

These elements are foundational to B2B commerce and typically live in ERP, procurement and finance systems.

Because UCP does not model or integrate with this layer today, it cannot support B2B buying beyond isolated, non-governed transactions.

What agentic commerce could unlock for B2B

Although the current version of UCP is not yet adapted to the full complexity of B2B commerce, the agentic commerce paradigm it introduces unlocks structural potentials that could become transformative if the protocol and ecosystem evolve in the right direction.

Autonomous procurement agents across suppliers

Imagine a procurement environment where the most time-consuming B2B tasks — supplier comparison, price negotiation, order placement — are handled by AI agents on behalf of organizations.

In a future iteration of agentic commerce:

• A corporate procurement agent could use UCP to query multiple suppliers with a single request, passing authenticated corporate credentials to fetch contract-specific terms.
• Suppliers would respond with real-time availability, negotiated pricing and delivery terms tied to that buyer’s profile.
• The agent could then compare alternatives, balance cost versus availability, and present the best options for human approval.

This model effectively turns multiple siloed supplier portals into one normalized discovery and procurement layer.

Automated replenishment and supply chain orchestration

In traditional B2B operations, reorder decisions often depend on manual monitoring or rigid automation tied to internal systems. With agentic commerce:

• IoT-enabled devices (e.g., smart shelves, factory sensors) could detect low stock and initiate a replenishment task.
• An AI agent — speaking UCP — could handle the entire transaction without bespoke integration into each supplier’s API.
• Because UCP standardizes the commerce interface, these transactions could be executed consistently across vendors without custom coding.

This is similar to emerging thinking about AI-assisted negotiation and autonomous transaction execution in procurement, although current research emphasizes privacy-preserving agent negotiation frameworks rather than specifically commerce protocols.

Intelligent quote and contract workflows

Legacy B2B commerce relies heavily on manual quotes and negotiation cycles. Agentic AI plus standardized commerce protocols could reshape this by:

• Letting agents request and compare quotes across suppliers using a common semantic layer.
• Returning structured offers that agents can score, reason about constraints and conditions, and even simulate negotiation outcomes.
• Integrating with human review loops where final approvals or adjustments happen in interface surfaces tied to procurement systems.

This aligns with broader research on AI negotiation workflows in B2B contexts, where agents autonomously screen, negotiate, and escalate only when needed.

Multi-stage payment and settlement extensions

Right now, UCP supports modern payment methods like wallet tokens or credit cards, which is sufficient for retail checkouts. B2B, however, demands richer payment orchestration:

• Purchase orders and invoicing cycles.
• Net terms, credit limits, split payments, milestone payments.
• Integration with accounting and ERP settlement systems.

A future where B2B agentic AI thrives might include commerce protocols that allow:

• PO fields and approval triggers built into transaction primitives, so agents can initiate a PO flow rather than a one-off charge.
• Deferred settlement tokens that represent future invoicing conditions.
• Hooks for financial workflow integration so that agents can initiate the transaction, but reconciliation, credit checks, and compliance steps are executed with back-office systems in lockstep.

This is consistent with industry thinking that agentic payments must balance autonomy with governance and auditability.

Data becomes the primary interface

In an agent-driven commerce model, the buying interface changes fundamentally. Humans no longer browse. Agents query.

As a result, data replaces UX as the primary interface to the buyer.

When AI agents are responsible for discovering products, comparing suppliers, and initiating transactions, visual design, navigation patterns, and persuasive content lose importance. What matters instead is whether an agent can understand, trust, and act on the data it receives.

Why data quality matters more than ever

B2B commerce already depends heavily on data. Agentic commerce raises the bar significantly.

For an AI agent, product and commercial data must be:

• complete
• consistent
• structured
• machine-readable

Any ambiguity becomes friction. Any inconsistency becomes a reason to move on.

Agents do not negotiate around missing attributes or unclear availability. They simply skip suppliers whose data cannot be reliably interpreted.

From “nice to have” to competitive requirement

In an agentic B2B context, competitive advantage shifts toward organizations that can provide:

• clean, well-structured product data
• stable identifiers across ERP, MDM, PIM and commerce
• accurate availability and lead-time signals
• pricing and terms that can be interpreted programmatically

This is not about exposing everything publicly. It is about being ready to respond with precise, trusted data when an authenticated agent requests it.

The hidden role of MDM and ERP foundations

This is where many B2B organizations will struggle. Agentic commerce does not tolerate:

• duplicated product records
• inconsistent units of measure
• fragmented price logic
• loosely synchronized stock data

What used to be internal data hygiene issues become external decision factors.

If agents cannot reconcile data across systems, suppliers effectively become invisible.

In that sense, Master Data Management and ERP integration stop being back-office concerns. They become revenue-critical capabilities.

Final thoughts

UCP will not transform B2B commerce overnight. It is not yet ready to support the full complexity of contract-driven pricing, procurement governance, or settlement workflows that define B2B today.

But it is a clear signal of where the market is heading.

As buying decisions are increasingly delegated to AI agents, the mechanics of competition change.

When agents buy:

• feeds matter more than frontends
• structure matters more than storytelling
• data quality determines discoverability

In that context, data stops being an internal concern and becomes a form of market access.

This does not mean B2B organizations should redesign procurement around UCP today. PunchOut, EDI and procurement platforms are not going away in the near term.

What does make sense now is preparation.

B2B leaders should focus on:

• auditing and cleaning product, pricing and availability data
• modelling identity, entitlements and contracts consistently
• reducing bespoke, brittle integrations that block automation
• closely tracking the evolution of agentic standards and protocols

UCP is not a roadmap. It is an early indicator of a structural shift.

In the future shaped by agentic commerce, success will not belong to those with the most sophisticated storefronts. It will belong to those whose data, rules and systems are ready to be understood — and trusted — by machines.

You might also like:

• Agentic Commerce, Part 1: Where Conversations Become Transactions » Learn more
• Agentic Commerce, Part 2: Winners, Losers, and the Forces Behind the Change » Learn more
• Agentic Commerce, Part 3: Strategic Priorities for Executive Leaders » Learn more
• Our expertise: Artificial Intelligence » Learn more
• Pharma’s AI Readiness Index: Who Leads the Race? » Learn more
• AI-Driven Personalization in B2B E-Commerce: From Recommendation Logic to Real-Time Experience » Learn more

Data ownership is no longer the key to value creation – those who want to unlock business potential must share.

Whether driven by regulation, sustainability goals, or efficiency gains, companies in regulated industries are increasingly required to share data in a structured, secure, and collaborative manner. But effective data sharing requires more than technology – it demands new mindsets, governance models, and clearly defined responsibilities.

Data Sharing ≠ Data Disclosure

A common misconception: Data sharing does not mean unrestricted access to all information. Rather, it involves providing controlled and contextual access to specific datasets, governed by well-defined roles, rules, and responsibilities. This is the only way to meet regulatory requirements and achieve strategic business outcomes.

1. Regulatory Data Obligations as a Starting Point

In regulated sectors, data sharing is no longer optional – it’s a legal obligation, often with high technical complexity and strict audit requirements.

Typical examples:

• Pharma & Life Sciences: Clinical trial and safety data submissions (EudraVigilance, SPOR, IDMP).
• Financial Services: Risk and ESG reporting under CRR/CRD and the new EBA reporting frameworks.
• Aviation & Maritime: CO₂ emissions data to ICAO, CORSIA, and THETIS-MRV.
• Chemical Industry: Joint data registration under REACH (SIEFs, REACH-IT).
• Agri-Food: Traceability via TRACES and compliance with EU Regulation 178/2002.

👉 IT service providers support with:

• Integration with regulatory submission platforms
• Automation of data extraction and validation
• Secure storage of audit-relevant data

Important: Without structured, traceable, and machine-readable data, regulatory compliance – and even market access – becomes impossible.

2. ESG Data as a Reputation Currency

Beyond compliance requirements like CSRD and the EU Taxonomy, new voluntary ESG standards are gaining influence – including CDP, SBTi, GRI, and TNFD.

Why it matters: ESG transparency is quickly becoming a prerequisite for B2B partnerships, investments, and financing. Companies unable to provide Scope 3 emissions data risk being excluded from tenders.

👉 IT service providers enable:

• Collection and aggregation of ESG-relevant data
• Transformation into standardized disclosure formats
• Integration with ESG portals and supplier systems

ESG reporting is shifting from static PDFs to continuous, API-based data flows. ESG is becoming an infrastructure task, not just a communications one.

3. Supply Chain Transparency as a Risk Lever

Effective risk management in procurement now requires structured data exchange beyond Tier 1 – such as:

• Production capacity insights
• Transport delays
• Material availability
• ESG and origin documentation

The automotive sector is leading the way with data-driven collaboration networks like Catena-X, a federated platform for secure data sharing across the value chain.

👉 IT service providers offer:

• Data capture and monitoring across the entire supply chain
• Integration with logistics or sustainability platforms
• Dashboards for real-time risk analytics

Key point: Resilient supply chains require connected data streams – not disconnected spreadsheets.

4. Data Accompanies Products Throughout Their Lifecycle

With upcoming Digital Product Passports (DPP) under the EU Green Deal, products become data carriers. From material composition and repairability to recyclability, machine-readable information must accompany products throughout their lifecycle.

👉 IT-Dienstleister leisten hier:

• DPP-compliant data models
• Integration with federated platforms (e.g. EUDPP)
• Semantic interoperability across supply chains

📌 Circular economy also means circular data. Without digital continuity, true circularity cannot be achieved.

5. Research & Innovation Platforms on the Rise

Some of the most forward-looking data spaces are being built not by mandate, but through voluntary cooperation:

• EHDS (European Health Data Space): Shared health data for research and care delivery
• Manufacturing-X: Industrial data sharing to accelerate AI, transparency, and sustainability
• Open Manufacturing Platform, IDSA, Gaia-X: Federated, open reference architectures for data ecosystems

👉 IT service providers contribute to:

• Building sovereign data spaces aligned with Gaia-X principles
• Defining roles, policies, and access controls
• Providing trust services and identity management

Core principle: Controlled openness instead of full disclosure. Data spaces enable selective sharing without loss of control.

Data Sharing: The Emerging Core Competency

Across all use cases, one insight prevails:

Owning data in isolation is no longer a competitive advantage. Value emerges through connected thinking, open architectures, and regulatory readiness.

To make this shift, organizations need:

• Data governance instead of silos
• Platform capabilities over point solutions
• And most of all: a clear roles-and-architecture strategy

🔜 Preview: Part 3 of our series

In the next article, we explore how organizations can technically, legally, and operationally prepare for data sharing – from selecting interoperable APIs to defining access control models and involving external partners.

Let us support you move from data potential to data performance.

You might also like:

• 7 key benefits of data sharing in e-commerce » Learn more
• Data sharing — Challenges and opportunities » Learn more
• Data governance – the linchpin of efficient data management » Learn more
• Digital Product Passport: Tips to Unlock New Opportunities in E-Commerce » Learn more
• Digital Product Passport: Data Management Challenges » Learn more

A year of AI reckoning, tech value correction, and human expertise revival.

2026 will be a year of correction — not collapse — for artificial intelligence. After a decade of soaring expectations, Forrester predicts a decisive shift toward value discipline, governance maturity, and renewed trust in human expertise. Budgets will tighten, talent shortages will intensify, and regulatory pressure will climb. Yet amid this friction lies the real story: AI is finally being forced to prove its worth.

In this article, we examine Forrester’s key predictions for 2026 and pair them with expert analysis from our leadership team. Their commentary goes beyond the headlines, offering concrete recommendations for CEOs, CIOs, CFOs, and business executives navigating one of the most pivotal transformations in enterprise technology.

IT Investment

PREDICTION:

25% of AI spend by enterprises will be delayed into 2027

Despite bold investment, AI returns remain elusive — only a fraction of firms can link AI to profit growth. As CFOs tighten oversight, projects with unclear ROI will stall, shifting one-quarter of planned budgets into 2027. This slowdown will expose the gap between marketing hype and business impact, forcing organizations to prioritize measurable value and renegotiate inflated vendor costs.

AI governance

PREDICTION:

1/4 of CIOs will be asked to bail out business-led AI failures

When AI initiatives falter, CEOs will turn to CIOs to restore governance, accuracy, and trust. As agentic systems spread, firms will realize only tech leaders can align teams, curate data, and enforce quality. The role of the CIO will expand beyond technology to include AI ethics, data strategy, and cross-department collaboration, preventing costly missteps before they escalate.

AI literacy

PREDICTION:

30% of large enterprises will mandate AI training to boost adoption and reduce risk

As AI literacy becomes essential for productivity and compliance, companies will launch structured training programs. Mandatory education will raise the organization’s artificial intelligence quotient (AIQ), support responsible use, and prevent costly errors in regulated sectors. Firms will partner with vendors to build ongoing learning ecosystems.

IT talent strategy

PREDICTION:

The time to fill developer positions will double

Hiring developers will become slower and more complex. Companies will seek candidates who can pair with AI tools or design AI-driven architectures, shrinking available talent pools. HR teams will face a flood of AI-generated applications and respond with stricter screening. To stay competitive, firms must invest in junior talent, train internally, and use AI to improve recruitment efficiency and accuracy.

Tech value management

PREDICTION:

2/3 of CIOs will need to justify budgets by linking tech spend to business value

As AI, cloud, and security costs rise, mapping spend to business impact becomes essential. Expect growing adoption of frameworks like technology business management (TBM) and practices like FinOps, supported by AI-driven analytics that automate attribution and reporting. In 2026, technical fluency must be matched by financial storytelling.

Shadow AI in B2B

PREDICTION:

Ungoverned generative AI in commercial apps will cost B2B companies more than $10 billion

Uncontrolled generative AI adoption across marketing, sales, and product teams will trigger data leaks, compliance breaches, and stock declines. Existing governance models can’t manage AI built into third-party tools. Organizations must raise employee “AI intelligence,” apply decentralized controls, and teach users to spot flawed outputs early.

B2B buyer interactions

PREDICTION:

Human expertise will rival generative AI in appeal as buyers seek deeper validation

As B2B buyers face information overload from generative AI tools, they’ll increasingly seek validation from human experts. Trust and nuanced understanding will make expert interactions critical again — especially in complex purchases. Vendors must upskill product specialists and customer success teams to offer insight that algorithms can’t replicate.

Machine customers in B2B

PREDICTION:

20% of B2B sellers will be forced to engage with AI-based buyer agents

B2B procurement agents will autonomously negotiate prices, terms, and service levels across hundreds of suppliers at once. These AI systems are tireless, data-driven, and fully compliant by design. Static pricing and manual quote handling will disappear, replaced by adaptive, agent-to-agent negotiations. Suppliers must respond with their own AI agents to stay profitable and competitive.

B2B payments

PREDICTION:

AI agents will be involved in about one-third of B2B payment workflows

B2B payments are becoming a proving ground for agentic AI. By 2026, one-third of transactions will involve autonomous agents managing invoicing, reconciliation, or spend control. These systems promise greater efficiency and fewer human errors than traditional automation, helping finance teams reallocate effort from manual checks to strategic decision-making.

EU legislation

PREDICTION:

EU law simplification will fail to deliver enterprise cost savings

The EU’s 2025 “Omnibus” plan promised to ease compliance costs for smaller firms, but those savings are unlikely to materialize. Large corporations still face full regulatory duties and will push due-diligence demands down their supply chains. As a result, smaller firms — though formally exempt from some rules — will need to stay compliant to keep business partners satisfied. Assess supplier impacts early to control costs.

AI adoption in Europe

PREDICTION:

Daily use of generative AI by Europeans will double, however, enterprise adoption will lag the US

Generative AI will become embedded in daily life, from devices to productivity apps, doubling everyday use across Europe. Yet enterprise rollout will trail the US by around 10%, slowed by regulation, skill gaps, and delayed vendor launches. To close the gap, European firms must invest in training, governance, and cross-border AI collaboration.

Conclusion

Forrester’s outlook for 2026 is clear: the AI era is entering a phase where discipline matters more than experimentation. The winners will not be the companies that spend the most, but the ones that align technology with measurable economic value, strong governance frameworks, and trusted human expertise.

Across every prediction — from agentic payments to shadow AI, from developer scarcity to European regulation — one theme is consistent: AI will only scale when organizations pair automation with accountability, and innovation with operational clarity.

As enterprises move into 2026, those that standardize their data, elevate their workforce, and build transparent, resilient AI ecosystems will transform uncertainty into competitive advantage. Others will spend the year catching up.

You might also like:

• Agentic Commerce, Part 1: Where Conversations Become Transactions » Learn more
• Agentic Commerce, Part 2: Winners, Losers, and the Forces Behind the Change » Learn more
• Agentic Commerce, Part 3: Strategic Priorities for Executive Leaders » Learn more
• Generative AI unlocks the value of unstructured data in insurance » Learn more
• Pharma’s AI Readiness Index: Who Leads the Race? » Learn more

Agentic Commerce is more than a technical innovation — it’s a structural change in how products are discovered, recommended, and purchased.

For senior leaders, the challenge is no longer whether to act, but how to build the technical, data, and governance foundations that will keep their businesses visible and trusted in AI-driven channels.

The checklist below outlines key strategic moves to prioritize in the coming quarters:

Structure product data for AI consumption

Your catalog is no longer just a marketing asset — it’s the language your brand speaks to AI agents.

Ensure every product includes complete, machine-readable attributes: dimensions, materials, compatibility, use cases, and high-quality imagery. The richer and more structured your data, the more likely your products are to surface in agentic recommendations.

Adopt an API-first architecture

Make core systems — catalog, inventory, fulfillment, pricing, and payments — accessible through standardized APIs. This modular approach allows AI agents to retrieve live data, initiate transactions, and confirm availability in real time without human intervention. Headless commerce isn’t just flexibility; it’s now a prerequisite for AI integration.

Integrate early with agentic protocols

Explore pilot integrations with the Agentic Commerce Protocol (ACP) or equivalent frameworks. Early movers will gain distribution advantages as AI platforms refine ranking models and expand merchant access.

Establish clear success metrics: AI-attributed conversion, order accuracy, and customer satisfaction.

Redefine attribution and revenue models

Traditional web analytics can’t track a purchase that happens entirely inside a chat.

Develop new models for measuring ROI, customer acquisition cost (CAC), and lifetime value in AI-led environments. Clarify how commissions or referral fees are shared across agents, merchants, and intermediaries.

Design loyalty beyond the website

When customers buy through AI, they may never visit your site. Build mechanisms that preserve post-sale relationships — email receipts, warranty registration, exclusive offers — so you can continue engagement after the transaction. Loyalty now begins after the checkout, not before it.

Strengthen trust, compliance, and transparency

Users must believe that recommendations are unbiased and payments are secure.

Implement tokenized payment flows, clear data-handling policies, and traceable audit logs. Stay ahead of evolving regulations around consumer protection, AI transparency, and sponsored content disclosure.

Monitor market behavior and consumer sentiment

Track how shoppers respond to AI-mediated purchasing. Some may embrace it for convenience; others may resist it out of mistrust.

Continuous monitoring of customer feedback, conversion funnels, and competitive positioning will help calibrate your investment pace and communication strategy.

Conclusion

Agentic Commerce marks a structural pivot in digital business. It merges product discovery, recommendation, and transaction into a single conversational flow — and shifts control of that journey to AI intermediaries. For executive teams, this demands not only technical readiness, but also a strategic rethinking of distribution, attribution, and customer ownership.

Key Takeaways

• The visibility of your products will soon depend less on search rankings and ad budgets — and more on the quality, accessibility, and structure of your product data.
• Competitive advantage will belong to organizations that expose their catalogs through open APIs and standardized protocols like ACP, allowing AI agents to “see” and transact with them seamlessly.
• Traditional performance metrics will lose relevance. New KPIs must capture AI-attributed conversions, tokenized transactions, and post-sale retention.
• The brand experience will increasingly unfold beyond owned websites. Securing loyalty after the sale will be as important as guiding customers before it.
• Governance, trust, and compliance will become core leadership topics — not back-office concerns — as companies face pressure to ensure that AI-mediated recommendations remain transparent and unbiased.

Agentic Commerce isn’t a passing trend — it’s the next logical step in how digital demand is created and captured.

Leaders who prepare now will shape the standards and economics of this new layer of commerce.

You might also like:

• Agentic Commerce, Part 1: Where Conversations Become Transactions » Learn more
• Agentic Commerce, Part 2: Winners, Losers, and the Forces Behind the Change » Learn more
• Our expertise: Artificial Intelligence » Learn more
• Pharma’s AI Readiness Index: Who Leads the Race? » Learn more
• AI-Driven Personalization in B2B E-Commerce: From Recommendation Logic to Real-Time Experience » Learn more
• More German companies see opportunities in artificial intelligence » Learn more

How to deliver seamless self-service and sustainable growth.

In subscription e-commerce, the real work begins after the customer clicks “Subscribe.” Winning brands don’t rely on a slick front end alone; they build a technology backbone that keeps every subscription accurate, compliant, and instantly responsive.

Modern buyers expect complete self-service: They want to upgrade or pause plans, add seats, review usage, and download invoices without contacting support. Meeting those expectations requires a well-integrated subscription management platform that automates billing, entitlement, communications, and data flow across the business.

Core capabilities every subscription platform needs

1) Recurring payments and automated billing

Reliable recurring billing is the backbone of subscription commerce. Your platform should:

• Comply with PCI DSS (payments) and GDPR (data protection).
• Support Strong Customer Authentication (SCA) where applicable.
• Offer intelligent dunning and retry logic for failed charges (e.g., smart schedules, card updater services, proactive notifications).
• Handle multi-currency and regional payment methods to support international growth.

This reduces involuntary churn, prevents revenue leakage, and builds trust by making payments predictable and transparent.

2) Flexible subscription management

Customers expect real-time control over their plans:

• Upgrades/downgrades with accurate pro-rata adjustments.
• Pause/suspend options without full cancellation.
• Add-ons/bundles that extend functionality on demand.
• Billing frequency changes (monthly, quarterly, annual) with correct recalculation of amounts and dates.

Behind the scenes, robust business logic must handle overlapping periods, proration rules, plan dependencies, and entitlements – so what customers see in the portal is exactly what finance records in the ledger.

3) Self-service account administration

A high-performing self-service subscription portal lowers support load and increases satisfaction. Essentials include:

• Edit personal and billing details in real time.
• Manage multiple subscriptions under one account with consolidated billing.
• Access a complete change history of plan modifications, pauses, renewals, and refunds.
• Download invoices/receipts instantly (PDF, XML).
• Manage roles, permissions, and license allocations in multi-user models.

This level of transparency improves retention because customers feel in control and supported, without needing a ticket.

4) Automated invoicing and documentation

Every billing event should trigger a legally compliant invoice, synchronized with your ERP or finance system. Invoices must be generated automatically, stamped with correct tax rules, and made available immediately in the portal. For B2B, consider supporting purchase orders, credit notes, and collective invoices for complex account structures.

5) Usage-based billing

Where consumption varies – API calls, storage, seats, or utilities – your platform needs real-time metering and rating:

• Accurate, tamper-resistant event collection.
• Threshold alerts and spending caps.
• Auto-pause when limits are reached.
• Transparent dashboards for usage and costs.

Done well, usage-based billing aligns price with value and can increase Net Revenue Retention (NRR) by enabling gentle, data-driven upsell paths.

6) Entitlement and feature management

Entitlement systems must reflect subscription status instantly: upgrade to Premium, features unlock; suspend an account, access is restricted. This requires an event-driven design where plan changes propagate immediately to your product, identity, and licensing services.

Standard modules vs. custom development

Most major platforms — including Shopify, Adobe Commerce (Magento), Salesforce Commerce Cloud, SAP Commerce — offer subscription modules or marketplace apps. They are excellent starting points, but they can struggle with:

• Hybrid models (physical goods + SaaS under one contract).
• Complex billing rules (multi-currency, multi-tax, fixed + usage fees).
• Industry compliance (healthcare data, financial regulations, export controls).

You will face a strategic choice: adapt your model to the platform’s limits or extend/build capabilities. Use a simple decision lens:

• Complexity: Are your billing/entitlement rules exotic enough to require code?
• Time-to-market: Can you launch on standard modules and phase in custom work later?
• Maintainability: Do you have the resources (internal or partner) to own a custom codebase for years?
• Total cost: Compare license + integration spend against the long-run cost of building and maintaining bespoke components.

Compliance and risk management: non-negotiables

Subscription platforms touch sensitive data and money flows. Treat compliance as a product feature:

• PCI DSS for cardholder data; avoid storing raw PAN where possible—use tokenization.
• GDPR/CCPA for personal data, consent management, and subject rights.
• SCA/3-D Secure for certain European payments.
• Tax compliance across jurisdictions (VAT/GST), ideally via automated calculation services.
• Auditability: Immutable logs of billing events, entitlements, and communications.

The payoff: fewer disputes, faster audits, and greater enterprise buyer confidence.

Architecture: the integrations that make self-service real

A portal without deep integrations is a façade. Real self-service relies on consistent, synchronized data across your stack:

Core systems to connect

• ERP/Financials: pricing, tax, revenue recognition, invoicing.
• Billing engine: recurring and usage charges, dunning, credits.
• Payment gateways: tokenization, retries, fraud checks.
• CRM: account hierarchy, engagement, lifecycle communications.
• WMS/Inventory: for physical components of subscriptions.
• Logistics: label generation, carrier updates, returns flows.
• Licensing/Entitlement: instant feature activation/deactivation.
• Support/ITSM: SLAs tied to plan and status.
• Analytics/CDP: events for churn prediction, upsell, and LTV analysis.

Integration patterns that scale

• Prefer event-driven messaging (e.g., “PlanUpgraded,” “InvoiceIssued”) over brittle point-to-point calls.
• Introduce an API layer or gateway for consistent authentication, throttling, and versioning.
• Use idempotency keys for payment and billing actions to avoid duplicates.
• Maintain a single source of truth for subscription state; let downstream systems subscribe to changes.

Example: a mid-cycle upgrade without friction

A customer on a €20 Basic plan upgrades mid-month to €40 Premium:

1. Upgrade request: The portal emits an event; ERP and entitlement subscribe.
2. Proration: ERP calculates the partial charge and posts a pending invoice.
3. Payment: The gateway runs an immediate charge (idempotent, with retry policy).
4. Entitlement: On payment confirmation, Premium features unlock instantly.
5. Communication: CRM updates the account and triggers confirmation + onboarding tips.
6. Analytics: The upgrade event feeds dashboards for NRR, ARPA, and expansion tracking.

Because integrations are event-driven, customers see changes in seconds, not days—no tickets, no manual fixes, no broken promises.

Analytics and KPIs that matter

Track the metrics that reveal system health and commercial impact:

• MRR/ARR and ARPA (Average Revenue per Account).
• Churn (logo and revenue) and Net Revenue Retention (NRR).
• Involuntary churn from failed payments vs. voluntary cancellations.
• Dunning success rate and payment recovery time.
• Upgrade/downgrade velocity and time-to-value after plan changes.
• Support deflection: portal self-service actions vs. tickets raised.

Instrumentation should be native: every billing, entitlement, and account event should be analytics-ready.

Implementation playbook: from pilot to scale

1. Discovery & design

• Map products, plans, add-ons, billing rules, and entitlements.
• Define the single source of truth for subscription state.

2. Integration blueprint

• Identify systems, events, and APIs; choose event bus/middleware.
• Specify idempotency, retries, and dead-letter handling.

3. Security & compliance

• Scope PCI responsibilities, data retention, and access controls.
• Document consent flows and audit trails.

4. MVP scope

• Launch core: checkout, billing, invoices, basic self-service.
• Reserve complex usage rating or advanced entitlements for Phase 2 if needed.

5. Change management

• Update terms, billing policies, and knowledge base articles.
• Train support and finance on new flows.

6. Observability

• Implement logging, tracing, alerts, and dashboards before go-live.
• Test failure paths: payment declines, partial refunds, plan conflicts.

7. Iterate

• Use customer behavior data to refine pricing, bundles, and limits.
• Automate the common manual exceptions you still see post-launch.

Future trends shaping subscription platforms

• AI-assisted dunning and collections: Predict optimal retry windows, channels, and messaging to lift recovery rates without harming UX.
• Personalized plans and dynamic bundles: Use behavioral data to auto-propose the right add-ons at the right time.
• Real-time entitlement at the edge: Low-latency feature checks for global SaaS footprints.
• Usage transparency by design: In-product meters and forecasts reduce bill shock and increase trust.
• Composability: More vendors are adopting modular, API-first architectures, letting you swap billing, payments, or entitlement components without rebuilding the whole stack.

Buy vs. build: a simple decision framework

Choose standard modules when:

• Your model is close to common patterns.
• You need to launch quickly with proven building blocks.
• Your team prefers configuration over code.

Invest in custom development when:

• You require hybrid physical/digital subscriptions or intricate usage pricing.
• Compliance or data residency demands tight control.
• You want proprietary differentiation in billing or entitlement logic.

Most mature organizations land on a hybrid approach: a commercial billing core plus targeted custom services for the rules that truly differentiate the business.

Conclusion: build the backbone that keeps every promise

Subscription commerce isn’t about selling a plan once – it’s about delivering value cycle after cycle. The brands that win pair an elegant self-service experience with a disciplined, integrated back end:

• Robust recurring billing with compliance built in.
• Flexible plan management and transparent account administration.
• Instant entitlements and accurate usage metering.
• Deep integrations across ERP, CRM, payments, support, and analytics.
• Scalable, composable architecture ready for AI and new revenue models.

Get those foundations right and your portal stops being a façade. It becomes the operational heart of your business – fulfilling every change in real time, reducing cost-to-serve, and compounding customer lifetime value.

You might also like:

• E-Commerce Solution Selection Guide, Part 1: Technology Overview and Evaluation Process » Learn more
• E-Commerce Solution Selection Guide, Part 2: Evaluation Criteria » Learn more
• How to choose the right IT service provider » Learn more
• How buzzwords could put your IT at risk » Learn more
• Mastering software delivery: Overcoming development hurdles » Learn more

Some waves in technology build slowly. Others overturn the entire shoreline. Agentic Commerce belongs to the latter.

As AI agents step into the role once held by search engines, marketplaces, and retailers, the commercial landscape is turning upside down before our eyes. These agents now determine which products surface, which brands gain visibility, and which merchants ultimately win the transaction.

For some, this shift will unlock vast new channels of growth. For others, it could quietly dismantle the foundations of their business models.

Yet none of this should come as a surprise. For companies watching closely, the shift was visible long before it arrived. The clues were everywhere:

• in how consumers stopped searching, and started asking
• in how checkout flows shrank to a single click
• in how AI became part of daily life

So, Agentic Commerce didn’t appear overnight. It is the natural culmination of technological, behavioral, and economic forces that have been aligning for some time. Together, they explain why this transformation is happening now — and why its impact will be impossible to contain.

What’s Driving the Change

AI becomes the first point of contact

Increasingly, users turn to AI assistants for recommendations instead of starting with search engines or marketplaces. Asking “What’s the best cleaner for wine stains?” or “Which lamp fits a Scandinavian interior?” has become faster and more intuitive than browsing endless product grids.

The rise of zero-click commerce

In digital journeys, every click creates friction — and lost conversions. Agentic Commerce eliminates that friction entirely: discovery, comparison, and purchase now happen in a single interface. Analysts already describe this shift as zero-click search, where the first answer becomes the final transaction.

New monetization models for AI

After the initial wave of subscription revenues, AI providers are seeking sustainable business models. Transaction fees and merchant commissions offer a new source of income — one that aligns platform incentives with measurable commercial outcomes.

Standardization through open protocols

The emergence of frameworks like the Agentic Commerce Protocol (ACP) allows AI agents and merchants to communicate through a shared data language. This interoperability makes large-scale adoption possible without bespoke integrations for each platform.

Maturity of language model

The newest generation of AI models can interpret nuanced user intent, evaluate multi-dimensional product attributes, and deliver context-aware recommendations. Only now are they capable of guiding users through complex buying decisions with human-like precision.

Rising consumer expectations

Shoppers are growing accustomed to instant answers and one-step actions. Once people realize they can “ask and buy” within a single conversation, that level of convenience will become the new baseline — pushing brands and retailers to meet them there.

As all these forces gain strength, the balance of opportunity across the commerce ecosystem is shifting, creating a new competitive map. The winners will be those who can adapt their data, architecture, and business models to an AI-first environment. Those who can’t — or won’t — risk being left behind.

Who Stands to Win

Digitally mature merchants

Retailers and brands with modern, API-first systems, rich product metadata, and well-structured catalogs will be the first to benefit. Their data can easily feed into AI agents, making their products visible and purchasable in emerging conversational channels.

Niche and long-tail sellers

Smaller merchants with distinctive products stand to reach audiences they could never access through search or advertising alone. Within AI-driven recommendations, quality and relevance often outweigh brand size or marketing budgets.

FinTechs and infrastructure providers

Payment platforms such as Stripe, which co-developed the Agentic Commerce Protocol (ACP), are establishing themselves as the backbone of this new commerce layer. Those who build the rails for tokenized, agent-initiated transactions will own a critical share of the value chain.

AI and platform operators

Conversational interfaces that capture intent and facilitate checkout become the new gatekeepers of demand. They will control which products appear, how they’re ranked, and what commissions or fees are applied to each transaction.

Consumers

Shoppers will gain from a more seamless, low-friction journey — from asking a question to completing a purchase — without navigating multiple websites or interfaces.

Who May Lose Ground

Traditional marketplaces and intermediaries

As purchases begin inside chat interfaces, the role of marketplaces like Amazon or Allegro as discovery gatekeepers may weaken. Traffic and conversion could shift upstream to the AI layer, where agents recommend products directly.

Price aggregators and comparison engines

In a world where AI already compares products and pricing internally, external aggregators may struggle to justify their role. The assistant becomes the comparison layer itself — personalized, dynamic, and invisible.

Strong brands controlling their own storefronts

Agentic Commerce challenges the notion of brand-owned experiences. When a transaction happens through an AI interface, the brand loses control over layout, storytelling, and visual identity — its exposure depends entirely on algorithmic relevance.

Data-poor merchants

Companies with incomplete, inconsistent, or unstructured product data will find themselves invisible to AI agents. Without rich metadata and accessible APIs, even great products may never surface in a conversational recommendation.

Performance-driven marketers

If users bypass search engines and websites altogether, traffic-based performance marketing models will erode. SEO, paid clicks, and retargeting become less effective; investment will need to shift toward data quality, structured feeds, and AI channel optimization.

Conclusion: The Rules of Commerce are Being Rewritten

The rise of agent-led shopping is quietly rewriting the rules of visibility, value, and competition in digital commerce.

In this new landscape, it’s no longer the marketplace owners — or the size of your advertising budget — that determine who gets seen. AI algorithms now decide what surfaces, when, and to whom. Product data becomes a company’s real marketing currency, and adaptability its greatest advantage.

The early adopters aren’t just selling through AI; they’re teaching it to recognize, rank, and recommend them. Everyone else is fighting to stay visible in a marketplace they no longer control.

The message is clear: in the age of agent-led shopping, success won’t depend on who shouts the loudest, but on who speaks the language of AI fluently enough to be heard.

You might also like:

• Agentic Commerce, Part 1: Where Conversations Become Transactions » Learn more
• Agentic Commerce, Part 3: Strategic Priorities for Executive Leaders » Learn more
• Our expertise: Artificial Intelligence » Learn more
• Pharma’s AI Readiness Index: Who Leads the Race? » Learn more
• AI-Driven Personalization in B2B E-Commerce: From Recommendation Logic to Real-Time Experience » Learn more
• More German companies see opportunities in artificial intelligence » Learn more

The integration method you choose today defines the speed, resilience, and cost-efficiency of your business tomorrow. This article helps you navigate the trade-offs.

Following on from Part 2, which examined the strategic patterns of system integration, this article shows how these concepts can be implemented in practice using proven technologies and methods. The aim is to ensure that integration decisions are technically sound and future-proof.

1. File-based integration

In file-based integration, system A generates a file (e.g., CSV, JSON, XML) and stores it in a designated location (local storage, FTP server, or cloud bucket). System B periodically checks for new files, retrieves them, and processes the data. Scheduled jobs (e.g., cron jobs, Windows Task Scheduler) are typically used to automate this exchange.

Associated patterns:

• Batch Data Integration
• Pull Integration
• Point-to-Point Integration

Advantages: Simple and cost-effective to implement. No need for real-time connectivity between systems.

Disadvantages: No real-time synchronization, manual error handling, risk of version conflicts and data inconsistencies, requires additional validation and processing before data import.

When to use: For periodic data transfers where real-time processing is not critical (e.g., nightly batch imports).

Common Technologies: SFTP, AWS S3, Windows Task Scheduler, CSV, XML, JSON file formats.

2. Database-level integration

Database-level integration involves two systems accessing a shared database or synchronized databases. Data exchange occurs through SQL queries, database replication, stored procedures, or triggers. Alternatively, systems may use database views or ETL processes to read and write data independently.

Associated patterns:

• Synchronous Integration (for direct database reads/writes)
• Batch Data Integration (in replication scenarios)
• Tightly Coupled Integration

Advantages: Direct access, high performance, ability to leverage native database capabilities like triggers and replication.

Disadvantages: Strong dependency on database schema and structure, scalability and performance issues in large deployments, increased risk of breaking data integrity and transaction consistency, difficult to manage schema changes across dependent systems.

When to use: When systems need fast access to shared data and are built around the same database environment.

Common Technologies: SQL (PostgreSQL, MySQL, Oracle Database), NoSQL (MongoDB replication, Cassandra), data replication tools (e.g., Oracle GoldenGate, AWS Database Migration Service).

3. API-based integration

API-based integration connects systems via standardized APIs (typically RESTful or SOAP APIs). System A sends an HTTP request (GET, POST, PUT, DELETE) to System B’s API, which processes the request and returns a response, often in JSON or XML format. Authentication and authorization are enforced through methods like OAuth 2.0 or JWT.

Associated Patterns:

• Synchronous or Asynchronous Integration
• Real-Time Integration
• Loosely Coupled Integration
• Push Integration

Advantages: Real-time communication, high flexibility, cloud-friendly, strong control over access, security, and service versioning.

Disadvantages: Requires careful API lifecycle management and documentation, API outages or breaking changes can disrupt dependent systems, security vulnerabilities if APIs are not properly protected.

When to use: For integrating dynamic applications, cloud services, mobile apps, and microservices architectures.

Common Technologies: REST APIs, GraphQL APIs, gRPC, OAuth 2.0, OpenID Connect, JWT, Tools like Postman, Apigee API Management, AWS API Gateway.

4. Enterprise Service Bus (ESB)

An Enterprise Service Bus acts as a centralized communication backbone between multiple systems. Each system connects to the ESB, which handles routing, transformation, orchestration, and monitoring of data flows between participants. Advanced ESBs can also apply business rules or conditional routing based on data content or context.

Associated Patterns:

• Centralized Integration
• Loosely Coupled Integration
• Asynchronous Integration (often with support for messaging systems)

Advantages: Simplifies integration architecture by centralizing management, supports complex orchestration, data transformation, and error handling, high scalability for large enterprise environments.

Disadvantages: Higher cost and complexity compared to direct integrations, if not properly designed, the ESB can become a bottleneck or single point of failure, requires robust governance and monitoring to avoid “monolithic ESB syndrome.”

When to use: In large organizations integrating dozens or hundreds of heterogeneous systems.

Common Technologies: MuleSoft Anypoint Platform, WSO2 Enterprise Integrator, Red Hat Fuse, IBM App Connect.

5. Event-driven integration

In event-driven integration, systems communicate by producing and consuming events rather than making direct calls. System A publishes an event (e.g., “order created”) to an event broker, and subscribing systems react asynchronously. This approach decouples senders and receivers, improving scalability and resilience.

Associated Patterns:

• Asynchronous Integration
• Real-Time Integration
• Loosely Coupled Integration
• Push Integration

Advantages: Near real-time processing with low latency, highly scalable and resilient against system failures, reduces system coupling, promoting flexibility and adaptability.

Disadvantages: More complex event modeling and data flow management, requires robust event brokers and monitoring mechanisms, increased difficulty in tracing and debugging event chains across distributed systems.

When to use: In IoT systems, e-commerce platforms, financial transactions, and real-time analytics environments.

Common Technologies: Apache Kafka, RabbitMQ, AWS EventBridge, Azure Event Grid.

6. iPaaS – Integration Platform as a Service

iPaaS solutions provide cloud-based platforms for building and managing integrations between cloud and on-premise systems through pre-built connectors and low-code development environments. Users can orchestrate workflows visually, map data transformations, and automate processes without heavy custom coding.

Associated Patterns:

• Centralized Integration
• Loosely Coupled Integration
• Real-Time and Batch Data Integration
• Push and Pull Integration

Advantages: Fast time-to-market with prebuilt connectors and templates, supports multi-cloud and hybrid integration scenarios, reduces the need for deep technical expertise in early integration stages.

Disadvantages: Subscription costs may increase significantly with scale, may have limited flexibility for highly customized or performance-critical integrations, risk of vendor lock-in if proprietary components are heavily used.

When to use: When organizations need rapid, scalable, and flexible integration capabilities without building everything from scratch.

Common Technologies: SAP Cloud Platform Integration (SAP CPI), Informatica Cloud Application Integration (CAI), Dell Boomi, Workato, MuleSoft Composer, Microsoft Power Automate.

7. Hybrid Integration Architectures

A single integration approach rarely prevails in practice. Successful companies combine different methods depending on the application.

An overarching, event-driven approach often forms the backbone of the architecture. API calls, file exchange, or database access are used in a targeted manner within individual domains.

Example:

When an order is placed:

1. The Order domain publishes an OrderPlaced event on the Event Bus.
2. The Email domain listens for that event and triggers logic to build a confirmation email for the customer.
3. The Email domain uses an API-based integration to retrieve email content from the CMS system.

This approach illustrates how event-driven architecture acts as a backbone, while point-to-point API communication is used tactically where direct interaction is appropriate.

Hybrid integration allows organizations to stay modular, avoid unnecessary coupling, and leverage the strengths of multiple methods — provided the integration strategy is governed consistently.

Conclusion: From vision to implementation

Technical implementation plays a decisive role in determining the future viability of an integration strategy. Those who consciously deploy technologies in line with strategic patterns create robust and scalable IT ecosystems. Hybrid architectures provide the necessary flexibility to optimally support departments, business models, and innovation cycles.

You might also like:

• IT System Integration, Part 1: Benefits and Challenges of IT System Integration » Learn more
• IT System Integration, Part 2: Strategic Patterns of IT System Integration » Learn more
• IT Report: Digitization projects — 6 key findings on challenges and experiences with recommendations » Learn more
• Cost savings in software development » Learn more
• Mastering software delivery: Overcoming development hurdles » Learn more
• Low-Code Falls Short Without Full-Code » Learn more

We are witnessing another power shift in digital commerce. The next checkout won’t be a webpage — it will be the chat window where you ask for advice.

A new e-commerce model is emerging — often called Agentic Commerce — where purchases happen directly inside conversations with AI assistants. Some describe it as the next evolution of conversational commerce (Conversational Commerce 2.0) but this time, the agent doesn’t just recommend products. It buys them for you.

How It Works: From Conversation to Checkout

The mechanism behind agent-led shopping generally follows a clear sequence:

Step 1. Intent capture

The process begins when the user formulates a request — for instance, asking for a “minimalist desk lamp under 80 €.” The AI interprets the query, analyzes preferences and constraints, and retrieves relevant product data from integrated merchant feeds.

Step 2. Intelligent product selection

Using structured metadata such as price, material, color, availability, and customer reviews, the AI agent compiles a shortlist of matching products. Recommendations are typically ranked by contextual relevance and user fit, not by sponsored placement.

Step 3. In-chat checkout

Once the user chooses a product, the assistant enables checkout directly in the conversation window. Payment, shipping, and confirmation happen within the same interface — no external redirects, no multiple tabs. This step removes the friction of switching platforms and shortens the buying journey to a single interface.

Step 4. Secure payment authorization

Instead of exposing sensitive payment credentials, the AI agent relies on a tokenized or delegated authorization system provided by the payment processor.

Whether through Stripe’s ACP, Google’s AP2, or similar emerging standards, the core principle remains the same: transactions are verified without the AI ever storing card data. These protocols — and their growing role in building the agentic commerce infrastructure — are explored later in this article.

Step 5. Merchant-side fulfillment

The order and payment confirmation are transmitted via standardized APIs to the merchant’s back-end systems. The merchant validates the transaction, handles taxation, fulfillment, and customer communication, while the AI acts as an intermediary between user and seller.

Step 6. Continuous optimization

Each interaction generates insight into user intent, conversion behavior, and product relevance. Over time, this feedback loop enables both merchants and AI providers to refine recommendations and improve the efficiency of agent-driven transactions.

Reality Check: Agentic Commerce Goes Live

Agentic Commerce is no longer a thought experiment or a slide in a keynote deck.

Over the past months, several leading technology firms have begun rolling out real, functioning implementations of AI-led shopping — systems that turn conversation into transaction. What was once a speculative idea is now appearing in production: users can browse, buy, and pay directly inside chat interfaces.

The following examples show how quickly this new model is moving from theory to market reality — and how differently each player is trying to capture its share of the coming AI commerce layer.

OpenAI and Stripe

OpenAI has taken the boldest step so far. Its Instant Checkout feature in ChatGPT allows users in the United States to browse and buy products from Etsy merchants — with Shopify integrations expected to follow soon. For now, the rollout covers U.S.-based sellers only, but it marks the first large-scale implementation of true Agentic Commerce.

The foundation of this system is the Agentic Commerce Protocol (ACP), an open standard co-developed by OpenAI and Stripe. ACP defines how AI agents and merchants communicate — how they request product data, transmit orders, and reconcile payments. In practice, it replaces dozens of one-off integrations with a single, standardized API language between agents and commerce platforms.

A key technical innovation is Stripe’s Shared Payment Token (SPT) — a secure, one-time payment credential linked to a specific merchant and basket total. It enables the AI agent to trigger a payment without ever storing or exposing card details.

SPTs currently support all major card networks (Visa, Mastercard, Amex), debit cards, and digital wallets such as Apple Pay, Google Pay, and Link by Stripe.

For Stripe, this step extends its role beyond payments into the economic infrastructure of AI-driven commerce. By defining how transactions are initiated by autonomous agents, Stripe positions itself as the backbone of the agentic layer.

If ACP gains traction across merchants and other AI platforms, OpenAI and Stripe could effectively control the “grammar” of online transactions — a level of influence once held by search engines and marketplaces.

Google and the Agent Payments Protocol (AP2)

Google has taken a parallel path with its own open framework — the Agent Payments Protocol (AP2) — announced in mid-2025.

AP2 aims to standardize how AI agents initiate and authorize payments across Google’s ecosystem, including Gemini, Android, and the Chrome browser.

The protocol allows agents to act on behalf of users to complete purchases, using delegated payment credentials stored in Google Pay.

AP2 mirrors many of the ideas introduced by OpenAI and Stripe’s ACP — tokenized transactions, standardized merchant APIs, and transparent data exchange — but operates within Google’s existing global infrastructure and merchant base.

Initial implementations are being tested in North America and Western Europe, focusing on digital goods and small consumer purchases from verified Google Pay merchants.

Industry observers expect AP2 to integrate tightly with Google Shopping, Google Maps, and YouTube, effectively merging product discovery and transaction into a single AI-mediated experience.

If widely adopted, AP2 could make Google’s AI agents the primary interface for both online and offline shopping — unifying discovery, payment, and loyalty under one ecosystem.

The competition between ACP and AP2 is less about technology and more about control: whichever protocol becomes the industry default will define how the next generation of commerce operates.

PayPal Honey and OpenAI

While OpenAI builds the foundation, PayPal is fighting to remain visible in the new discovery layer. Its Honey extension, once known for surfacing coupons and cashback offers in browsers, is now integrating directly with ChatGPT and other AI platforms.

When users ask AI for product advice, Honey injects real-time pricing, merchant options, and available deals — even flagging retailers that the AI might have missed. This strategy aims to reinsert PayPal into the path-to-purchase at a time when AI platforms threaten to bypass traditional affiliate models altogether.

By merging its data on merchants, discounts, and transactions with conversational agents, PayPal is repositioning itself not just as a payment processor, but as a pricing intelligence layer inside AI commerce. The challenge, however, will be maintaining influence in systems increasingly controlled by the AI platforms themselves.

Microsoft and Perplexity

 Microsoft, through its Copilot Merchant Program, is enabling brands to set up in-chat storefronts — miniature shops living inside AI interfaces. Meanwhile, Perplexity has already rolled out similar in-chat shopping and payments, mirroring the OpenAI approach.

Both moves highlight the strategic race for the AI “front door.” Whoever controls that interface — the place where a purchase begins — controls discovery, data, and revenue. For merchants, this introduces both opportunity and dependency: access to massive traffic via AI assistants, but under rules set by the platform.

The competition among OpenAI, Microsoft, and Perplexity is not just about features; it’s about defining the default path to purchase in an AI-mediated economy. As more agents adopt standardized protocols like ACP, the barriers to switching between them may shrink — yet the platform with the strongest user trust and ecosystem will likely dominate.

Conclusion

Digital commerce is undergoing a fundamental shift: In the future, checkout will no longer happen on a website — it will happen directly in the chat. AI assistants will handle product discovery, selection, and payment without the user ever leaving the conversation.

Agentic Commerce represents the next evolution of e-commerce: a system in which standardized protocols like Stripe ACP and Google AP2 govern transactions — and in doing so, redefine the balance of power in digital commerce.

👉 Key Takeaway:

Agentic Commerce shifts control over product discovery, purchasing, and payment from shops and marketplaces to AI assistants. Whoever owns the interface between user and transaction today will shape the grammar — and the rules — of online commerce tomorrow.

You might also like:

• Agentic Commerce, Part 2: Winners, Losers, and the Forces Behind the Change » Learn more
• Agentic Commerce, Part 3: Strategic Priorities for Executive Leaders » Learn more
• Our expertise: Artificial Intelligence » Learn more
• Pharma’s AI Readiness Index: Who Leads the Race? » Learn more
• AI-Driven Personalization in B2B E-Commerce: From Recommendation Logic to Real-Time Experience » Learn more
• More German companies see opportunities in artificial intelligence » Learn more

System integration today is a maze of patterns, methods, and technologies — only with the right guidance can you design an architecture that truly scales and adapts.

As system integration becomes a strategic pillar of modern IT ecosystems, the question is no longer whether to integrate — but how to do it effectively.

This article highlights the most important integration patterns — their areas of application, advantages, and challenges.

1. Synchronous vs. Asynchronous Integration

Synchronous integration involves immediate communication between systems, where the sending system waits for a direct response before proceeding. It’s ideal for processes that require instant confirmation, like payment authorization or login validation.

Asynchronous integration, on the other hand, allows systems to exchange information without waiting — the sender transmits a message and continues operating while the receiver processes it independently. This model supports higher scalability and system decoupling, but adds complexity in handling delays or failed messages.

When to use:

• Synchronous: For critical operations requiring real-time feedback and low latency.
• Asynchronous: For long-running processes, high-volume data exchanges, or when systems operate at different speeds.

Potential challenges:

• Synchronous: Risk of bottlenecks under high loads.
• Asynchronous: Increased effort for error handling and message queues.

2. Real-Time vs. Batch Data Integration

Real-time integration ensures that data is exchanged and processed instantly as events occur. It supports dynamic operations like real-time inventory updates or fraud detection systems.

Batch integration involves collecting data over a period and processing it all at once. This approach suits non-time-sensitive tasks like nightly financial reconciliation or large-scale data migrations.

When to use:

• Real time: When immediate data visibility is critical for operations or customer experience.
• Batch: When timing flexibility is acceptable and operational efficiency is prioritized over instant updates.

Potential challenges:

• Real time: High infrastructure requirements.
• Batch: Potential data latency and synchronization issues.

3. Point-to-Point vs. Centralized Integration

Point-to-point integration creates direct, individual connections between each pair of systems. It is simple for a small number of systems but becomes exponentially complex as more systems are added, often leading to a “spaghetti architecture.”

Centralized integration uses a hub, bus, or integration platform to manage all connections through a single point, dramatically reducing complexity and easing maintenance.

When to use:

• Point-to-point: Suitable for small environments with a limited number of integrations.
• Centralized: Essential for scaling, flexibility, and managing complex, dynamic IT landscapes.

Potential challenges:

• Point-to-point: Not very scalable, high maintenance costs.
• Centralized: Risk of a single point of failure if the system is not adequately designed.

4. Out-of-the-Box vs. Custom Integration

Out-of-the-box (OOTB) integration leverages pre-built connectors or integrations offered by system vendors. It’s faster to implement, requires less technical expertise, and is ideal for common use cases (e.g., integrating Salesforce with Outlook).

Custom integration involves building tailored solutions to meet unique business needs, offering maximum flexibility but at a higher cost and complexity.

When to use:

• Out-of-the-box: For standard, widely supported integrations where speed and cost-efficiency are priorities.
• Custom: When business processes require specific, non-standardized interactions that OOTB solutions cannot address.

Potential challenges:

• Out-of-the-box: Low flexibility, potential vendor lock-in.
• Custom: High maintenance costs, need for experts.

5. Tightly Coupled vs. Loosely Coupled Integration

Tightly coupled integration means systems are directly dependent on each other’s data models, interfaces, and availability. Changes in one system often require changes in the others.

Loosely coupled integration decouples systems, using standard protocols, event-driven architectures, and independent services that interact without hard dependencies.

When to use:

• Tightly coupled: For simple, stable environments where control and performance are more important than flexibility.
• Loosely coupled: For scalable, resilient, and adaptable architectures that can evolve over time without breaking dependent systems.

Potential challenges:

• Strongly coupled: Low adaptability to change.
• Loosely coupled: Requires precise planning and modern integration patterns.

6. Pull vs. Push Integration

Pull integration means the receiving system periodically requests (polls) data from the sending system, which can lead to delays and increased load during polling intervals.

Push integration has the sending system proactively deliver data to the receiving system whenever an event occurs, supporting more real-time or near-real-time communication.

When to use:

• Pull: When receiving systems should control the frequency and timing of data synchronization.
• Push: When timely information delivery is critical and sender-driven communication is preferred.

Potential challenges:

• Pull: Load during high query frequency.
• Push: Need for robust transmission mechanisms and error handling.

Conclusion: Choose the right blueprint

Integration patterns are the foundation of any IT architecture. Companies should choose them carefully, based on their processes, goals, and scaling strategy.

In part 3 of this series, we will examine the technological methods that can be used to successfully implement these patterns in practice.

→ Continue to part 3: Technologies and Methods for the Practical Application of IT System Integration

You might also like:

• IT System Integration, Part 1: Benefits and Challenges of IT System Integration » Learn more
• IT System Integration, Part 3: Technologies and Methods for the Practical Application of IT System Integration » Learn more
• IT Report: Digitization projects — 6 key findings on challenges and experiences with recommendations » Learn more
• Cost savings in software development » Learn more
• Mastering software delivery: Overcoming development hurdles » Learn more
• Low-Code Falls Short Without Full-Code » Learn more

In today’s agile, high-speed development cycles, roles are evolving — and sometimes overlapping. But what if they didn’t just overlap? What if one of them could disappear entirely?

At Striped Giraffe, based on years of hands-on project experience across multiple industries, we’ve seen growing evidence that a competent, empowered QA Manager can fulfill — and sometimes outperform — the traditional role of a Release Manager. This observation raises a thorny question: Is the Release Manager still a necessary role in modern software delivery?

To explore this question, we sat down with two of our experts: Aleksandra Neukirch, QA Lead, and Magdalena Osmańska, Senior Software Tester. Here’s what we’ve learned from that conversation — and why we believe this is a debate worth having.

An expert view by

When QA manages the release – and makes everything much easier

At first glance, the merging of QA and Release Manager responsibilities might sound radical. But in practice, especially in agile or CI/CD-driven environments, it often happens naturally — and works surprisingly well.

“It’s not just about speeding things up,” says Aleksandra Neukirch. “When QA leads the release, you eliminate friction. The person who understands the risk is also the one deciding when to go live.”

Among the key advantages:

• Faster Release Cycles: QA can make informed go/no-go calls immediately, without waiting for external validation.
• Cost Efficiency: One person instead of two — particularly valuable for startups or lean teams.
• Stronger Focus on Automation: The merged role tends to drive better adoption of CI/CD pipelines and post-deploy smoke tests.
• Improved Process Feedback Loops: Bugs or blockers encountered during testing feed directly into release planning and vice versa.

“QA has the full picture,” notes Magdalena Osmańska. “We see the trends, the flaky tests, the areas of recurring failure — and we’re in the best position to use that insight to make a safe release decision.”

But wait — isn’t this a conflict of interest?

Of course, not everyone agrees. One of the most common criticisms of this model is the inherent conflict of interest.

QA’s role is to protect quality. The Release Manager’s job is to ship. When both responsibilities fall to one person, can we really expect objective decisions?

“The risk is real,” admits Aleksandra. “But it’s manageable — if you rely on measurable, automated release criteria.”

She recommends practices like:

• Requiring 95%+ pass rates on regression tests
• Blocking releases with any critical bugs
• Enforcing code coverage thresholds or performance benchmarks

Rather than rely on gut feel or deadline pressure, this approach builds objectivity into the process itself.

A question of skillsets — or of culture?

Another counterargument is that QA and Release Managers have fundamentally different skill profiles.

Release Managers excel in stakeholder communication, deployment orchestration, and risk management. QA specialists are deep in test strategies, automation frameworks, and UX validation.

So does merging the roles mean compromising both?

“Only if you expect one person to do everything,” says Magdalena. “The trick is to define clear responsibilities — and encourage collaboration. QA doesn’t need to be the DevOps expert, but we can take ownership of the release with support from the team.”

In other words: this isn’t a call for heroic multitasking. It’s a call for leaner, smarter processes, where release readiness is a shared responsibility — and quality is everyone’s concern.

When this model works — and when it doesn’t

Like most ideas in software development, this model isn’t universal. But it’s powerful in the right conditions:

• Complex or frequent releases (SaaS, agile, CI/CD-heavy environments)
• Mature automation and observability pipelines
• Teams with a strong quality culture
• Cross-functional collaboration and trust
• A QA lead with solid technical and communication skills

Where these conditions are met, the results can be transformative.

“It’s about empowering QA,” says Aleksandra. “When you trust us with release decisions, you get faster cycles, fewer bugs in production, and a tighter feedback loop. Why add another layer of management if you don’t have to?”

Final thoughts: Is it time to rethink the Release Manager?

We’re not claiming that Release Managers are obsolete. But in many modern teams, they may be optional rather than essential — especially when QA leaders are equipped with the right tools, authority, and skills.

This is not just a debate about roles. It’s about how we structure our teams to move fast, build quality software, and stay lean.

So what do you think?

• Have you worked in environments where QA led the release process?
• Do you see risks in merging these responsibilities — or untapped potential?

Let’s talk.

You might also like:

• How buzzwords could put your IT at risk » Learn more
• Mastering software delivery: Overcoming development hurdles » Learn more
• How to choose the right IT service provider » Learn more

Behind every seamless online payment stands a complex integration of systems, data, and processes — here’s how we built it for a global B2B leader using SAP Digital Payments.

In the first part of our series, we explained what SAP Digital Payments (SAP DP) is and highlighted its business value for companies managing payments across multiple markets.

Now, in the second part, we open the “technical kitchen.” We will show how to actually integrate SAP DP with an e-commerce system and design workflows that ensure security, flexibility, and compliance — all without disrupting the customer journey.

To make this tangible, we take you behind the scenes of a real project we delivered for one of our clients: a global B2B giant running e-commerce operations in numerous local markets. The company needed a solution that could flexibly adapt payment methods and providers to each country’s specific business and regulatory requirements.

Why is this topic so important? Because payments in e-commerce are among the most complex processes to handle. They involve user experience, order management in the backend, ERP accounting, reconciliation, regulatory compliance, and customer service. A single weak link can lead to serious operational and financial risks. And in B2B — where order values are often much higher than in B2C — the stakes are even greater.

Through the intricacies of SAP Digital Payments, we will be guided by our expert — Damian Woźniak, Technical Architect at Striped Giraffe.

Two Integration Models

As we mentioned in the first part of this series, there are two integration models that SAP offers to connect an e-commerce platform with SAP Digital Payments:

1. Payment Page
2. Direct PSP Connection

Both models have different implications for architecture, user experience, and operational processes. Our customer decided to implement and test both approaches to evaluate their strengths and weaknesses.

Model 1: Payment Page

The Payment Page model is PSP-agnostic. This means that the e-commerce application does not need to know which specific payment service provider (PSP) is being used. The entire customer payment process is handled through SAP DP interfaces.

Once a payment transaction is completed, all required payment data is retrieved from SAP DP and sent along with the order data to the SAP S/4HANA system. SAP S/4HANA can then execute further actions for the transaction, such as refunds.

Figure 1. SAP DP Payment Page Integration Model

SAP provides a JavaScript library called DPJSLIB. It is responsible for rendering the payment page in a dedicated window. The content inside this window is served directly by the integrated PSP. This means that every new PSP integration requires coordination with the PSP team to build such a page according to business and UX/UI requirements.

DPJSLIB also enables the application to load all available payment methods configured in the SAP DP administration panel. Based on these options, the user can select the payment method they prefer.

In addition, SAP provides REST endpoints that allow the integrated system to initiate a payment transaction and check its status.

Model 2: Direct PSP Connection

In the Direct PSP Connection model, the e-commerce application integrates directly with a specific PSP, without SAP DP in the middle of the transaction process. SAP DP operates only on the backend layer, supporting financial operations after the order has been placed.

SAP DP is informed about transactions executed directly with the PSP via dedicated endpoints exposed by SAP. These are used to pair the transaction with SAP DP through an adapter.

In this model, the entire payment process is controlled by the PSP. As a result, the UI and overall user experience depend on the PSP chosen.

Figure 2. SAP DP Direct PSP Integration Model

Integrating SAP DP Payment Page with the E-commerce System

Since our client already operated on SAP Commerce, this project was based on that platform. Still, it’s important to note that the same approach could be successfully implemented with other e-commerce systems such as Adobe Commerce/Magento, Shopify or Salesforce Commerce Cloud.

Step 1: Loading Payment Methods

We injected the DPJSLIB script into the frontend application. This is a mandatory step to load digital payment methods and render the payment page served by the PSP.

Step 2: Communication with SAP DP

To communicate with SAP DP endpoints, we developed a REST service (Microservice #1). This centralized unit manages all communication and is especially useful when planning to integrate multiple applications across the architecture landscape with SAP DP.

The service connects to AWS Secret Manager, where SAP client credentials are securely stored. Importantly, the commerce system itself is not connected to the PSP directly. This separation makes it easier to maintain and switch PSPs based on business needs.

Figure 3. Payment Page Solution Architecture

Step 3: Order Placement in ERP

Microservice #1 and SAP Commerce work together to check payment transaction status. In case of success, they trigger order placement.

To place an order in SAP ERP, we leveraged an existing middleware REST service (Microservice #2), which we customized to meet the project’s requirements. This service is responsible for various order-related actions and can be called by SAP Commerce when needed.

Step 4: Communication Flow

The frontend communicates with Microservice #1 and SAP Commerce via GraphQL. This wrapper simplifies data aggregation and API call execution.

Figure 4. Payment Page Sequence Flow

(Hover over the diagram to zoom into specific areas. A high-resolution version is also available for download in PDF and PNG formats.)

How the Payment Page Flow Works

1. Checkout page opens

The customer navigates to the checkout page, which requires payment to complete the order. At this stage, the frontend simply renders the already available content.

2. Initiating the payment transaction

Using GraphQL and Microservice #1, the frontend calls the relevant SAP DP endpoints to initiate the payment. In return, it receives two key elements: the transaction data and the DPJSLIB URL.

3. Injecting DPJSLIB

The frontend injects the DPJSLIB script into its codebase.

4. Loading available payment methods

Dedicated functions from DPJSLIB are then used to load all the available payment methods configured in SAP DP.

5. Presenting payment options to the customer

The payment methods are displayed directly in the customer’s browser.

6. Payment selection

The customer chooses their preferred payment method.

7. Order confirmation

Once the customer confirms the order, the frontend executes a dedicated DPJSLIB function that opens a payment window served by the PSP.

8. Draft order

At the same time, the system proactively submits a draft order to Microservice #2 via GraphQL and SAP Commerce. This order is not yet finalized — it is waiting for payment confirmation. Meanwhile, SAP Commerce triggers the payment finalization thread in Microservice #1.

9. User completes actions in PSP window

After the customer finishes all actions in the PSP payment window, a JavaScript promise is sent back to the frontend.

⚠️ Important: This promise only confirms that the user’s interaction has ended — it does not guarantee a successful payment. Final confirmation still has to come from SAP DP via dedicated endpoints. If the transaction fails early, the JS promise will return an error status, enabling the system to notify the customer immediately.

10. Backend finalization thread

At this point, the payment finalization thread in Microservice #1 checks the transaction status directly through SAP DP.

• If successful → the final order is placed in SAP ERP (via SAP Commerce and Microservice #2).
• If the status is “in progress” → the system retries the request until a final outcome is received.
• If unsuccessful → the system triggers customer alerts and notifications.

11. Frontend status checks

In parallel, the frontend also begins checking the final payment status via GraphQL and Microservice #1. Depending on the status retrieved, the system responds accordingly:

• Displays a success page if the payment is confirmed.
• Shows an error page or triggers another status check if the result is still pending, often accompanied by a loading animation to keep the customer informed.

Potential Pitfalls in the Payment Page Model

Even with careful planning, several challenges can occur:

Browser or Network Failures

The JavaScript promise may fail if the customer’s browser crashes, if there’s an internet outage, or if a frontend bug occurs. To prevent lost orders, we developed logic to proactively submit draft orders.

Without this safeguard, payments could be processed without triggering an order submission — creating risks for both the business and the customer. A better long-term solution would be for SAP DP to provide a callback mechanism to notify the system about completed payments.

Payment Method Filtering

Currently, available payment methods are loaded only on the frontend via DPJSLIB. For complex B2B scenarios, it is often necessary to filter payment options per customer or region.

A more scalable solution would be for SAP DP to offer a REST endpoint that returns all available methods. The integrated system could then apply filtering logic at the backend level.

Payment Finalization

At present, the system must continuously poll SAP DP endpoints to confirm payment status. This can create unnecessary load. A more efficient solution would again be callback notifications, where SAP DP pushes the final status to the backend.

Other Important Considerations

• SAP DP is not plug-and-play – Each PSP requires dedicated integration and close collaboration.
• PSP software testing – Every PSP delivers its own software, which must be thoroughly tested to ensure functionality.
• Local regulations – Each country has unique rules that must be complied with.
• Transaction times vary – Performance depends on PSP software, network conditions, and integration specifics.

Integrating SAP DP Direct PSP with the E-commerce System

In this model, the frontend integrates directly with the PSP. There is no universal guideline because every PSP works differently. To decouple business logic, we implemented a dedicated middleware REST service (Microservice #1) between the e-commerce platform and PSP.

Figure 5. Direct PSP Solution Architecture

Step 1. Payment completed at the PSP

The payment process is executed entirely by the chosen PSP. The final payment status is usually communicated via dedicated PSP webhooks, which confirm whether the transaction was successful or not.

Step 2. Making SAP DP aware of the transaction

Since the payment bypasses SAP DP during the checkout, the system does not automatically know that the transaction has taken place. To solve this, SAP provides dedicated endpoints that can be used to register (or “pair”) the PSP transaction with SAP DP.

Step 3. Middleware handles the pairing

In our customer’s solution, this step is handled by a dedicated Microservice #3. It listens for the PSP’s webhook notifications, collects the necessary payment details, and calls the SAP DP endpoints to create a corresponding transaction record inside SAP DP.

Step 4. Transaction created in SAP DP

Once paired, SAP DP creates a new internal transaction object representing the payment. At this stage, SAP DP enriches the record with its own identifiers such as tokens, authorization numbers, and other metadata needed for further financial processing.

Step 5. Pushing data to SAP ERP

After the transaction has been successfully created in SAP DP, all payment details are forwarded to SAP ERP, along with the corresponding order data. This flow is orchestrated by the e-commerce platform in combination with Microservice #2.

Step 6. Enabling financial operations in ERP

With both order and payment data now present in SAP ERP, the SAP DP module ensures smooth post-payment management. This includes activities such as reconciliation, accounting entries, or initiating refunds if necessary.

Figure 6. Direct PSP Sequence Flow

(Hover over the diagram to zoom into specific areas. A high-resolution version is also available for download in PDF and PNG formats.)

Our Findings About Direct PSP

• Every new PSP requires new development work.
• However, you gain a more personalized UX/UI compared to the Payment Page integration model, and you can use all PSP features because you are not limited by any intermediary interfaces.
• Clearer responsibility: the PSP owns the entire payment journey, reducing ambiguity in case of defects.

Conclusion

Our client’s project proved that both models have their place.

• Payment Page → faster rollouts, easier PSP switching, less complexity for commerce teams.
• Direct PSP Connection → maximum control, richer PSP-specific features, but higher development effort.

In either case, SAP Digital Payments remains the backbone, ensuring that all payment data flows securely into SAP ERP. This enables consistent accounting, refunds, and compliance — no matter how many PSPs are involved.

In a world where seamless digital payments define customer experience and operational agility, SAP Digital Payments enables enterprises to unify, secure, and scale transactions across markets and payment providers.

Digital payments have become a cornerstone of both B2C and B2B commerce. Customers expect speed, flexibility, and security at checkout, while enterprises need consistency, transparency, and compliance across geographies. Meeting these requirements is increasingly complex, given the multitude of payment service providers (PSPs), regulatory landscapes, and customer preferences.

SAP Digital Payments addresses these challenges by acting as a central hub between SAP S/4HANA, e-commerce systems, and external PSPs like Stripe, PayPal, or Adyen. The solution simplifies integration, standardizes financial processes, and provides full transparency across the Order-to-Cash cycle — making digital payments not only a technical enabler but also a strategic differentiator.

The business overview provided in this article is based on insights shared by two of our experts at Striped Giraffe: Mariusz Święs, Chief Architect, and Damian Woźniak, Technical Architect.

What is SAP Digital Payments?

SAP Digital Payments add-on is a secure cloud service that connects your SAP S/4HANA system with external payment providers. Instead of building separate integrations for each country or payment method, SAP Digital Payments acts as a central hub. Customers can pay with their preferred methods — credit card, PayPal, Apple Pay, Google Pay, or local PSPs — and S/4HANA automatically receives the results for further processing.

How it works

Digital payments are not isolated transactions; they are part of an end-to-end process that spans from checkout to financial clearing. SAP Digital Payments ensures that each step — from authorization to refund — is seamlessly connected and transparently reflected in SAP S/4HANA.

1. Customer checkout – A buyer places an order in your e-commerce system and selects a payment method.
2. Payment execution – The payment request is securely transmitted to the payment service provider, either via SAP Digital Payments interfaces or through a direct connection, depending on the setup.
3. Authorization / Direct funds capture – Depending on the integration setup, the PSP either authorizes the payment (reserving the amount on the customer’s account) or directly captures the funds. In both cases, the PSP may also decline the transaction if the payment cannot be processed. Authorization provides flexibility for later settlement, while direct capture immediately transfers the money. Companies choose the approach based on their business model, customer experience goals, and legal requirements.
4. Confirmation in S/4HANA – The result is sent back to your ERP. Orders are posted together with payment details, ensuring that finance immediately sees the correct status.
5. Financial follow-up – SAP Digital Payments enables further actions like refunds, cancellations, or financial clearing within S/4HANA, with full transparency for customer service and accounting teams.

Integration models

Enterprises differ in how they approach payment integration. The choice of integration model depends primarily on market coverage, regulatory obligations, customer expectations, and the degree of control a company wants over the checkout process.

SAP Digital Payments supports two complementary approaches, giving organizations the flexibility to select the most effective model for each business line or geography.

Payment Page

In the Payment Page model, the checkout window is displayed via SAP Digital Payments, without exposing the underlying PSP to the e-commerce system. This PSP-agnostic approach means the merchant does not have to build or maintain individual PSP integrations. Sensitive payment data is processed by the PSP and never touches the e-commerce application or ERP, which reduces compliance overhead and risk.

This model makes PSP switching or onboarding easier, with minimal IT effort. It is particularly beneficial for global players expanding into new countries with diverse payment methods.

Example:

A global fashion retailer operating across Europe and Asia frequently switches PSPs to adapt to local regulations and transaction costs. With the Payment Page model, these changes are transparent to customers and require no redevelopment of the e-commerce platform. The company benefits from faster rollouts into new markets, while ensuring regulatory compliance and keeping checkout consistent worldwide.

Direct PSP Connection

In the Direct PSP Connection model, the e-commerce system connects directly to a chosen PSP. SAP Digital Payments is still used in the background for financial posting, reconciliation, and transparency in SAP S/4HANA.

This approach offers maximum control over the checkout experience, allowing merchants to tailor and fully brand the flow according to their requirements. It also supports specialized use cases such as recurring payments or subscription billing.

In addition, negotiating directly with PSPs can lead to more favorable transaction fees, providing both operational flexibility and potential cost advantages.

However, this model requires dedicated integrations for each PSP, which can increase implementation effort and maintenance complexity. Businesses must weigh the benefits of a fully customized and branded checkout against the additional resources needed for integration and ongoing support.

Example (B2C):

A subscription-based video streaming provider integrates directly with a PSP to optimize recurring billing. Direct control allows the company to manage retries, failed payments, and customer notifications in a way that aligns with its retention strategy.

Example (B2B):

A manufacturing supplier selling spare parts to industrial clients chooses direct integration with a PSP to support large-value transactions and invoice-based payments. By integrating directly, the company ensures that its B2B customers can use agreed credit lines and deferred settlement terms, while SAP Digital Payments handles backend posting and reconciliation. This model offers the supplier full flexibility to align payment processing with contractual agreements and financing structures.

Challenges

Adopting a digital payment hub solves many issues but enterprises need to be aware of several challenges that still require planning and governance:

• Transaction Reliability: Payment confirmations can fail due to network or browser issues, making fallback mechanisms essential.
• Provider-Specific Complexity: Each PSP has its own rules, software, and compliance requirements, which means that onboarding a new one still requires coordination.
• Processing Delays: Transaction settlement times vary across payment providers. While SAP Digital Payments standardizes workflows, it cannot eliminate provider-side delays.

Business benefits

For enterprises, the value of SAP Digital Payments extends beyond payment execution.

• Customer experience: Customers get modern, local, and secure payment options without extra manual handling.
• Efficiency: Finance teams no longer need to reconcile payments manually — transactions and refunds are processed automatically in S/4HANA.
• Compliance & security: Companies can rely on secure payment processing without storing sensitive card data (such as PAN or CVV) in SAP ERP (S/4HANA). Instead, tokens issued by the PSP are used, significantly reducing risks related to PCI-DSS and data protection regulations (GDPR).
• Flexibility: Companies can integrate multiple PSPs across countries with a unified approach.
• Transparency: Sales, customer service, and finance all see real-time payment statuses within the Order-to-Cash process.

Fit into S/4HANA processes

Payments are not a back-office detail; they sit at the heart of the Order-to-Cash cycle, directly impacting sales, finance, and customer service. SAP Digital Payments ensures that transactions are seamlessly embedded into S/4HANA processes, strengthening revenue recognition, improving cash flow visibility, and enabling rapid customer response.

• Sales Order Management: Payment authorization is tied directly to order creation.
• Billing & Accounts Receivable: Payments are matched automatically, reducing delays.
• Customer Service: Refunds and cancellations are handled directly in S/4HANA, improving response time.
• Finance & Controlling: Real-time payment postings improve cash flow visibility and reporting accuracy.

Compatibility with non-SAP systems

ERP systems

SAP Digital Payments add-on is tightly bound to SAP S/4HANA. It cannot function as a standalone payment hub for third-party ERP systems, because its architecture is designed to post, reconcile, and track payments directly in S/4HANA financial modules. Any attempt to connect SAP Digital Payments with a non-SAP ERP would require extensive custom integration and would not provide the same level of reliability or supportability.

E-commerce platforms

In contrast, e-commerce systems are far more flexible. SAP Digital Payments can be triggered from virtually any e-commerce or digital sales platform — SAP Commerce or otherwise — through API calls to SAP Business Technology Platform (BTP). This allows companies to preserve their existing customer-facing systems while still leveraging the secure payment orchestration and reconciliation capabilities of S/4HANA. What is required, however, is an integration layer or middleware to handle the API communication and ensure that transaction results flow back correctly into the ERP.

Pricing and commercial models

SAP Digital Payments add-on is delivered as a cloud-based subscription with a flat annual fee, offering unlimited transaction usage in its standard form. In some vendor or implementation contexts, pricing may be tailored based on transaction volumes or deployment scope, per individual agreement. Additionally, adapters (e.g., Stripe) are typically free, but standard charges for the use of SAP Digital Payments still apply.

Conclusion

SAP Digital Payments is more than a technical integration layer — it is a strategic enabler of global commerce. By centralizing payment processes, reducing compliance risks, and embedding transactions into the Order-to-Cash cycle, it delivers measurable value for both finance and customer experience.

Enterprises evaluating SAP Digital Payments should:

• Assess compatibility with their current e-commerce and ERP landscape.
• Define integration strategies for each market (Payment Page vs. Direct PSP).
• Consider operational cost savings and compliance benefits when evaluating ROI.

Next, we will dive into the technical implementation of both integration models, illustrated by a real-life project delivered for one of our global clients.

→ Continue to part 2: How to Integrate SAP Digital Payments into Your E-Commerce Architecture

The adoption of cloud computing in the pharmaceutical industry means operating in one of the world’s most stringent and heavily regulated environments. However, companies that are mastering this complexity are already demonstrating that innovation and compliance can go hand in hand — provided that cloud computing is used strategically and not viewed as a panacea.

Executive Summary

• Over 80% of leading life sciences companies have already migrated critical workloads to the cloud.
• The life sciences cloud market is expected to reach USD 9 billion by 2030.
• Key drivers include exploding data volumes, on-demand high-performance computing, AI-enabled research, global collaboration, and stringent regulatory requirements.
• Cloud adoption is not only about cost and flexibility – it has become a strategic enabler for innovation, compliance, and new business models in pharma and life sciences.
• Hybrid approaches that combine cloud agility with the control of local systems are becoming increasingly popular.

Global studies show that more than 80% of top life sciences companies have already moved critical workloads into the cloud, ranging from research pipelines to enterprise resource planning systems. Analysts expect the life sciences cloud market to surpass USD 9 billion by 2030, driven by the exponential growth of data, the urgency of faster R&D, and the demand for real-time collaboration across a highly distributed ecosystem.

Unlike many other industries, pharma faces unique challenges:

• strict regulatory oversight
• intellectual property protection
• the need to handle sensitive patient data with the highest standards of confidentiality

Cloud adoption is therefore not just about cost or flexibility. It is about enabling new ways of working — from AI-driven drug discovery to federated data sharing across global consortia — while staying compliant with EMA, FDA, and GDPR requirements.

This article highlights the key drivers of cloud adoption in pharmaceuticals and life sciences—and shows why hybrid models are the most realistic option for many companies.

Exploding Data Volumes

Pharmaceutical research generates unprecedented volumes of data — from genomics and proteomics to real-world evidence and connected manufacturing equipment. Traditional on-premises infrastructures struggle not only with the scale but also with the heterogeneity of these sources.

The challenge is compounded by the rise of unstructured data such as lab notes, medical images, sensor feeds, and electronic health records. These datasets are essential for discovery and clinical insight, yet they cannot be efficiently managed with traditional relational databases or siloed storage.

Cloud advantage

Cloud-native data architectures provide the flexibility to ingest, organize, and analyze both structured and unstructured datasets at scale. Data lakes and data lakehouses enable pharma companies to centralize genomics files, imaging datasets, and real-world evidence into a single, queryable environment.

Combined with scalable compute, these models allow researchers to run advanced analytics and AI pipelines without overwhelming local infrastructure. The result is a more unified data foundation that supports faster discovery and better downstream decision-making.

Challenge

Data classification, governance, and security models must be carefully defined to meet regulatory requirements.

High-Performance Computing for Research

Drug discovery increasingly depends on compute-heavy workloads such as protein folding, molecular modeling, and advanced simulations. Building and maintaining such infrastructure in-house is costly and inflexible.

Cloud advantage

Cloud high-performance computing (HPC) provides scalable clusters that researchers can access on demand, dramatically reducing the time and cost of computational experiments.

Boundary

For very stable, consistently high computing loads, on-premises HPC systems may still be more cost-effective.

Advanced Analytics and AI

Artificial intelligence and machine learning are central to modern pharma, from predicting drug-target interactions to analyzing imaging data. These workloads require both scalable compute and advanced data governance.

Cloud advantage

Cloud providers offer managed ML platforms, automation pipelines, and AI-specific services that make these processes faster to deploy and validate.

Speed and Agility

Time is critical in the pharmaceutical industry. Whether it is launching a new drug, validating a manufacturing process, or running a clinical trial, delays translate into lost revenue and slower patient access to treatments.

Cloud advantage

Cloud infrastructure can be provisioned in hours instead of weeks, giving pharma organizations the agility to experiment, test, and deploy new solutions faster.

Boundary

Critical core systems often remain deliberately local in order to ensure regulatory security and business continuity.

Global Collaboration

Pharmaceutical research and development is by nature distributed. Large pharma companies work across multiple sites and time zones, often in partnership with contract research organizations, academic institutions, technology providers, and regulators. Ensuring that these diverse stakeholders can collaborate effectively is essential to shorten development timelines and improve research quality.

Cloud advantage

Cloud-based platforms enable this kind of cooperation by providing controlled environments where teams can access standardized data and tools in real time, regardless of geography or organizational boundaries.

Pragmatic approach

Hybrid models ensure that particularly sensitive data remains local and only results are shared.

Secure Data Sharing

Data sharing has become a critical enabler of progress in today’s pharma. Whether in precompetitive collaborations, public–private research consortia, or multi-sponsor clinical trials, no single organization can generate the diversity and scale of data required to fuel modern science. Sharing data accelerates discovery, improves the robustness of clinical evidence, and allows companies and institutions to pool insights that would be impossible to achieve in isolation.

At the same time, sharing sensitive research data — clinical records, compound libraries, or patient information — requires careful handling to protect confidentiality, intellectual property, and compliance. The challenge lies in enabling knowledge exchange across organizations without exposing raw datasets. Federated learning is one approach that helps achieve this balance.

Cloud advantage

Federated learning itself is a machine learning paradigm, but in pharma it is typically deployed on cloud infrastructure that provides the secure orchestration, elastic compute power, and regulatory-grade compliance required for multi-party collaboration.

In federated learning, instead of pooling data into a single repository, each organization keeps information within its own environment. Algorithms are sent to the data, trained locally, and only the learned parameters are shared back to a central model hosted in the cloud.

By running training locally while using the cloud to coordinate and aggregate models, companies can unlock collective insights without ever moving raw data.

Challenge

Only cloud infrastructures that comply with regulatory requirements are suitable for this purpose.

Compliance and Auditability

Pharma operates under some of the world’s most demanding compliance regimes, including GxP, 21 CFR Part 11, and EMA guidance.

Cloud advantage

Cloud platforms support compliance with automated audit logs, traceable data flows, and infrastructure as code — reducing the manual burden of validation and increasing transparency.

Challenge

Companies remain responsible for ensuring that cloud environments are recognized and validated by regulators.

Manufacturing and Quality Control

In pharmaceutical production, quality assurance is mission critical. Manual inspection processes are slow, resource-intensive, and prone to human error.

Cloud advantage

Cloud-enabled machine learning models can automate anomaly detection, improving both efficiency and product safety.

Pragmatic approach

Critical production systems remain local, while the cloud is used for analytics.

ESG and Supply Chain Transparency

Investors, regulators, and patients increasingly expect visibility into environmental, social, and governance (ESG) practices. For pharma, this means tracking emissions, energy use, and supply chain performance.

Cloud advantage

Cloud platforms allow data from disparate systems to be consolidated, analyzed, and presented in real time, turning ESG reporting into a tool for both compliance and operational improvement.

Boundary

For stable, predictable workloads, on-premises may be more economical.

Cost Efficiency and Flexibility

Maintaining large on-premises infrastructures ties up capital that could otherwise fund R&D.

Cloud advantage

Cloud’s pay-as-you-go model shifts investment to operational expense and allows organizations to scale spending in line with actual usage. This flexibility is especially valuable for companies with fluctuating compute demands, such as during peak trial phases.

Conclusion

Cloud adoption in pharma and life sciences is not a uniform journey but a balancing act between innovation, compliance, and risk management. The industry has demonstrated that the cloud can handle the scale and sensitivity of genomic research, clinical trials, manufacturing, and ESG reporting — provided that deployments are carefully designed with security and regulatory frameworks in mind.

For CIOs and senior leaders, the lesson is clear: cloud is not simply about replacing on-prem infrastructure but about rethinking operating models. Success requires hybrid strategies that combine the scalability of cloud with the control of local systems, strong partnerships with trusted providers, and a clear roadmap for data governance.

Those who master this balance will not only reduce costs and accelerate R&D but also create new opportunities for collaboration, transparency, and patient trust.

You might also like:

• Artificial Intelligence in Pharma and Healthcare (Booklet) » Read more
• Pharma’s AI Readiness Index: Who Leads the Race? » Read more
• ESG – It is our responsibility (E-Book) » Read more
• Data sharing — Challenges and Opportunities » Read more
• Data governance – the linchpin of efficient data management » Read more

The adoption of digital technologies in construction has evolved from isolated experiments into a defining factor of business performance.

While the industry has long been considered skeptical of technology, recent years have shown that digital technologies can fundamentally change how buildings are designed, constructed, and operated.

Today, the question is no longer whether to adopt them, but how to leverage them most effectively to deliver value at scale.

Benefits of Digital Technology Adoption

According to Deloitte’s “State of Digital Adoption in the Construction Industry 2025,” companies that have embraced digital tools report the most common benefits as higher productivity (60%), increased profitability (58%), improved quality of goods and services (55%), and reduced operational and management risk (55%). These benefits go beyond short-term cost savings and strengthen long-term competitiveness, project quality and work safety.

Figure 1. Benefits observed since introducing new technologies

Source: State of Digital Adoption in the Construction Industry 2025, Deloitte, February 2025

Note: Businesses could select up to three benefits.

Digital adoption also contributes to revenue growth and better project delivery. Deloitte’s analysis shows that the introduction of just one additional digital technology increases the likelihood of a project being delivered on time and under budget, while also raising expectations for annual revenue growth. For an average construction business with $100 million in revenue, this translates into $1.14 million in additional growth per year.

Another critical impact is on safety. Digital tools help identify risks early, digitize approval and monitoring processes, and enable earlier interventions, directly improving worker safety and compliance outcomes.

The 10 most important digital technologies in construction in 2025

1. Building Information Modeling (BIM)

Building Information Modeling is a digital representation of physical and functional characteristics of a building or infrastructure. Modern BIM operates within Common Data Environments (CDEs), increasingly standardized through ISO 19650. Advanced levels extend BIM into 4D (adding time for sequencing and scheduling) and 5D (integrating costs), enabling real-time simulation of how a project evolves over time and budget.

Benefits:

• Better interdisciplinary collaboration.
• Early detection of planning errors and collision checking.
• Support for sustainable construction methods.
• Increased productivity through prefabrication and visualization.

Key use cases:

• Early-stage conceptual modeling for cost estimation and risk detection.
• Clash detection and multi-discipline coordination (architecture, structural, MEP).
• 4D/5D project planning and budget control.
• Facilities management and lifecycle monitoring after handover.

Real-world example:

On London’s Crossrail, one of Europe’s largest infrastructure projects, BIM was used to integrate design, construction, and asset data across the entire lifecycle. The digital workflows enabled better coordination, reduced errors, and improved handovers, generating measurable efficiencies and savings in the millions of pounds.

2. Digital Twin

A digital twin is a virtual replica of a physical asset, process, or system, continuously fed by real-time data from sensors and operations. In construction, this typically involves three layers: the physical twin, the operational twin (monitoring performance), and the intelligent twin (predictive, enhanced with AI/ML).

Benefits:

• From reactive to proactive project management.
• Improved operational efficiency and lifecycle optimization.
• Early risk minimization and increased safety.

Key use cases:

• Design validation and process optimization before execution.
• Real-time monitoring of building systems (HVAC, structural integrity, energy use).
• Predictive maintenance through feedback loops between engineering and operations.
• Augmented reality overlays for site inspections.

Real-world example:

Helsinki created a city-scale digital twin that allows planners to simulate everything from solar energy potential to wind flow around new developments. The model cut turnaround times for studies, improved stakeholder engagement, and made planning more transparent and data-driven.

3. Robotics and Automation

Robotics in construction ranges from autonomous material transporters and robotic welders to layout-planning bots and collaborative robots (cobots). When combined with AI, these systems can adapt to new tasks and operate in hazardous or repetitive environments.

Benefits:

• Relief from the shortage of skilled workers.
• Higher precision and round-the-clock operation.
• Improved occupational safety and resource efficiency.

Key use cases:

• Autonomous robots for material handling on construction sites.
• Precision welding and assembly in high-risk environments.
• Automated layout marking and surveying.
• Prefabrication factories with robotic assembly lines.

Real-world example:

At a quarry site in Sweden, a collaboration between Volvo CE and Skanska tested an autonomous “electric site” where robotic, battery-powered machines performed heavy operations. The pilot demonstrated that automation could maintain productivity while cutting energy costs drastically and reducing CO₂ emissions by over 90%.

4. Artificial Intelligence (AI)

AI in construction analyzes vast amounts of project data to support cost forecasting, risk management, quality control, and safety monitoring.

Benefits:

• More accurate cost and schedule planning, improving overall project predictability.
• Reduced delays and risks through proactive data-driven insights.
• Automated document review and compliance checks, ensuring faster approvals and regulatory adherence.

Key use cases:

• AI-powered cost estimation based on historical project data, market pricing, and productivity benchmarks.
• Predictive analytics for project scheduling and risk management.
• AI vision systems for quality control and safety inspections.
• Natural Language Processing (NLP) for document analysis and contract compliance checks.

Real-world example:

On large Nordic projects, cameras mounted on workers’ helmets were connected to an AI system that compared actual site progress against BIM and schedules. This approach created an objective digital record of construction activities, enabling earlier detection of delays and discrepancies, and helping project teams keep delivery on track.

5. Generative AI (GenAI)

Generative AI is a subset of AI capable of creating new content — from design alternatives to material compositions — by learning from vast datasets. In construction, it integrates with BIM and digital twins to produce new design options, optimize site layouts, and automate documentation.

Benefits:

• Faster planning processes through design alternatives.
• More efficient approval and compliance checks.
• More sustainable and innovative solutions.

Key use cases:

• Generative design of 3D building components or layouts.
• Generative materials design for sustainability and performance optimization.
• Site plan optimization based on constraints such as resources, climate, and codes.
• Automated content summarization from technical documents and drone surveys.

Real-world example:

In the Netherlands, a contractor leveraged a generative design platform to assess thousands of potential layouts for new residential neighborhoods. The system intelligently optimized the trade-offs between land utilization, energy efficiency, and financial feasibility, providing planners with rapid, data-driven insights into alternative designs and enabling more sustainable, well-informed development decisions.

6. Augmented Reality (AR), Virtual Reality (VR), and Mixed Reality (MR)

AR overlays digital models on the physical environment, VR immerses users in a fully virtual environment, and MR combines both approaches. In construction, these technologies are increasingly tied to BIM and digital twins.

Benefits:

• Fewer planning errors through improved visualization.
• Faster coordination with clients and authorities.
• Safer training in virtual environments.

Key use cases:

• Virtual project walkthroughs for clients and teams before breaking ground.
• AR-enabled inspection of unbuilt or underground structures in situ.
• Safety training simulations in VR environments.
• Wearable MR devices (e.g., XR10 helmets) for field teams to access real-time project data.

Real-world example:

Irish construction companies use MR headsets to overlay BIM models directly on construction sites, resulting in fewer errors and smoother installations.

7. Drones (Unmanned Aerial Vehicles – UAVs)

Drones in construction provide aerial imagery, LiDAR, and video data for surveying and inspections.

Benefits:

• Fast, cost-effective data collection.
• Reduced risk for workers in hazardous areas.
• Real-time monitoring of construction progress and safety.

Key use cases:

• Topographic surveys and terrain mapping.
• Real-time site inspections and monitoring.
• Inventory and material tracking.
• Drone-based photogrammetry for digital twins.

Real-world example:

On UK infrastructure projects, drones are now routinely deployed for surveying and inspections. They provide high-resolution imagery of large and hard-to-access areas in minutes instead of days, reducing the need for manual surveys, improving safety by keeping workers out of hazardous zones, and cutting inspection costs.

8. Internet of Things (IoT) and Sensors

IoT refers to interconnected devices and sensors deployed on job sites and within building systems. They capture data such as vibration, temperature, or energy consumption and transmit it in real time for analysis.

Benefits:

• Predictive maintenance and optimized resource use.
• Enhanced safety through wearables and sensors.
• Sustainability through continuous monitoring.

Key use cases

• Monitoring machinery usage and performance.
• Wearable IoT for worker safety and tracking.
• Smart building systems (HVAC, lighting, energy monitoring).
• Structural health monitoring in bridges, tunnels, and buildings.

Real-world example:

Concrete pours on major projects in Ireland were instrumented with wireless maturity sensors that transmitted real-time data on curing progress. This enabled engineers to make evidence-based decisions about when to remove formwork or proceed with the next stage, shortening schedules and reducing reliance on traditional, slower testing methods.

9. Cloud Computing

Cloud platforms provide scalable storage, computing power, and collaboration environments for construction data and applications, including BIM, digital twins, and AI.

Benefits:

• Shared data foundation (“Single Source of Truth”).
• Lower IT costs and improved collaboration.
• Real-time access to BIM, IoT, and analytics data.

Key use cases:

• Centralized BIM and CDE platforms.
• Mobile-first collaboration tools for contractors and subcontractors.
• AI and analytics workloads run at scale.
• Integration of IoT sensor data into digital twins.

Real-world example:

An Irish contractor adopted a cloud-based common data environment to manage design and construction workflows. By centralizing documents, models, and field data, the project team was able to improve information flow, reduce errors, and achieve faster coordination between office and site.

10. 5G and Private Cellular Networks

5G provides high-speed, low-latency wireless connectivity, crucial for data-heavy applications such as real-time BIM collaboration, drone streaming, and IoT-based site monitoring.

Benefits:

• Reliable data flows for AR/VR and autonomous machines.
• More efficient site monitoring.
• Increased safety and productivity.

Key use cases:

• Real-time AR/VR field applications.
• Continuous IoT monitoring of equipment and structures.
• High-resolution drone video streaming.
• Remote control of autonomous equipment.

Real-world example:

In Sweden, pilot programs in collaboration with telecom providers have equipped construction sites with private 5G networks. The connectivity allowed real-time streaming from drones, remote operation of heavy equipment, and instant access to AR/VR applications in the field, resulting in safer operations and more efficient site management.

Conclusion: Make Digital a Core Strategy, Not an IT Project

The evidence is clear: digital technologies in construction boost productivity, quality, safety, and sustainability.

But it’s not just about choosing the right tools — it’s about strategically integrating them into core business processes. Companies that treat digitalization as a mere IT project leave significant potential untapped. Successful organizations align it with business objectives, governance, and employee training.

For construction firms, this means identifying the top technologies, launching scalable pilot projects, and building capabilities for holistic integration.

Digital transformation is no longer a vision for the future—it is an essential reality for successful, profitable, and safe construction.

Artificial intelligence is becoming a strategic imperative in the pharmaceutical industry, rather than just an experimental add-on.

The latest AI Readiness Index by CB Insights shows which global giants are best positioned to turn algorithms into competitive advantage.

The ranking underscores a turning point: leadership is no longer about “getting ready” for AI but about executing at scale, with drug discovery, clinical development, and manufacturing as the primary battlefields.

The Top Contenders in 2025

According to CB Insights’ latest analysis, Eli Lilly tops the list, followed closely by Merck KGaA, Bayer, AstraZeneca, and Roche. The narrow gap between second and fifth place (just 3.9 points) shows how intensely competitive the AI race has become.

Figure 1. Pharma AI readiness index – The top 20 companies

¹ Score is out of 100.

Source: “Pharma AI readiness: How the 50 largest companies by market cap stack up,” CB Insights, July 2025

1. Eli Lilly – From capital windfall to AI leadership

Eli Lilly’s booming revenues from its GLP-1 drugs — a class of treatments for diabetes and obesity — have given the company fresh resources to expand into AI. According to CB Insights, Eli Lilly has made 13 AI-related investments between August 2023 and May 2025 — more than any other pharmaceutical peer. Key moves include:

• Drug discovery: partnership with Insilico Medicine to accelerate target identification through generative chemistry.
• Regulatory automation: collaboration with Yseop for AI-driven clinical report generation.
• Medical devices: investment in RetiSpec for early Alzheimer’s detection.

With new AI-integrated research facilities in Indiana and Ireland ($7.3B), Lilly is blending infrastructure with algorithmic discovery at unmatched scale.

2. Merck KGaA – Digital governance as key to success

Despite a smaller market cap, Merck KGaA ranks second, driven by 10 AI investments during the analyzed period and a robust digital governance strategy.

• Its AI materials center in Japan ($70M) integrates AI into next-gen material sciences and bioprocessing.
• In R&D, the company partners with startups to apply AI in immuno-oncology and protein engineering.
• Merck’s data stewardship programs in Darmstadt have been cited as benchmarks in pharma for enterprise-wide data governance and architecture.

3. Bayer – Partnership powerhouse

Bayer excels in external collaboration, recording 21 AI partnerships, second only to Roche. When it comes to AI investments, Bayer ranks just behind the podium — holding fourth place among leaders with six deals during the analyzed period.

• The company’s AI platform for digital pathology (in partnership with Aignostics) enables faster biomarker discovery.
• Bayer’s Crop Science unit also feeds advanced data capabilities back into pharma, strengthening its AI architecture.
• Its €1.5B EU facilities are designed with AI-driven production and predictive maintenance systems.

4. AstraZeneca – Scaling AI globally

AstraZeneca is making AI central to its R&D operations.

• It invested $2.5B in an AI-powered R&D center in China, focusing on precision oncology and respiratory diseases.
• Through partnerships with BenevolentAI and Insilico, it applies machine learning to both small-molecule and biologics discovery.
• Its US and EU facilities integrate AI into clinical manufacturing, enabling faster scale-up of biologics.

5. Roche – Collaboration as a core strategy

Roche leads all peers in AI partnerships (22 in total during the analyzed period).

• The company’s Genentech unit works with Recursion and nVIDIA to power large-scale biological image analysis.
• Roche Diagnostics collaborates with PathAI to expand AI in pathology.
• It has pledged $50B over five years into AI-integrated R&D facilities in the US, positioning itself as a long-term ecosystem orchestrator.

The Next Five: Diverse AI Pathways

Beyond the top five, the remaining leaders also make significant bets:

• Merck & Co., Inc. – Building an AI-enabled vaccine manufacturing facility in North Carolina and deploying enterprise-wide LLMs.
• Sanofi – Developing CodonBERT, an AI platform for mRNA design, while committing $20B in US operations integrating AI and IoT.
• Pfizer – Deploying Vox, its Amazon-powered internal LLM, to streamline R&D data queries.
• Amgen – Partnering with Microsoft and nVIDIA on AI-accelerated supercomputing for protein modeling.
• AbbVie – Advancing ARCH, its AI-based target discovery platform, while investing $10B in AI-ready manufacturing facilities.

Preparing the Data Foundations for AI

AI readiness in pharma ultimately rests on the foundation of high-quality, well-governed data. Advanced algorithms and generative models are only as effective as the datasets they are trained on — a fact that becomes critical when AI is entrusted with research, clinical decision support, and operational optimization.

Beyond sheer volume, data must be accurate, consistent, and structured in ways that allow secure integration across clinical trials, manufacturing, and real-world evidence. This makes data quality the single most important determinant of successful AI deployment in the pharmaceutical sector.

Pharma companies therefore face three interrelated challenges:

1. Data governance at scale – Establishing enterprise-wide standards for data lineage, provenance, and compliance (e.g., Merck KGaA’s global data stewardship model).
2. Data integration and interoperability – Unifying siloed clinical, genomic, imaging, and real-world data into AI-ready architectures. Roche and Bayer are both building cloud-native data lakes to address this.
3. Data literacy and culture – Training thousands of scientists to interpret and apply AI outputs responsibly; AstraZeneca has rolled out global AI literacy programs across its R&D workforce.

A rising theme is cross-industry data sharing, particularly in oncology. Consortia like Project Data Sphere or Roche’s collaborations in cancer genomics show that the most promising AI applications will require shared, harmonized datasets. Yet data sharing remains bound by privacy regulations, IP protections, and competitive dynamics — a balancing act that defines the next frontier of AI collaboration in pharma.

Conclusion

The Pharma AI Readiness Index 2025 makes it clear:

• AI is the new currency of competitiveness in pharma.
• Leading companies are those that combine capital, governance, partner networks, and data-driven strategies.
• Data maturity will be decisive: those who optimally control, integrate, and share data will shape the medical breakthroughs of tomorrow.

You might also like:

• Artificial Intelligence in Pharma and Healthcare » Learn more
• Data governance – the linchpin of efficient data management » Learn more
• Data sharing — Challenges and Opportunities » Learn more
• Cloud Data Warehouse – Next-Generation Data Management (e-book) » Learn more

In B2B e-commerce, search is no longer a utility — it’s a sales engine.

Yet, according to Forrester, only 49% of B2B websites offer product search at all. Meanwhile, CXL Institute found that those who optimize it have seen sales from search-driven sessions jump by up to 43%.

With 70% of B2B leaders calling product discoverability a top priority, the message is clear: search isn’t a backend feature. It’s a commercial differentiator.

To see how B2B companies are meeting (or missing) this challenge in 2025, Zoovu analyzed the search experience across 30 enterprise-level B2B e-commerce websites.

The best performers share one thing in common: They build for how buyers think and buy, not how internal systems are organized.

Here’s what they’re doing differently:

1. Core search functionality

As B2B product catalogs grow and buyer expectations evolve, search functionality must be flexible enough to handle part numbers, industry-specific language, or vague problem descriptions. Whether it’s a procurement officer or a field technician, users expect intelligent results that adapt to their input — including misspellings, synonyms, or incomplete terms. Core functionality lays the groundwork for the entire experience.

What top performers are getting right:

• Fastenal handles typos and alternate phrasing with ease.
• 3M and Thermo Fisher support keywords, part numbers, and use cases.
• Caterpillar gets you from part number to relevant product in seconds.
• AO Smith stands out with its powerful filtering and sorting tools.

2. Product discovery & guidance

Finding a product is just the beginning — evaluating its fit is where the real decision-making starts. In B2B, buyers often face complex requirements: compatibility constraints, multi-component systems, or regulatory considerations. That’s why the best platforms go beyond filters, offering guided selling tools that help users compare options, build solutions, and explore alternatives. The goal isn’t to push products — it’s to support informed, confident decisions at every step.

What top performers are getting right:

• Schneider Electric and John Deere integrate guided selling tools directly into search.
• Schaeffler dynamically groups results and clearly displays compatibility.
• Straumann excels with compatibility logic and related items baked into results.
• Caterpillar stands out with compatibility indicators and robust bundling options.

3. Product detail experience

In B2B, a product isn’t just a SKU — it’s a bundle of specifications, certifications, documentation, and dependencies. Buyers need detailed, credible, and easily accessible information to evaluate options and minimize risk. That includes everything from CAD files and compliance data to stock availability and price calculators. A rich product detail experience not only builds trust — it accelerates decisions by reducing the need for back-and-forth with sales.

What top performers are getting right:

• Molex, Thermo Fisher, and Kennametal provide CAD files, spec sheets, and compliance docs right from the results page.
• Graco and Ingersoll Rand offer interactive product imagery and instant access to product details.
• Caterpillar provides price calculators, inventory info, and strong product tags.

4. Performance & experience

Speed and stability are crucial for B2B buyer satisfaction. In enterprise settings, users are often under time pressure and working across devices or departments. A slow or clunky interface becomes a barrier to purchase — especially when expectations are shaped by seamless B2C experiences. Responsive design, persistent filters, and fast loading times aren’t extras — they’re enablers of flow and focus in complex buying journeys.

What top performers are getting right:

• ABB preserves previous queries across sessions.
• 3M, Uline, and Neogen deliver lightning-fast and responsive search journeys.

5. Conversational search

As product complexity increases, so does the value of guidance. Conversational search brings a human-like layer to digital discovery, allowing users to express intent in natural language, clarify queries in real time, and explore solutions without navigating rigid taxonomies. Especially in B2B scenarios where needs are nuanced, conversational interfaces powered by AI can bridge the gap between catalog logic and customer logic — creating faster, more intuitive paths to the right solution.

What top performers are getting right:

• Graco has layered AI-powered chat directly into the search journey.
• Rockwell Automation integrates AI chat to support next-step guidance.
• Ingersoll Rand lets users ask configuration and compatibility questions in natural language.

“Designing a great search experience isn’t just about providing the right features. It’s equally about the quality of the product and customer data that power them,” says Danielle Rietsch, Head of Marketing at Striped Giraffe. “AI, personalization, and intelligent content delivery can add tremendous value — but only when built on a strong data foundation and aligned with how B2B buyers actually think and decide.”

How does your on-site search stack up — and are your data foundations ready to support it?

We’d be happy to talk. Contact us at experts@striped-giraffe.com.

You may also like:

• E-Book: B2B Commerce — More than just an online shop
• Study: B2B Commerce, Part 1 — Caught between economic stress test and digital efficiency
• Study: B2B Commerce, Part 2 — The Right Approach to Digitization
• Expertise: B2B E-Commerce

Generative AI (GenAI) has the potential to reshape the insurance industry more profoundly than any previous wave of automation.

“As an industry, insurance runs on data — but most of it isn’t cleanly structured,” says Łukasz Cempulik, DWH and BI Architect at Striped Giraffe. “We’re talking about thousands of scanned documents, handwritten forms, medical reports, legal correspondence, photos, videos, voice recordings — data that has historically required manual handling.”

With large language models (LLMs) and multimodal AI, insurers can now tap into this ocean of unstructured information to drive faster, smarter, and more consistent decisions — while reducing costs and improving service levels.

Cempulik adds: “GenAI doesn’t just automate existing processes — it introduces an entirely new level of cognitive capability. It can read a 40-page medical record, cross-reference it with actuarial tables, and summarize key risks in seconds. That’s enabling a different scale and depth of insight.”

How exactly are insurers putting these new capabilities to work? Let’s explore the key use cases:

1. Risk Assessment & Underwriting

Instead of sampling limited data points, GenAI reviews full electronic health records (EHRs), lab results, financial disclosures, wearable sensor logs, and even lifestyle indicators. It interprets hundreds of pages of medical records and cross-references them with actuarial data to surface risk signals no human would spot.

Impact: More accurate, objective underwriting — completed in hours, not weeks.

Traditionally, underwriting depended on predefined checklists and summary data. GenAI changes that by consuming full electronic health records (EHRs), diagnostic reports, financial statements, lifestyle surveys, wearable sensor logs, and even social media behavior (where permitted by regulation).

In practice, a GenAI model can:

• Parse hundreds of pages of clinical documents using natural language understanding.
• Extract relevant biomarkers, pre-existing conditions, or risk flags.
• Compare findings with actuarial models and propose a risk score.
• Draft underwriting recommendations with explainable reasoning.

This leads to more accurate and objective risk assessments — and far faster turnaround times.

Real-world example: A leading life insurer in Germany now uses a GenAI-based underwriting assistant that evaluates full medical histories and lab reports. Underwriters report a 60% reduction in time needed for complex case reviews, with improved risk segmentation outcomes.

2. Claims Processing

Claims often involve unstructured content: handwritten notes, PDFs, phone transcripts, photos or videos from property inspections, and even drone footage. GenAI excels at synthesizing such data.

Insurers deploy it to:

• Digitize and categorize scanned forms.
• Generate structured claims summaries based on adjuster narratives and evidence.
• Detect gaps, inconsistencies, or duplicate submissions.
• Auto-suggest decision pathways based on historical resolutions.

Integration with OCR (Optical Character Recognition), computer vision, and document classification tools enhances this pipeline.

Real-world example: A large multiline insurer in the Netherlands has integrated GenAI with its claims triage system. The solution reviews thousands of property damage claims weekly, cross-checking adjuster notes with photographic evidence and flagging inconsistencies. Claim resolution times have dropped by over 40%.

3. Fraud Detection

Fraud schemes often involve subtle patterns across communications, claims, and behavioral anomalies. GenAI brings a semantic lens to fraud detection.

Its capabilities include:

• Analyzing historical cases to uncover fraud indicators across text, image, and voice.
• Identifying linguistic inconsistencies across documents and emails.
• Spotting behavior shifts in claimant communication or submission frequency.
• Collaborating with machine learning models for anomaly detection.

Real-world example: A Scandinavian insurer integrated LLMs into their anti-fraud program. The system surfaced coordinated fraud rings based on similar phrasing in multiple claims. Within six months, the detection rate of suspicious cases increased by 28%.

4. Product Development

In a rapidly shifting market, GenAI enables insurers to innovate faster and more precisely.

It can:

• Analyze customer feedback from call center logs, surveys, and online reviews.
• Benchmark competitor offerings using scraped policy content.
• Generate ideas for new insurance products or dynamic pricing models.
• Simulate customer responses via synthetic personas.

Paired with traditional analytics, this leads to customer-centric innovation and faster go-to-market cycles.

Real-world example: A mid-sized insurer in Central Europe used GenAI to develop micro-insurance products for gig workers and freelancers. By analyzing usage patterns, customer complaints, and social media threads, the company introduced two new offerings that reached profitability within nine months.

5. Customer Service

Perhaps the most visible application, GenAI powers AI assistants that understand context, nuance, and policy complexity.

These systems:

• Retrieve answers from policy documents, prior correspondence, and regulatory guidelines.
• Generate personalized responses across channels.
• Understand sentiment, intent, and urgency to escalate when needed.
• Offer multilingual support for international markets.

GenAI also assists live agents by summarizing conversation history and suggesting next-best actions.

Real-world example: A major health insurer in the US deployed a GenAI-powered chatbot across web and mobile apps. Within weeks, it was handling over 70% of member inquiries without human intervention — maintaining a 92% satisfaction score.

Further application scenarios

Beyond the five areas outlined above, GenAI is beginning to influence:

• Regulatory reporting: Drafting compliance reports and explaining decisions in natural language.
• Broker enablement: Equipping sales reps with AI-generated product comparisons and selling points.
• Policy management: Simplifying coverage changes and policy renewals via conversational interfaces.

Transformation Starts with Data

The success of all of the above use cases hinges on a critical enabler: unstructured data readiness for AI.

“Unlocking GenAI’s full potential depends on how well insurers prepare their data,” emphasizes Cempulik. “You can’t just drop PDFs into a model and expect magic. You need to clean, classify, and annotate your unstructured content. That’s where proper data pipelines and governance come in.”

To prepare unstructured data for GenAI, insurers are increasingly investing in:

• Document digitization and OCR optimization.
• Data labeling and metadata enrichment.
• Knowledge graph integration for contextual understanding.
• AI governance frameworks ensuring explainability and compliance.

Done right, these investments not only improve GenAI performance but also lay the foundation for more transparent, auditable, and secure AI systems.

Conclusion

Generative AI is not a distant promise — it’s already reshaping how insurance works, from the underwriting desk to the customer help line. By unlocking the value in messy, unstructured data, GenAI gives insurers a competitive edge in speed, precision, and personalization.

But that edge won’t come by itself. It requires robust data strategies, interdisciplinary collaboration, and a strong understanding of GenAI’s capabilities and limits. For insurers ready to rethink their approach to information, the rewards are clear: faster insights, better products, and more satisfied customers.

“The real disruption isn’t just AI implementation — it’s what becomes possible when you prepare your data so AI can finally understand it and make use of it,” Cempulik adds.

You may also like:

• E-Book: Digitalization in the financial sector
• Blog article: Top 7 challenges of digitization in the financial sector
• Expertise: AI Readiness

Digital transformation unfolds its true value not within isolated applications but through the intelligent interconnection of systems and data flows.

In many organizations, IT systems operate in parallel without effectively communicating with each other. The result: redundant data entry, fragmented processes, lack of transparency, and hindered decision-making. In fast-moving markets, this can lead to a significant competitive disadvantage.

The alternative is an integrated IT approach: systems are seamlessly connected, data flows automatically, and processes run synchronously across departments. This kind of system integration is not only technologically feasible today – it is a key enabler of scalability, efficiency, and customer centricity.

Business Value of IT System Integration

1. Automation of Core Business Processes

Integrated systems enable the automation of standardized processes such as order processing, invoicing, or inventory management. Manual effort is minimized, turnaround times are shortened, and error rates are significantly reduced.

Example: Automatically transferring online orders from the webshop to the ERP system speeds up invoicing and reduces the need for manual corrections.

2. Improved Data Consistency and Governance

Centralized data storage and automated synchronization reduce the risk of duplicates and inconsistencies. Uniform data sources form the foundation for reliable analyses and informed decision-making.

Example: Connecting CRM and ERP prevents conflicting customer records and ensures a consistent view across sales, customer service, and accounting.

3. Informed Real-Time Decision-Making

Real-time data from interconnected systems enables leaders to act faster and with greater precision. Information silos are eliminated, and dashboards provide consolidated insights as a foundation for data-driven management.

Example: An integrated BI tool immediately analyzes point-of-sale data and supports proactive pricing and assortment adjustments.

4. Enhanced Service Quality and Customer Retention

A 360-degree view of the customer across all touchpoints becomes a reality through system integration. Customer data, interactions, and transactions can be consistently tracked and leveraged.

Example: Integrating the helpdesk with the e-commerce system gives the support team direct access to order information, resulting in faster response times.

5. Optimized Supply Chain Processes

In an interconnected supply chain, integrated systems reduce inventory levels, prevent overselling, and improve planning accuracy. This leads to more efficient operations and enhanced delivery performance.

Example: Warehouse management, online shop, and shipping platforms are connected – inventories are automatically updated and bottlenecks identified early.

6. Scalability and Future-Readiness

A modular integration architecture makes it easier to add new systems, locations, or services. Companies remain technologically agile and can respond to growth or market changes with flexibility.

Example: An e-commerce company expands internationally and integrates new payment providers without modifying the existing ERP system.

7. Reduction of Operational Costs

Fewer manual tasks, lower error costs, and more efficient processes lead to reduced operational expenses. Resources can be allocated more effectively to value-generating activities.

Example: Automated invoice generation from order data replaces manual entry and accelerates the billing process.

8. Secure and Compliant Data Processing

Centralized control of access rights, data flows, and audit trails not only improves IT security but also supports compliance with regulations such as GDPR or HIPAA.

Example: In the pharmaceutical industry, integrating clinical data platforms, LIMS, and ERP systems ensures compliant documentation in line with EU-CTR.

9. Enabler for Omnichannel Strategies

Consistent customer experiences across all channels require seamless system integration. Only then can context, preferences, and history be used effectively across web, app, in-store, and service environments.

Example: A loyalty program is linked to the online shop and POS system – customer benefits apply automatically across all channels.

Challenges in IT System Integration

Despite the numerous advantages, implementing a successful integration strategy is complex. Organizations often face challenges ranging from outdated system compatibility to data inconsistencies and security risks. These issues can not only delay integration but also drive up costs significantly.

Below are key challenges along with practical solutions:

1. Technical Complexity

Systems are often based on different technologies, architectures, and standards. Older or proprietary applications in particular may have incompatible protocols, security mechanisms, or data structures.

Impact: Without thoughtful planning, technical complexity can delay projects, increase costs, and raise risk levels.

Solution: Conduct a comprehensive system analysis and develop a detailed integration roadmap to identify potential incompatibilities early. Using standardized interfaces (e.g., REST APIs or message brokers) simplifies future integrations.

2. Legacy Systems

Legacy systems are often difficult to connect to modern platforms due to missing interfaces, outdated data formats, or their critical role in daily operations.

Impact: They act as integration bottlenecks and increase complexity and cost.

Solution: Middleware, API gateways, or integration layers can “wrap” legacy systems, making their functions accessible without changing the core applications.

3. Lack of Communication Standards

Different systems often use incompatible protocols, message formats, or authentication methods.

Impact: Point-to-point integrations emerge that are maintenance-intensive and error-prone – particularly during system updates.

Solution: Consistent integration is supported by industry standards (e.g., REST, SOAP, OAuth 2.0), internal guidelines, and a central integration layer — such as SAP CPI or Informatica Cloud Application Integration — which wraps system-specific APIs and exposes standardized interfaces.

4. Incompatible Data Formats

Varying data models, fields, and validation rules complicate consistent integration.

Impact: Inaccurate data mapping leads to faulty reports, poor decision-making, and degraded customer experiences.

Solution: A unified data model, supported by ETL (Extract, Transform, Load) processes for asynchronous transformation and integration platforms like SAP CPI for real-time mapping in synchronous scenarios, ensures structural consistency and reliable data exchange across systems.

5. Security and Data Privacy Risks

Each additional interface increases the potential attack surface. Missing safeguards can expose sensitive data.

Impact: Data breaches lead to compliance violations, reputational damage, and financial losses.

Solution: End-to-end encryption, secure API gateways, strict access controls, and regular audits reduce exposure to threats. Integration should follow a “security by design” approach, with each system receiving only the data necessary for its function, in line with authorization standards.

6. Lack of Integration Expertise

Integration requires in-depth knowledge of APIs, middleware, data architecture, security, and cloud services – skills that are often lacking in-house.

Impact: Projects may be delayed, error-prone, or overly dependent on external providers.

Solution: Invest in internal training, build expertise, and ensure strategic knowledge transfer when working with external partners.

7. Limited Scalability

What works well today may be overwhelmed tomorrow by increasing transaction volumes or system complexity.

Impact: Performance degradation, outages, or reduced customer satisfaction may follow.

Solution: Design modular, cloud-native, and event-driven integration architectures that can scale elastically.

8. Cloud Integration Barriers

Cloud and on-premises systems differ in terms of security, data, and availability models.

Impact: Poorly orchestrated integrations lead to data silos, inconsistent user experiences, and compliance gaps.

Solution: Modern integration platforms with hybrid capabilities provide secure, unified management across both environments.

Conclusion

Integrating IT systems is a strategic lever for companies aiming to achieve digital efficiency, scalability, and customer proximity. It forms the technical foundation for end-to-end processes, reliable data analysis, and forward-looking business models.

Organizations that invest early in integration expertise, architectural design, and trusted technology partners gain not only operational benefits – they also position themselves for long-term digital success in a competitive landscape.

→ Continue to part 2: Strategic Patterns of IT System Integration

You might also like:

• IT System Integration, Part 2: Strategic Patterns of IT System Integration » Learn more
• IT System Integration, Part 3: Technologies and Methods for the Practical Application of IT System Integration » Learn more
• How to Integrate SAP Digital Payments into Your E-Commerce Architecture » Learn more
• IT Report: Digitization projects — 6 key findings on challenges and experiences with recommendations » Learn more
• Cost savings in software development » Learn more
• Mastering software delivery: Overcoming development hurdles » Learn more
• Low-Code Falls Short Without Full-Code » Learn more

Architecture of Modern Personalization Solutions

Personalization in B2B must be dynamic, context-sensitive, and cross-system – including custom pricing, relevant product suggestions, or tailored content based on customer segment, industry, or purchase history. Modern B2B companies therefore rely on decoupled, API-first system landscapes:

• Customer Data Platform (CDP): Serving as the central data hub, the CDP links account, user role, and interaction data across all touchpoints, enabling a consolidated view of business customers and their decision-making networks.
• Recommendation Engines: AI-based engines identify relevant cross- and upselling opportunities – based on purchase behavior, seasonal demand, or industry-specific patterns.
• Experience Layer: Shop portals, self-service tools, or offer configurators access the personalization system via APIs.
• AI & Predictive Models: These forecast reorder probabilities, sales opportunities, or customer churn and enable real-time, data-driven interactions.

This architecture creates the foundation for scalable, efficient, and precise personalization across complex B2B customer journeys.

Best Practices for Data Strategy and Quality

In B2B, data is often distributed across multiple systems and user roles – turning personalization into a data challenge. Key success factors include:

• Centralized Customer Identities: Company, contact, and role data must be accurately linked and maintained.
• First-Party Data Strategy: All relevant usage, transaction, and communication data should be centrally aggregated.
• Role-Specific Segmentation: Deliver differentiated content to buyers, technicians, or decision-makers.
• Compliance & Data Privacy: Even in B2B, GDPR-compliant consent management is essential.

Want to learn how to successfully implement personalization in your B2B business? Get in touch with our experts.

You may also like:

• E-Book: “Personalization 01” – Customer Experience
• E-Book: “Personalization 02” – E-Commerce Use Cases

JACEK POLKOWSKI
Marketing & Communication Manager
Striped Giraffe

DPP Accessibility for Online Consumers

While the exact regulatory requirements for e-commerce platforms are not yet determined, it is certain that online retailers will be obliged to provide some form of access to the digital passport directly on the product page.

If information is not presented in direct connection with the product, crucial details about sustainability, composition, or recyclability might only become available after delivery and scanning the QR code. That would defeat the very purpose of the DPP, which aims to empower consumers with transparency upfront.

With that in mind, let’s explore what the DPP means for e-commerce businesses and the opportunities it presents in an increasingly sustainability-driven market.

Flexible Implementation Approaches

While e-commerce platforms will be required to provide access to DPP data, the specific method of implementation will likely remain at their discretion.

Two primary approaches stand out: linking to an external source, such as an official EU registry, or integrating and displaying the full DPP data directly in the online store.

A simple hyperlink to an external repository is the most straightforward solution, ensuring compliance with minimal technical effort. For that reason, it is optimal especially for smaller online stores, which usually have limited personnel and financial resources as well as often rely on simpler SaaS e-commerce platforms with significantly fewer customization capabilities.

On the other hand, many larger retailers may prefer to import DPP data and present it within their own platforms to retain customer engagement and provide a seamless shopping experience. This approach ensures that shoppers do not need to leave the site to access sustainability and product lifecycle details elsewhere.

Optimized Product Page Design

For the best user experience, larger e-commerce platforms may want to integrate DPP data directly into product pages — either as a dedicated block of information in the product description or as a separate tab.

Some online retailers may take this a step further by creating a dedicated section on their platform that consolidates digital passports for all items they sell. In this case, a product page would simply include a link to its corresponding DPP within the same digital environment.

The final scope of information that the EU will require to be shared with consumers through DPPs has not yet been defined. One thing is certain: if a retailer decides to host DPPs directly on their e-commerce platform, they will need to publish all mandated data. However, displaying all the information directly on the product page could clutter it with excessive content, as not all details may be equally important to every customer.

In such cases, a hybrid approach could emerge as the most effective solution: displaying key sustainability highlights directly on the product page while linking to a more detailed breakdown for those who want to explore further. This ensures that the most critical information is immediately accessible, while the full DPP remains available for deeper insights.

Enhanced Visualization for Consumer Engagement

While the EU will require e-commerce businesses to share DPP data with consumers, it will not dictate how this information should be presented visually. This means retailers will not be obligated to display raw data in plain text and will have the freedom to develop their own methods for visually enhancing sustainability attributes.

Online retailers can leverage visual elements such as icons, infographics, interactive charts, or images to make the data more engaging. For example, a clothing retailer can use a graphical scale to indicate where a particular product falls in terms of water consumption during production.

Dynamic elements, such as animations or comparative visualizations, could further enhance the shopping experience. A furniture store, for instance, could create an interactive module showing how different sofas rank in terms of recyclability or material sourcing.

Product Comparisons Based on Sustainability Criteria

With standardized DPP data, online businesses can enhance their product comparison tools by adding sustainability metrics alongside traditional parameters like price, dimensions, or technical specifications.

For example, a customer comparing washing machines could evaluate models not just by energy consumption or capacity, but also by factors such as CO2 emissions during production, recyclability, or the percentage of recycled materials used. Similarly, a shopper looking at smartphones could compare repairability scores or the carbon footprint of each device.

This integration of DPP data into comparison tools empowers consumers to make more informed, sustainability-driven purchasing decisions.

Personalized Shopping Experiences

Many e-commerce platforms already utilize personalization engines to tailor product recommendations based on browsing behavior. Integrating DPP data into these systems can enable sustainability-conscious consumers to receive product recommendations aligned with their values.

For instance, a customer who frequently purchases organic or eco-friendly items may see a curated selection of products with the lowest environmental impact.

Relevant filters can also be applied automatically during a search, ensuring that a shopper looking, for example, for home appliances sees energy-efficient options first.

Promoting Eco-Friendly Choices

Sellers can use DPP data to highlight and promote products that meet the highest sustainability standards. Dedicated sustainable product categories could be introduced, such as Zero-Waste Fashion, Non-Toxic Cleaning Products or Most Recyclable Electronics.

Product filtering and sorting options as well as searching criteria could also include sustainability features, enabling customers to prioritize the most eco-conscious choices.

Additionally, e-commerce platforms could display sustainability insights at the checkout stage.

For instance, a customer purchasing multiple items could see an aggregated summary of the environmental impact of their cart, along with suggested swaps for greener alternatives.

It is All About the Data

Every idea presented above relies on one key element: the effective use of data from digital passports. For online businesses, this means not only accessing and processing this data but also integrating it seamlessly into their operations.

One thing is clear: retailers will not be required to collect product lifecycle data or create DPPs for the products they sell. That responsibility lies primarily with manufacturers. Retailers’ role is to ensure consumers can access this information. However, ambitious businesses will likely go further, leveraging DPP data to build competitive advantages.

To access this data, retailers will need to establish integrations between their systems (such as ERP or PIM) and external systems hosting official DPPs. But this is not a one-time task. DPP data is not static and may need to be updated every time some factors change, such as production methods or materials used.

This means retailers cannot simply pull the data once and forget about it. They will need mechanisms to synchronize and verify that their internal systems reflect the most current information.

Retailers already store some DPP-related data in their PIM or ERP systems, such as technical specifications, energy consumption, or product origin. However, integrating this with DPP data may lead to conflicts. For example, if a TV’s energy consumption in the DPP differs from the retailer’s existing data, the DPP information should take precedence. Retailers will need to reconcile these differences and ensure their systems align with the official DPP.

The DPP will not apply to all product categories at once. The EU is introducing it gradually, starting with specific sectors like electronics and textiles, with others to follow. This means retailers with diverse product portfolios must monitor which categories fall under DPP requirements and when.

Moreover, the regulatory landscape is still evolving. While the DPP framework is taking shape, many details remain undefined. Retailers must stay agile, as updates and new guidelines are likely to emerge once the system is in practice.

You may also like:

• Blog article: “Digital Product Passport: Data Management Challenges”
• E-book: “ESG – It is our responsibility”
• Booklet (PDF): “Digital Product Passport: All you need to know”

Understanding the DPP

The Digital Product Passport (DPP) is a groundbreaking initiative by the European Union designed to enhance transparency, sustainability, and accountability across product lifecycles.

At its core, the DPP will be a digital record that provides detailed information about a product’s environmental impact, materials, and recyclability. Accessible via QR codes, RFID tags, or web-based platforms, it will empower consumers, businesses, and regulators to make informed decisions that support the transition to a circular economy.

The primary goal of the DPP is to promote sustainable consumption and production. By offering clear and reliable data, it enables consumers to choose products with lower environmental footprints, while encouraging manufacturers to design for durability, repairability, and recyclability. This aligns with the EU’s Circular Economy Action Plan and European Green Deal, which aim to reduce waste, cut carbon emissions, and foster resource efficiency.

The DPP will include a wide range of information, such as the product’s material composition, carbon footprint, supply chain details, and end-of-life recycling instructions.

Initially targeting high-impact sectors like electronics, textiles, and batteries, the DPP is expected to expand to other industries over time, becoming a universal standard for product transparency.

For businesses, the DPP represents both a challenge and an opportunity. While it requires significant adjustments to data collection and management processes, it also offers a chance to build trust with consumers, differentiate from competitors, and drive innovation in sustainable practices.

As the DPP rolls out, it will play a pivotal role in reshaping how products are designed, manufactured, and consumed in the EU and beyond.

Data Challenges

At its core, the Digital Product Passport is a data-driven initiative, making its successful implementation dependent on a company’s ability to manage, integrate, and govern vast amounts of product information. This shift will have far-reaching implications for data strategy, IT infrastructure, and regulatory compliance, requiring businesses to rethink how they collect, store, and share information both internally and across supply chains.

In the following sections, we will explore the key data challenges, from standardization and interoperability to security and real-time updates.

Data Collection and Integration

Data for the DPP can come from multiple sources, including internal systems (e.g., ERP, PIM, PLM), supplier inputs, or third-party certifications. For example, a textile manufacturer might collect data on fabric composition from suppliers, energy usage from production facilities, and certifications from third-party auditors. Companies must establish robust processes and infrastructure to collect and integrate this information seamlessly.

Data Standardization

Implementing the Digital Product Passport (DPP) requires standardized data formats to ensure consistency and interoperability across industries and supply chains. However, the European Union has yet to finalize these specific technical standards, creating uncertainty for businesses preparing to comply with future requirements.

In the absence of official guidelines, companies must navigate fragmented industry practices and varying reporting frameworks. Existing standards like ISO 14040 (Life Cycle Assessment) and ISO 14067 (Carbon Footprint of Products) offer temporary reference points, but they may not fully address the broader scope of future DPP requirements.

Another challenge lies in ensuring data consistency across complex supply chains, where multiple stakeholders operate with different levels of technological readiness. Without clear protocols, discrepancies in data collection, formatting, and sharing can lead to inefficiencies and compliance risks.

To mitigate these challenges, businesses should prioritize flexible data management strategies that can adapt to evolving standards. Engaging in industry initiatives and fostering collaboration with supply chain partners will also be critical in aligning data practices ahead of formal regulatory mandates.

Data Verification

Ensuring the accuracy, reliability, and timeliness of data collected for the DPP is one of the most critical challenges for businesses. With multiple stakeholders contributing data across complex supply chains, inconsistencies and errors can easily arise.

Companies must establish robust internal data governance frameworks, including regular audits and cross-departmental collaboration, to validate information at every stage. This ensures that data reflects the true environmental impact, material composition, and recyclability of products.

To further enhance data credibility, businesses may need to engage third-party auditors to verify key aspects of the DPP, such as the origin of materials, production processes, and end-of-life recyclability.

Additionally, adopting standardized data formats and integrating advanced data management systems can streamline the collection and validation process, reducing the risk of errors.

Data Interoperability

Businesses must ensure that their data management systems can communicate seamlessly with relevant systems across the supply chain, as well as with EU-managed platforms (such as the central DPP registry) and third-party DPP service providers.

To enable seamless communication between systems, businesses will need to adopt standardized protocols for data exchange. These protocols will likely be defined by the EU as part of the DPP framework and may include specific data formats, APIs, or communication standards.

Real-Time Updates

For certain product categories, such as batteries for electric vehicles, each individual unit will have its own unique digital passport. In addition to static data — such as specifications, manufacturer details, and material composition — the passport will also include dynamic data, which evolves throughout the product’s lifecycle and must be updated in real-time. These data points may include usage history, performance metrics, carbon footprint tracking, and repair and maintenance records.

Managing and integrating real-time data into DPPs presents a significant challenge. Ensuring data accuracy, consistency, and security requires robust digital infrastructure capable of continuously collecting, processing, and updating records.

Data Security and Privacy

With multiple stakeholders granted access to DPP data under the Ecodesign for Sustainable Products Regulation (ESPR), ensuring security and controlled access is essential. While some DPP information will be publicly available, other data will be restricted based on user roles, requiring robust access management protocols to prevent unauthorized disclosure or misuse.

Balancing transparency with data protection is a key challenge. Sensitive information — such as supplier networks, production methods, and material compositions — must be shielded from competitors while remaining accessible to authorized entities, including market regulators, repairers, and recyclers. Implementing granular permission controls will be crucial to safeguard proprietary data without compromising the DPP’s intended benefits.

Compliance with GDPR and other data protection regulations further adds complexity. While most DPP data pertains to products, occasional inclusion of personal details (e.g., supplier contacts) necessitates strict security measures, encryption, and clear data governance policies. Businesses must ensure that only the right stakeholders can access specific datasets while maintaining full regulatory compliance.

Conclusion

The Digital Product Passport marks a major shift in how product information is collected, shared, and utilized. While it presents numerous data management challenges, it also unlocks opportunities for businesses to enhance transparency, build trust, and drive sustainability. For companies, success will depend on proactive data strategies, flexible digital infrastructure, and collaboration across supply chains.

You may also like:

• Blog article: “Digital Product Passport: Tips to Unlock New Opportunities in E-Commerce”
• E-book: “ESG – It is our responsibility”
• Booklet (PDF): “Digital Product Passport: All you need to know”

Reducing return rates with technological solutions

Environmental impact of returns

Nowadays, returns are a standard aspect of retail, often safeguarded by consumer protection laws worldwide. However, the return process poses serious environmental challenges. These are due to the use and choice of return packaging and labelling, reconditioning, storing as well as additional transportation required to pick up, sort, repack and redistribute products, especially if return centers are far away from the customer’s location.

In addition, many returned products frequently end up as waste, despite being new. High costs of restocking and refurbishing often make it uneconomical to return these products to the sales cycle, resulting in their discarding or even destruction.

A global view of returns

Evidence suggests that online shopping fuels higher return rates compared to traditional retail. For instance, according to the National Retail Federation (NRF), in the U.S., online returns reached 17.6% in 2023. In contrast, the return rate for brick-and-mortar purchases was significantly lower at 10.02%, excluding instances where online orders are returned in-store (often termed as BORIS – Buy Online, Return In-Store).

In Europe, Switzerland had the highest return rate per online buyer in 2023, at 62%, according to the Statista Consumer Insights survey. Germany closely followed with 55%, and the United Kingdom completed this podium with 54% of returned products. In France and Spain, this rate was slightly lower, at 46% and 44%, respectively.

Technology for reducing returns

Digital tools can help e-commerce businesses manage and minimize returns. In this regard, UNCTAD recommends two key actions:

• Detailed product information: Providing customers with more comprehensive and precise information about a product, including color, weight, size, and other pertinent dimensions.
• Consistency between the presentation of the product and its actual appearance: Ensuring that images and descriptions accurately reflect the real product minimizes returns due to unmet expectations.

In line with the second recommendation, Augmented Reality (AR) is emerging as a breakthrough technology that can significantly reduce return rates. By empowering virtual try-ons, spatial planning, product visualization, and measurement assistance apps, AR enhances the user experience and allows for immersive virtual environments.

A market survey conducted among over 4,000 shoppers across France, Saudi Arabia, the United Kingdom, and the United States revealed that around two-thirds of consumers who used AR technology were less likely to return their purchases.

Beyond UNCTAD’s recommendations, there are other technological solutions that can help minimize returns. As a company specializing in e-commerce technology, we recommend the following strategies, among others:

• Using automated and intelligent digital solutions to manage returns and streamline logistics and storage processes. For example, by optimizing reverse logistics, businesses can minimize waste and improve efficiency.
• Collecting and analyzing data on returns helps businesses identify trends and areas for improvement. For example, understanding which products have higher return rates can guide better product descriptions or adjustments in product offerings.
• Leveraging data and AI/ML for CX personalization. For example, by offering customers more relevant product recommendations, companies can increase the likelihood of product satisfaction and minimize returns caused by poor purchase decisions.

Looking for more practical ways to reduce returns and cut costs in e-commerce using digital technologies? Discover key strategies in our blog article.

Conclusion

E-commerce is often perceived as a virtual space, but its environmental impact is anything but intangible. While the act of purchasing may take place in a digital realm, the consequences of returns and the logistics that follow are very real and material. As UNCTAD emphasizes, the path to sustainable e-commerce requires concerted action by companies, governments, and consumers. From curbing free returns to adopting fair use policies and employing digital tools, the industry has several levers to pull. Ultimately, the goal is clear: to reduce the environmental impact of returns and enable more informed, conscious consumer choices in the online shopping sphere.

This article delves into proven strategies for cost savings in data management.

Centralizing Data in a Unified Repository

Many organizations today struggle with fragmented data ecosystems, which hinder efficient management and decision-making.

“Medium-sized companies in particular are increasingly tending to stock up on various ready-to-use cloud solutions, which unfortunately leads to a highly fragmented IT environment”, explains Łukasz Cempulik, DWH and BI Architect at Striped Giraffe.

These organizations often rely on multiple cloud-based tools, each serving a specific purpose. However, the lack of standardization between systems from different providers means that businesses frequently resort to manually preparing and uploading data to ensure these tools can perform tasks like generating personalized product recommendations or launching marketing campaigns.

Fragmented data systems lead to significant costs:

• Redundant data storage: Multiple departments maintain duplicate records, inflating data storage and processing expenses.
• Inefficient processes: Data preparation and manual transfers waste valuable time and resources.
• Missed opportunities: Without a unified view, businesses struggle to generate actionable insights and miss chances for growth.

Centralizing data in a unified data platform addresses these inefficiencies and reduces costs. It enables organizations to:

• Eliminate redundancy: Consolidate duplicate records, cutting storage costs.
• Streamline operations: Automate data flows, reducing time spent on manual tasks.
• Ensure compliance: Simplify regulatory reporting and avoid costly penalties.

Governance plays a crucial role in this process. A robust data governance framework ensures that data is accurate, consistent, and aligned with business needs.

“When selecting technology solutions, we always start with a thorough analysis of the client’s existing systems and specific requirements. This allows us to recommend the most suitable tools,” says Łukasz Cempulik. “Among the platforms we often recommend are solutions from Talend, Informatica, Snowflake, Microsoft, Google or AWS, as they offer flexibility and scalability for organizations aiming to centralize and govern their data effectively.”

Investing in Data Virtualization

In contrast to centralizing data in a single repository, Data Virtualization offers a more agile approach by creating a virtual layer that integrates data without moving it physically. This strategy is particularly effective in architectures like Data Mesh, where decentralized data ownership aligns with organizational domains.

Key advantages of Data Virtualization:

• Cost Efficiency: By eliminating the need for large-scale data migrations, virtualization significantly reduces storage and infrastructure costs.
• Scalability: Virtualization supports dynamic environments, integrating new data sources seamlessly.
• Enhanced Quality Management: Tools within the virtualization layer ensure consistency across datasets, even when spread across multiple systems.
• Faster Data Provisioning: Data virtualization accelerates the availability of data for critical applications, enabling quicker access to insights and reducing the time needed for traditional data preparation processes.

Automating Data Management Processes

Automation remains a cornerstone of cost-efficient data management. From ingestion to reporting, automating routine data processes reduces manual errors and frees up resources for strategic initiatives.

Key applications of automation in data management:

• Data Ingestion: Automating the capture and integration of data from disparate systems or data sources. This process ensures that data is continuously gathered from multiple locations without manual intervention, enabling a seamless flow of information into central repositories.
• Data Transformation: Applying business rules to data to ensure it is in the correct format or structure for analysis. For example, calculating aggregates or adjusting time stamps.
• Data Quality Management: Automated data quality checks at scale, including identifying and resolving duplicate records, missing values, or inconsistencies.
• Analytics and Reporting: Automating the generation of reports and dashboards to provide near-real-time insights to stakeholders. This enables faster decision-making and eliminates the manual effort of pulling and analyzing data for business reports.
• Data Backup and Archiving: Automating data backup and archival processes to ensure data is stored securely and efficiently. This is especially important for disaster recovery and meeting retention policies.

“Organizations often underestimate the ROI of automation in data workflows,” notes Łukasz Cempulik. “Not only does it streamline processes, but it also ensures consistency across systems, reducing hidden costs from errors and inefficiencies.”

Organizations leveraging automation have reported up to 40% reductions in operational costs related to data management, according to recent industry benchmarks.

Cloud-Based Solutions for Data Management

Migrating data management to the cloud offers a powerful alternative to traditional on-premises infrastructure. Platforms like Amazon Web Services (AWS), Google Cloud, and Microsoft Azure provide scalable tools for storing and processing large volumes of data, making them ideal for organizations looking to optimize their operations.

How it reduces costs:

• Eliminates on-premises infrastructure costs: By moving away from local servers, businesses can cut down on the expenses of maintaining physical infrastructure, including server hardware, software updates, and IT personnel.
• Pay-as-you-go model: Cloud services allow businesses to only pay for the resources they actually use, reducing wastage and enabling cost-efficient scaling.
• Scalability: As data volumes grow, cloud platforms provide easy scalability, eliminating the need to constantly invest in new hardware to keep up with demand.

Despite the clear cost advantages, many companies — particularly in highly regulated industries — are hesitant to move their data to public clouds due to security concerns. According to Łukasz Cempulik, while it is understandable that organizations in sectors like finance and healthcare are cautious, leading cloud platforms are often more secure than traditional on-premises setups.

Cloud providers invest heavily in security infrastructure, offering robust encryption, continuous security monitoring, and compliance certifications that many internal IT departments may struggle to match. Furthermore, leading platforms are equipped with disaster recovery protocols and advanced backup solutions, ensuring business continuity in case of disruptions.

For highly regulated industries, cloud services now offer features like dedicated cloud instances and private cloud setups to meet stringent compliance standards (e.g., GDPR, HIPAA). These solutions ensure that data is stored in secure, isolated environments, addressing concerns around data privacy and security.

Conclusion

Effective data management is not just a technical challenge — it is a financial imperative. By centralizing data, implementing virtualization, automating workflows, and transitioning to cloud-based solutions, organizations can significantly reduce costs while enhancing operational efficiency.

Success lies in aligning these strategies with the unique needs of the organization. After all, cost savings are only meaningful if they lead to better insights, smarter decisions, and a stronger competitive edge.

Is your organization ready to transform its approach to data management while realizing cost savings?

Let’s talk!

E-commerce has come a long way from the monolithic setups that once dominated the industry. In traditional architectures, like the earlier versions of SAP Commerce, every component – backend and frontend – was tightly bound together. This meant that even the smallest change could require a full redeployment. The use of legacy technologies like JSP only amplified these limitations. But as customer expectations for personalization, new features and experiences skyrocketed, the rigidity of monolithic systems became an obstacle.

Enter headless commerce. By decoupling the backend from the frontend and bridging them through APIs, headless architecture offers a flexibility that meets modern needs. Backend teams can focus on core functionalities without needing to tailor everything for a specific frontend, making it easier to support multiple channels like mobile, web, and even IoT, with faster release cycles and fewer dependencies.

SAP recognized this shift and began moving its Commerce Cloud platform towards a more composable, API-driven model. This transformation now reaches a key juncture: as of Q2 2027, SAP will release its final update that includes the Accelerator UI components, marking the start of a new era where Accelerator-based storefronts will no longer receive support. Starting from Q3 2028, support will end entirely, leaving users with a choice: adapt or maintain the deprecated code on their own.

So, what options do current Accelerator users have?

With a range of new, adaptable frontend strategies on the horizon, there is no shortage of solutions to help them make a smooth transition.

As an SAP Gold Partner specializing in SAP Commerce, we recommend four key strategies for a successful transition:

• leveraging SAP’s Composable Storefront
• exploring other ready-made storefront solutions
• implementing a CMS centralized solution
• building a custom frontend tailored to specific business needs.

Option 1: SAP Composable Storefront

SAP Composable Storefront, formerly known as Spartacus, represents SAP’s modern alternative to the legacy Accelerator extensions and addons.

Choose this option if your team is still getting up to speed with modern frontend technologies and would benefit from a solution that is already developed and maintained by SAP, providing a ready-made framework with built-in best practices.

Modern Framework: Built on a contemporary JavaScript application framework, SAP Composable Storefront delivers a single-page application experience, allowing smooth transitions without page reloads. This approach mimics the responsiveness of a mobile app and greatly enhances the user experience.

Simplified Customizations: Unlike its monolithic predecessor, Composable Storefront minimizes the need for complex customizations, making it easier to implement, iterate, and maintain over time. Teams can focus more on customer-facing updates rather than backend dependencies.

CMS Integration: While a CMS is required, it plays a lighter role in managing the storefront. This setup enables teams to focus on dynamic content creation and branding, with less dependency on backend-heavy configurations.

Rapid Feature Updates: With a headless, API-driven structure, teams can independently update the frontend experience, speeding up the delivery of new features and optimizations. This flexibility allows teams to experiment with new user-facing features faster, supporting a more agile approach.

Omnichannel Ready: As customer engagement expands across platforms, SAP Composable Storefront is optimized for multi-channel strategies, allowing easy access to new channels like mobile, social commerce, and more.

Option 2: Ready-Made Storefront

Ready-made storefronts offer an out-of-the-box, pre-configured frontend solution designed to work seamlessly with backend systems, including SAP Commerce Cloud. This approach is ideal for businesses that want to transition to a modern frontend experience without extensive development or customization efforts.

Choose this option if your team wants a straightforward, modern frontend solution without the burden of extensive development and maintenance work. This option suits organizations looking for a flexible, future-proof approach to storefront management while keeping backend integration simple and manageable.

Quicker Start with Modern Frontend: Ready-made storefronts come prebuilt on contemporary frontend frameworks, making it easy for teams to adopt modern frontend practices without the complexities of building from scratch.

Backend for Frontend (BFF) as a Service: With BFF delivered as a Platform-as-a-Service (PaaS), the setup reduces the backend development work, handling key integrations while giving your team more flexibility to focus on frontend updates and user experience improvements.

Streamlined Maintenance: While there are typically higher licensing fees, ready-made storefronts reduce the ongoing maintenance and monitoring workload, as much of the infrastructure and backend integration is managed by the provider.

Prebuilt Integrations: Many ready-made storefront solutions come with prebuilt integrations for a wide range of platforms, including CMS, e-commerce, payment gateways, and Product Information Management (PIM) systems. This setup can significantly reduce the time and cost involved in connecting to essential systems and launching the storefront.

Ease of Technology Replacement: Thanks to their integration flexibility, ready-made storefronts simplify the process of swapping out backend components. Whether it is e-commerce platforms, payment providers, CMS solutions, or PIM systems, this architecture supports easy technology transitions if you decide to switch providers down the line.

Option 3: CMS as a centralized solution

The CMS upfront approach places the CMS as the central hub for managing both content and e-commerce pages. Here, all web pages — including product and transactional pages — are hosted within the CMS, which then retrieves e-commerce data directly or through a middleware layer. This setup can streamline content management and ensure consistency across your digital experience.

Choose this option if you have a reliable CMS solution in place that suits your needs, manage large volumes of content frequently, and run a more standardized e-commerce model without complex customization requirements. This option is ideal for businesses looking to streamline content management while keeping e-commerce capabilities straightforward.

Unified Content Management: With a CMS handling all pages, you gain a single, centralized platform for managing both e-commerce and content pages. This can simplify updates, particularly if your site is content-heavy and requires frequent modifications.

Direct or Middleware Integration: This setup can directly connect to your e-commerce solution, ensuring smooth and real-time data flow for things like product details, pricing, and inventory. Alternatively, a middleware layer can be used to mediate connections, offering an additional layer of customization and control over data handling.

Optimized for Standard E-Commerce Features: This approach is ideal for e-commerce setups that rely on widely used, straightforward features without extensive customization or unique self-service capabilities. It is a practical solution for businesses whose needs align with standard e-commerce functionality, without the demand for highly tailored content layouts or complex user interface flows.

Cost-Efficient Solution: By managing most of the frontend through a CMS, you can reduce development costs, as well as maintenance associated with complex e-commerce-specific frontends. If your team is already proficient with your current CMS, this option allows you to leverage that existing expertise.

Option 4: Custom Frontend (Greenfield)

Building a custom solution from scratch allows organizations to create a storefront tailored to their unique needs, rather than relying on the building blocks offered by software vendors.

Choose this option if your organization has specific, non-standard frontend needs, your team is skilled in a different tech stack, or you prefer not to be constrained by vendor strategies that may not align with your business vision.

Tailored Frontend Experience: This option is ideal for businesses whose frontend requirements significantly diverge from standard offerings. It enables the creation of a truly unique user interface and user experience that aligns perfectly with your brand and business goals.

Alignment with Team Expertise: If your development team has expertise in technologies that differ from the stacks proposed by vendors, a custom solution empowers them to leverage their skills fully. This choice minimizes the need for extensive retraining and allows the team to work more efficiently with familiar tools and technologies.

Independence from Vendor Strategies: A custom approach gives you the freedom to avoid aligning your business strategy with that of your e-commerce vendor. Vendors may adapt to trends that do not necessarily fit your vision, leading to additional costs and potential disruptions. This option allows you to develop a strategy that suits your budget and long-term objectives without the pressure of vendor timelines or updates.

What’s next?

As you navigate this transition, remember that there is no one-size-fits-all solution. Your company’s specific context plays a crucial role in determining the best approach.

Selecting an option is just the beginning; you will also need to consider the following next steps:

• Migration Strategy Plan: Develop a comprehensive strategy to ensure a smooth transition from your current setup to your chosen option.
• Timeline Estimation: Create realistic timelines for each phase of the transition to manage expectations and keep stakeholders informed.
• Communication Plan: Establish clear agreements with the business regarding timelines and the scope of changes to ensure everyone is aligned.
• Implementation: Execute the plan with diligence and adaptability, keeping a close eye on potential challenges and adjustments.

We have years of experience in developing and configuring e-commerce systems for leading European companies. They all prefer a flexible e-commerce solution for the reasons already mentioned. However, their requirements and conditions vary widely.

Are you also interested in a new solution? Write to us at experts@striped-giraffe.com and let’s get started.

There are numerous ways to control software development costs. In this article, we focus on three selected areas: balancing automated test coverage within continuous integration (CI), using feature flags to roll out features incrementally, and implementing smart automation practices in continuous delivery (CD).

For more ideas on how to optimize the software delivery process, read our article “Mastering Software Delivery: Overcoming Development Hurdles.”

Align coverage with automated testing

Automated testing is essential to ensure that code changes are tested early and often. IT teams often aim for 100% unit test coverage and 80 to 90% integration test coverage. However, high coverage is not a guarantee of quality.

Unit testing

Mariusz Święs, Chief Software Architect at Striped Giraffe, warns against aiming for too much code coverage in unit testing.

“While these tests validate the structure of the code, they may overlook critical business functions. Furthermore, high coverage can be costly, often without a sufficient return on investment.”

Recommendations:

• Quality over quantity of tests: Instead of focusing on test coverage, invest in high-quality tests that verify how the code fits into the larger code base.
• Prioritize tests by code impact: Not all code created is equal. Prioritize comprehensive testing in high-impact areas such as billing and data.
• Focus on frequently updated components: Automate tests for frequently changed functions to ensure greater stability as your code evolves.
• Trust your team’s expertise: If your team has a strong track record of writing high-quality code, comprehensive testing of all minor functions may not be necessary.
• Set practical coverage goals: Aim for about 80% coverage; the last 10-20% can require a disproportionate amount of effort.

Integration testing

Integration testing, which validates entire workflows, often adds more value to a business than isolated unit testing. However, these tests are often poorly managed and focus too much on code structure rather than functionality.

For example: a team we consulted had developed and continuously executed 1,100 integration tests in Java, but only reviewed the results every three weeks. About 400 tests failed each time, mainly because the tests verified the code structure rather than the business logic.

“We recommended switching from writing tests in Java code to developing behavior-driven tests in Cucumber and its descriptive language Gherkin. This made it possible to simply write test scenarios in English and ensure that they included the verification of the business logic. Additionally, we recommended involving QA specialists to validate all test scenarios.”

This change improved the relevance of the tests and reduced maintenance.

Feature flags for controlled rollouts

Feature flags help balance rapid feature rollouts with system stability. They enable developers to release new features incrementally and maintain continuous delivery without major disruptions.

Key use cases:

• Incremental feature rollouts: New functionality is initially released to a small group of users, allowing teams to identify and troubleshoot issues. If successful, access is gradually extended to larger groups until the feature is fully rolled out to all users.
• A/B testing: Teams can monitor the performance of different versions of a feature or UI components to ultimately select the iteration that is most stable or well-received by users.
• Targeted user access: Flags provide precise control over which user groups can access certain features. This approach enables exclusive access, personalized experiences, and targeted feedback without additional code changes.
• Instant rollback: If a new feature causes problems or falls short of user expectations, you can instantly revert it to the previous version using feature flags.

Feature flags accelerate the delivery of new features by allowing developers to incorporate code changes without making them immediately accessible to end users. This enables “silent” testing in a live environment. This approach mitigates the risks of traditional deployment, which can result in costly downtime – as much as $5,600 per minute, according to Forbes – due to bugs that make it into production. Teams can also identify and fix problems incrementally, reducing the need for expensive rollbacks.

Strategic Automation for Continuous Deployment

In general, continuous deployment (CD) automates and regularly releases code changes, increasing agility and reducing human error. However, it is important to decide strategically what exactly should be automated.

Optimal use cases for automation:

• Environment deployment: Automate the creation and configuration of development, test, and production environments.
• Monitoring: Implement automated monitoring for real-time performance insights.
• Rollback procedures: Automate rollback processes to quickly return to stable versions.
• CI/CD integration: Automate the integration of CI processes into CD workflows.
• Performance testing: Automate performance testing to quickly identify issues.

Processes to approach with caution when automating:

• Critical feature releases: Features that are business-critical or require precise coordination across multiple teams should be deployed manually.
• Complex database migrations: Manual processing may be required for large-scale migrations, especially for critical data transformations, to minimize risks associated with data integrity.
• User Acceptance Testing (UAT): Manual testing remains essential to gather qualitative data and understand user experience.
• Edge-case handling: Manual monitoring of rare edge cases can be more efficient and enable teams to avoid complicated automation scripts that provide limited value.

By strategically selecting which components to automate and manage manually, organizations can optimize their continuous delivery efforts and increase productivity while ensuring high-quality releases.

Conclusion

Streamlining software development processes is essential to reducing costs and increasing efficiency. By focusing on automated testing, feature flags, and strategic automation in continuous delivery, organizations can reduce unnecessary costs and improve product quality. By prioritizing high-quality testing and controlled rollouts, organizations can optimize development and accelerate time to market, benefiting both the business and its users.

Optimize your software delivery with our proprietary 5-step program!

Download our exclusive offer here.

As both individual and business customers increasingly prefer online experiences for product discovery and purchases, e-commerce is emerging as a powerful tool, reshaping how vehicles, parts, and services are marketed and sold.

This article examines how various players in the automotive sector are adapting to this evolving landscape, showcasing innovative e-commerce strategies and business models.

Transformation to B2B e-commerce

Automakers are increasingly recognizing the importance of B2B e-commerce in transforming their interactions with traditional distribution networks. A prime example is Mercedes-Benz’s B2B Connect platform.

B2B Connect is a comprehensive digital tool designed for independent workshops and service providers who work with Mercedes-Benz vehicles. This platform offers a centralized solution for accessing a wide range of resources, including ordering genuine parts, utilizing diagnostic tools, and accessing up-to-date repair and maintenance information.

Some key features include:

• Integration with the WebParts system: It allows users to quickly identify and order Mercedes-Benz genuine parts and remanufactured parts directly in B2B Connect, ensuring that workshops can efficiently find the correct parts for any specific vehicle.
• Digital Service Booklet (DSB): This feature offers VIN-based service history and documentation, facilitating management of service records and helping workshops maintain accurate and up-to-date service histories for the vehicles they work on.
• Diagnostic tools: The platform provides access to advanced diagnostic tools, ensuring that workshops can perform precise and effective vehicle technical inspections.
• Workflow Integration: The platform is designed to support the entire workflow of a workshop, from vehicle diagnosis to repair and maintenance, streamlining processes and improving overall efficiency.

The Mercedes-Benz B2B Connect is continuously updated to include new features and capabilities, with recent improvements focusing on empowering workshops to access all the information, tools and products they need in one place.

Direct-to-consumer models

For decades, car manufacturers relied on vast networks of dealerships to reach customers. However, the rise of e-commerce has given them the power to sell directly to consumers, bypassing traditional intermediaries.

Tesla is a perfect example. The electric vehicle pioneer has embraced a direct-to-consumer (D2C) model, allowing customers to order custom vehicles online, track their production, and arrange for delivery — all without stepping foot in a dealership. This approach has not only cut costs but also provided Tesla with direct access to customer data, enabling personalized marketing and product development.

Tesla operates two distinct e-commerce platforms: its main site, tesla.com, and shop.tesla.com. The main site was built in-house to support the sale of Tesla’s electric vehicles, as no off-the-shelf solution could meet the company’s specific requirements. The system leverages a headless commerce architecture, where Tesla has decoupled the front-end from the back-end for greater flexibility. This setup enables API-driven customization, allowing for streamlined customer experiences such as fast delivery, flexible refunds, and better customer service.

And shop-tesla.com, Tesla’s store for accessory and lifestyle products, runs on Shopify Plus, a customizable SaaS e-commerce platform. This integration allows Tesla to efficiently manage a wide range of products, from EV chargers to branded apparel, while providing a reliable and scalable solution to handle the company’s high traffic and transaction volumes.

Tesla is a rare example of a pure D2C model among automakers. However, established players like Volvo, BMW, and Volkswagen are leaning towards hybrid D2C known as the agency model. These companies have long relied on and built relationships with a network of dealers and now seek to integrate these partners into new market realities.

Volvo’s recent transition to a direct-to-consumer approach in the UK illustrates this shift. As of June 2023, Volvo has closed its wholesale channel in the UK, making it the first market to adopt a fully online sales model. Under this new structure, Volvo itself acts as the retailer, while franchise partners operate under agency agreements. This change positions dealers as service providers who facilitate test drives, assist in the purchasing process, and handle vehicle delivery and servicing.

The core of Volvo’s sales operation now lies in its online store, volvocars.com, which offers transparent, fixed prices with no haggling or hidden fees. In addition, a centralized digital inventory system enables Volvo to manage stock nationwide efficiently, ensuring customers can quickly locate and obtain their vehicles.

The move away from a wholesale model is driven not only by cost savings — typically around 30% of the vehicle’s retail price — but also by the desire for complete access to customer data. Volvo realized that investing in advanced data analytics capabilities would help it better understand customer preferences and enhance the purchasing experience.

Volvo plans to extend this direct sales model to Germany starting from 2026, with further expansions in Sweden and Norway anticipated. The transition in Germany will be more gradual due to the need to convert operations at 240 retail locations, but the goal remains to create a unified and customer-centric sales process.

Volvo’s e-commerce platform is built on Salesforce Commerce Cloud, providing a robust solution for both online vehicle sales and parts integration with its dealer networks. With Salesforce CRM and Marketing Cloud embedded into the platform, Volvo has a unified system for personalization and customer management.

The Salesforce Interaction Studio plays a crucial role in real-time behavior tracking across the customer lifecycle, enhancing conversion rates by adapting interactions based on user activity. This setup offers strong omnichannel capabilities, ensuring a seamless transition between online and offline channels, driving not only customer satisfaction but also sales.

BMW Group and Volkswagen Group present similar approaches.

BMW Group has already rolled out an agency model in January 2024 for its MINI brand in Italy, Poland, and Sweden. The change will gradually expand to more European markets, and eventually include the BMW brand in 2026.

Volkswagen Group, in turn, is preparing to shift to this model with all six of its brands: Volkswagen Passenger Cars, Skoda, Seat, Cupra, Audi, and Volkswagen Commercial Vehicles. Ireland is one of the first markets to be affected by this change from 2026.

In the new model, Volkswagen’s dealerships will function as agents whose responsibilities will focus on the physical handover of vehicles and after-sales services such as maintenance and repairs. For these services, dealerships will receive a fixed fee instead of the variable profit margins they traditionally earned on sales.

From the customer’s perspective, the new model offers great flexibility and increased transparency in pricing and sales practices. Buyers can choose to complete their purchase entirely online, or they can visit a dealership to experience the vehicle in person.

B2B2C model

To balance the benefits of D2C with the advantages of their existing dealership networks, car manufacturers are also increasingly turning to the B2B2C model. This strategy still allows them to sell directly to consumers online, while integrating dealers into the transaction process, thereby maintaining strong relationships with these crucial intermediaries.

In the B2B2C model, manufacturers provide the products — vehicles and related services — while dealers handle the logistics, customer interactions, and final transactions. Consumers benefit from a seamless experience where they can browse, customize, and purchase vehicles online, with the support of local dealers who facilitate test drives, financing, and delivery.

One prominent example of this approach is Hyundai’s “Cl!ck to Buy” platform. This innovative B2B2C model enables customers to complete every aspect of the car-buying process online, from vehicle selection and customization to financing and final purchase.

The process begins with customers selecting or configuring online their preferred car model. The platform provides real-time inventory updates, allowing buyers to see exactly what is available and make informed decisions.

Once the vehicle is configured, customers choose a participating dealer — typically one nearby — ensuring that the dealer is part of the process from the start. This selection not only connects customers with a dealer for personalized support but also allows the dealer to provide tailored offers based on the customer’s preferences.

Next, the platform allows customers to explore pricing and various financing options, which can be customized to fit their financial situation. Customers receive personalized quotes and in some markets they can apply for credit online if needed.

Throughout this entire journey, Hyundai provides live support to assist with any questions, ensuring a smooth and stress-free experience.

After reviewing and accepting the final offer, customers can finalize their purchase entirely online. This includes completing necessary paperwork and choosing how they wish to receive their new vehicle. Customers can either pick up their vehicle from the selected dealership or opt for home delivery, adding an extra layer of convenience.

In some markets, Hyundai’s platform also allows for scheduling test drives, either at the dealership or at the customer’s home, further enhancing the buying experience.

This comprehensive approach combines the ease of online shopping with the trusted support of traditional dealerships, ensuring that customers enjoy a seamless, modern car-buying process while still benefiting from the expertise and service of their local dealers.

Hyundai’s Cl!ck to Buy e-commerce platform is powered by Amazon Webstore and Adobe Experience Manager (AEM). This unique combination allows Hyundai to offer an optimized and user-friendly online buying experience. AEM supports the creation of Hyundai’s interactive car configurators, enabling customers to easily personalize their purchases. With Amazon Webstore handling the backend, Hyundai benefits from a scalable and reliable system capable of managing large transaction volumes, while AEM ensures dynamic content delivery and personalized digital experiences for prospective buyers.

Conclusion

The future of the automotive industry is undeniably digital, with e-commerce reshaping how vehicles, parts, and services are marketed and sold. Tesla’s pure D2C model exemplifies this shift, while traditional automotive giants like Volvo, BMW, and Volkswagen are integrating online platforms with traditional dealership networks, ensuring that the customer experience remains seamless and supportive.

As all players in the industry embrace digital strategies, the blending of online and physical sales and customer service channels will likely become the new standard, offering both B2B and B2C customers a more flexible and personalized experience.

In e-commerce, there are many areas with significant cost reduction potential that can drastically increase a company’s profitability.

In this blog article, we will focus on three key areas with enormous potential for cost reduction: optimizing the e-commerce architecture, implementing self-service functions for customers, and tools for reducing the number of returns.

Optimizing the e-commerce architecture: a hybrid approach instead of replatforming

To meet increasing customer expectations, implement new business requirements and make business processes as efficient as possible, e-commerce systems must be continuously adapted.

In particular, monolithic systems or SaaS solutions that do not allow for customization reach their limits quite fast. In this case, it is recommended to optimize the current e-commerce architecture, starting with a comprehensive evaluation of the technology stack and integration capabilities.

There is no need to set up a new system; instead, certain functions can be separated and set up individually. This makes the system more agile and significantly more flexible. Custom requirements can be implemented as desired alongside ongoing operations. This saves resources and also improves the customer experience.

Digital self-services + automation + AI

Self-service functions in e-commerce significantly reduce operating costs because customers can independently perform tasks such as product configuration or after-sales support. AI-powered chatbots also provide them with immediate assistance. As a result, valuable resources are under less strain on the customer service or sales side.

Automation also reduces manual tasks, which lowers labor costs and further minimizes the risk of errors. In addition, the integration of APIs and robust software solutions ensures a smooth flow of data between systems, further increasing efficiency.

Ultimately, the key to a successful self-service strategy is to reduce friction – making it as easy as possible for customers to achieve what they need without contacting support, while providing a highly personalized, expedient experience. By doing so, organizations can achieve significant cost savings and increased operational efficiency overall.

Relevant content, data analytics and augmented reality to prevent returns

Management systems and visual aids

Technologies such as Product Information Management (PIM), Product Content Management (PCM) and Digital Asset Management (DAM) ensure accurate and consistent use of product data across all platforms.

This helps to avoid misunderstandings in advance. Combining this with a realistic presentation of the products using visual aids such as 3D views, high-quality images and product videos reduces the likelihood of returns.

Augmented Reality

AR tools offer significant cost-saving opportunities in e-commerce by enabling customers to virtually experience and interact with products, reducing the costs of on-site demonstrations and decreasing return rates through better customer understanding.

Although AR requires upfront investment, the long-term savings from reduced support, travel, and returns processing costs can lead to a quick return on investment (ROI).

AI and ML

Personalization engines based on AI and machine learning help reduce the cost of returns by offering products that are more likely to meet customer preferences. By leveraging customer data platforms (CDPs), companies can efficiently integrate customer data to make more accurate product recommendations, minimize unnecessary purchases, and reduce associated costs.

Automation in returns management

Automated returns management systems streamline the returns process, saving companies money. They enable customers to initiate returns online, offer alternatives such as exchanges, and optimize logistics.

By routing defective products directly to manufacturers or processing in-store exchanges, companies can significantly reduce their shipping and handling costs, thereby increasing overall operational efficiency.

Summary

By strategically investing in technology and optimizing key areas of their operations, e-commerce companies can significantly reduce their costs. At the same time, they can increase customer satisfaction, ultimately paving the way for sustainable growth and higher profitability.

Introduction

Benefits of partnering with an IT service provider

Setting requirements

Key evaluation criteria

Evaluation

Conclusion

Finding the perfect IT service provider is like finding a business soulmate. They bring specialized skills, cost efficiency, and flexibility, allowing your team to concentrate on what they do best.

However, on the other hand, choosing an inappropriate IT partner could result in the failure of your digitization projects in the future.

As our survey of German B2B companies conducted jointly with ECC KÖLN last year showed, 80% of respondents have canceled or “put on hold” at least one IT project within the last five years. Overall, around 13% of the projects initiated were canceled at some point.

Figure 1: Reasons for discontinued digitization projects

Question: Why were your digitization projects canceled?
Info: Respondents whose digitization and IT projects were canceled within the last five years; Multiple answers possible.

Source: “B2B Commerce – Caught between economic stress test and digital efficiency”, ECC KÖLN & Striped Giraffe, April 2023

Unfortunately, the majority of companies (68%) admitted that the wide variety of service providers and complex selection process are often overwhelming.

So, how do you find the right IT service provider?

In this article, we will dive into the benefits of partnering with an IT service provider, explore the different types of providers, and guide you on how to research and evaluate potential partners. By the end, you will have the tools to make a confident decision and secure a partnership that drives your business forward.

Benefits of partnering with an IT service provider

Let’s take a detailed look at the key advantages and benefits of companies working with external IT partners and technology providers.

Access to expertise

Partnering with an IT service provider grants you access to specialized skills as well as industry knowledge. They can suggest the right solutions to solve your problems and meet your challenges. Here is how:

• Specialized skills: IT partners bring specialized skills that might be lacking in your in-house team.
• Industry knowledge: They have extensive experience both in yours and other industries and can apply best practices to your projects.
• Optimal technology recommendations: They can recommend the best and most efficient technologies to achieve your business goals.

Cost efficiency

Reducing overhead costs compared to maintaining an in-house team is a significant benefit. In particular, IT service providers help by:

• Lowering recruitment costs: Avoid the high costs associated with hiring and training new employees.
• Reducing infrastructure costs: Minimize the need for investing in expensive hardware and software.

Scalability and flexibility

IT service providers offer the ability to adapt to business growth and changing needs. They usually provide:

• Scalable resources: Quickly scale resources up or down based on your project requirements.
• Flexible solutions: Adapt solutions to meet your evolving business needs.

Quick response to issues

While your IT department may be overwhelmed with other tasks, an IT partner is contractually obligated to respond promptly. Benefits include:

• Guaranteed response times: Service level agreements ensure timely responses to issues.
• Dedicated support: A reliable support system ensures you receive the attention needed to resolve issues swiftly.
• Proactive monitoring: Continuous monitoring to identify and address issues before they impact your business.

Acquiring new knowledge and skills

Learning from your IT partners can enhance your team’s competencies. Advantages include:

• Skill transfer: Gain new skills and expand your team’s knowledge by working alongside experienced IT professionals with a broad range of expertise.
• Innovative perspectives: Exposure to the latest industry trends and cutting-edge technologies can inspire innovative thinking within your internal team.

Improved project methodologies and practices

IT service providers bring modern methodologies and practices, improving project success rates. This includes:

• Agile approaches: Implementing agile methodologies that can enhance efficiency and adaptability.
• Best practices: Learning and adopting the latest industry best practices for IT project management.

Focus on core business

Allowing internal teams to concentrate on primary business activities is crucial. Benefits are:

• Operational peace of mind: Rest assured that your IT systems are in expert hands, with potential issues proactively addressed.
• Resource optimization: Free up valuable time and resources to focus on revenue generation, marketing strategies, and other critical business functions.
• Minimized downtime: Improved system reliability and reduced downtime enable uninterrupted focus on core operations.

Partnering with an IT service provider offers numerous benefits that can address common challenges and propel your business forward. By leveraging their expertise, you can ensure your technology investments deliver the desired results, allowing your internal teams to thrive in their primary roles.

Setting requirements

Before you decide on a service provider, you should define your requirements as precisely as possible. There are two possible scenarios:

• Searching for a contractor to support or carry out a specific IT project, or
• Looking for a long-term partner for comprehensive and permanent IT support

Depending on which of the two you have in mind, there will be different requirements and evaluation criteria. However, there are many universal aspects that should be considered in both situations.

General questions to consider

Here is a list of sample non-trivial questions you should answer at this stage:

• What were the reasons for the failure or cancellation of your past IT projects?
• Have your software development practices and methodologies worked so far?
• What are the non-negotiable priorities for your business?
• What must-have characteristics and qualities should your partner represent?
• What practices or behaviors will you not tolerate under any circumstances?
• Do you need a partner just for software development or implementation or do you also expect their support in IT consulting and advisory?
• Are you flexible and open-minded enough to evolve your IT practices, structures, and processes?
• Are you ready to change your tools and frameworks to optimize the process of delivering new software?
• What level of support and responsiveness do you expect?
• How do you measure success in your IT initiatives?

Case-specific questions to consider

When looking for an IT partner for different scenarios, tailor your questions to each situation:

Specific project support:

• What is the scope of your project?
• What core functional requirements must the project meet?
• What level of expertise and services are needed?
• What are the business constraints, including budget and timeframe?
• What technologies would you like to use for this project?
• How do your individual requirements differ from standard market solutions?

Long-term IT partnership:

• How do you envision technology contributing to your overall business strategy and growth?
• What are the areas, such as cybersecurity, cloud services, data analytics, or e-commerce, where you require ongoing support and innovation?
• What are your expectations for continuous improvement and innovation?
• Where do you excel in your IT capabilities, and where do you face challenges or gaps?
• How critical is scalability and flexibility in your IT infrastructure?
• What are your expectations for innovation and staying ahead of technology trends?
• How important is data security and compliance to your operations?
• What is your overall budget for IT services, and how do you plan to allocate it?

Key evaluation criteria

There are several key evaluation criteria to consider when choosing the right IT service provider, each playing a vital role in ensuring that the partnership aligns with business goals and delivers value.

Broadly, these criteria can be categorized into qualitative and quantitative aspects. Qualitative aspects focus on the provider’s ability to understand and adapt to the business’ unique needs, while quantitative aspects emphasize measurable outcomes and capabilities.

According to Deloitte’s survey, the top qualitative aspects valued in a service provider highlight the importance of relational and strategic attributes. Transparency emerged as the most valued aspect, followed by trustworthiness and understanding of the business. Proven results and clear communication were also highly rated, indicating that businesses seek partners who not only deliver but do so with clarity and reliability (Figure 2).

Figure 2: Top qualitative aspects companies value most in their service providers

Question: What are the top qualitative aspects you most value in a service provider?

Source: Deloitte Global Outsourcing Survey 2022

In the following sections, we will delve into some criteria in detail, exploring how each aspect contributes to selecting the right IT partner and ensuring a successful, productive and long-term partnership.

Scope of services and expertise

Ensure that the range of services and skills offered by a potential IT partner can meet both your current and long-term IT needs. The ability to combine diverse competencies, such as in e-commerce and data management, can prove invaluable to the success of your future projects.

Industry expertise

An IT partner with industry-specific knowledge can tailor solutions to your unique challenges and anticipate future needs.

Technological adaptability and innovation

Your IT partner should not only be proficient in current technologies, but also be able to use new technologies. This includes evaluating different technological options, recommending optimal solutions and explaining their pros and cons.

Software development capabilities

Let’s face it: even with low-code/no-code platforms, there will inevitably be programming tasks that you need to solve.

Therefore, if your business requires custom software solutions, ensure your IT partner has robust software development skills, including building new systems from scratch and customizing ready-made solutions.

Consulting proficiency

A competent IT partner should provide a comprehensive range of services, including strategic advice, risk assessments, and project management methodologies. They should be capable of conducting thorough IT audits, offering tailored solutions for your business challenges, and ensuring your IT infrastructure aligns with your business requirements.

Transfer of knowledge

Ensure the IT partner can effectively transfer knowledge to your team, facilitating smooth operation and maintenance of implemented solutions. This includes providing comprehensive training as well as detailed system documentation.

References

Verify the alignment of the IT service provider with your company’s objectives and profile, considering their typical clientele. Check specific project details via references and case studies to validate their practical expertise, ensuring their capabilities are rooted in real-world experience.

Locations and staff availability

For multinational companies, the geographic distribution and availability of specialists are crucial factors when selecting an IT service provider. Ensuring that the partner can provide on-site support when needed, as well as having the flexibility for remote collaboration, can greatly impact project success.

Cultural fit

Ensure seamless and effective collaboration by assessing the cultural alignment between your organization and a potential IT partner. Beyond technical expertise, prioritize compatibility in communication style, work ethic, and values to foster synergy and shared understanding.

ESG values

Ensuring alignment with Environmental, Social, and Governance (ESG) principles is a priority when evaluating potential IT partners, considering the increasing significance of these issues driven by both regulatory requirements and corporate reputation.

Maturity

The maturity level of a company, which is often assessed on the basis of market presence, growth trajectory, number of employees and successful projects, is an important indicator of credibility. It is also important to pay attention to the experience and expertise of employees.

Financial stability

Ensuring the financial stability of a potential IT partner is vital for mitigating risks and guaranteeing consistent, high-quality service delivery. Financially stable vendors have the necessary resources to honor contractual commitments, invest in research and development, and maintain operations without disruption. Assessing financial health through credit history checks, references from previous clients, and other financial evaluations can help ensure you choose a reliable partner.

Security and compliance

Evaluate a potential IT service provider’s approach to security and compliance. A robust security framework ensures protection against cyber threats, while adherence to compliance standards mitigates legal risks and adheres to industry-specific regulations.

Response times and support

Find out about the provider’s support structure, their response times and their approach to urgent problems. Reliable support ensures that any issues are resolved quickly to keep your IT systems running smoothly and minimize downtime.

Commercial terms

When selecting an IT service provider, understanding the commercial terms is crucial for managing costs and ensuring a transparent, mutually beneficial relationship. This involves scrutinizing costs, billing methods, and contract terms. A clear and fair contract can prevent misunderstandings and protect both parties’ interests, while flexible billing methods and transparent cost structures ensure you get value for your investment.

Evaluation

Assigning priorities to requirements

Once you have a comprehensive list of requirements, prioritize them based on their importance to your organization.

Assign weights to all requirements according to their importance to your organization and its objectives. This helps in creating a balanced evaluation framework.

Scoring potential partners

Use a simple scoring scale to evaluate how well the company meets each requirement, for example, ranging from 1 (low) to 5 (high), so you can easily identify the best and worst outliers.

If you are unsure about an evaluation at any point, hold off and try to obtain more of the required information and data.

It is very important that you evaluate each supplier based on the same criteria, otherwise comparisons will not be equal.

Create a short list of partners and re-evaluate them before making a final selection.

During the course of the evaluation, you may want to add, change or remove some requirements. You may also want to add entirely new companies to your list. This is perfectly fine if it is to help you achieve your goal of finally selecting the one or two best companies, which you will further test and evaluate more rigorously before making a final decision.

Only after comparing all the vendors and determining who can best meet your organization’s needs can you choose an IT partner and negotiate final terms.

Stakeholder alignment

To ensure everyone is on the same page, review the prioritized requirements and their category weighting with all stakeholders before starting the vendor evaluations. This alignment ensures that the evaluation process reflects the collective priorities and goals of your organization.

Conclusion