SAP and Microsoft: Why AI requires a new approach to integration

AI is changing how employees access business data and processes. For organizations, this raises a fundamental question: how should SAP and Microsoft work together, both technically and organizationally?

Whether responding to a customer inquiry, approving an order, or reviewing KPIs, employees often move between two worlds. Business data and processes reside in SAP, while day-to-day work happens in Microsoft 365, Teams, Outlook, or Power BI. Copilot and other AI assistants are changing how employees access this information. SAP data and process information can be surfaced directly within the task at hand, combined with other information, and placed in the relevant working context.

As a result, AI is increasingly becoming an interaction layer across traditional system boundaries. For employees, where information resides becomes less important. What matters is whether it is available at the right moment and with the business context needed to use it effectively.

A technical connection between SAP and Microsoft alone is not enough to support these scenarios. Data, process logic, identities, and permissions need to work consistently across systems so that AI can use them reliably. This becomes particularly important when an AI assistant prepares actions or supports individual process steps. An approval, status change, or update should only take place when the relevant business context, process logic, and access rights are correctly applied.

Integration becomes an architecture question

Connecting SAP and Microsoft is nothing new. For years, the focus has largely been on transferring data between systems, connecting applications, and synchronizing identities. AI is shifting that focus. When employees need to access SAP context from Microsoft 365, or AI applications need to bring together information from different enterprise systems, integration becomes a broader architecture question: How should data, processes, and access be organized across both platforms?

At the same time, many SAP landscapes are already approaching major modernization decisions. Mainstream Maintenance for SAP Process Integration and SAP Process Orchestration 7.5 ends at the end of 2027, with Extended Maintenance available until the end of 2030. SAP Identity Management 8.0 also reaches the end of Mainstream Maintenance at the end of 2027, with an Extended Maintenance option through 2030.

These changes provide an opportunity to align existing integration and identity architectures with new ways of working and emerging AI scenarios. Organizations need to determine which SAP data and processes should be available across systems, how identities should be mapped, and how business context and governance can be maintained.

The central architecture question is therefore: How can organizations make SAP context available so that both people and AI applications can use it securely and effectively? Connecting the systems technically is only one part of the answer.

AI needs business context

What matters most is what remains intact as information moves between systems. For AI, simply making SAP information available is not enough. Its business meaning must also be preserved outside the source system.

Even a basic order status illustrates the point. Interpreting it requires context about the specific order, delivery dates, responsibilities, approvals, and other process information. The user's permissions also need to be taken into account. An AI assistant must understand which information that person is authorized to access and which actions they are permitted to perform within a process. Only when these relationships remain intact can AI interpret SAP data correctly and use it meaningfully outside the source system.

Bringing SAP into the flow of work

In practice, this could mean a sales employee checking the status of a customer order directly from their Microsoft work environment. A manager might access SAP KPIs there to inform an upcoming decision. Pending approvals or other process steps can appear where the responsible person is already working.

The integration of SAP Joule and Microsoft 365 Copilot illustrates what this can look like. Users can access SAP context from Microsoft 365 Copilot and Microsoft Teams, while Joule can incorporate information from Microsoft 365. This brings SAP information together with the user's current work context in one place.

The value extends beyond access to information. A query can lead directly to the next step in a process. Employees can continue working without switching applications, while the underlying business logic remains anchored in the SAP process.

This is also where integration requirements increase. Once AI begins preparing or initiating actions, identity, permissions, process logic, and governance need to work together reliably.

Connecting data, identity, and AI

SAP and Microsoft are also becoming more closely connected at the data layer. SAP Business Data Cloud Connect for Microsoft Fabric is designed to make SAP data and its business context available in Microsoft Fabric without creating additional copies of the data. Its bidirectional zero-copy approach provides a foundation for combining SAP data with other enterprise data and using it for analytics and AI scenarios.

In parallel, Microsoft is developing the semantic layer for AI. Fabric IQ is designed to enable AI to work not only with individual data values, but also with their meaning and relationships. This involves connecting business entities, their relationships, rules, and metrics with the underlying enterprise data. Work IQ adds context from everyday work, including collaboration and workflows.

For organizations, the key implication is that AI can increasingly access data together with its business context and connect information across different systems.

Data access, identities, and permissions therefore need to be addressed consistently across the platforms involved. Organizations should define early on which information is available to which roles, how access is governed, and which actions can be performed from a Microsoft or AI environment.

With SAP Identity Management approaching end of maintenance, organizations also need to decide how to manage identity lifecycle management and access rights across systems in the future. Microsoft Entra is one relevant option: SAP and Microsoft support the migration of identity management scenarios from SAP IdM to Microsoft Entra. SAP Cloud Identity Services remain part of the SAP identity landscape and can work alongside Microsoft Entra.

This leads to an architectural principle that also aligns with the clean core approach: processes, data, and business rules anchored in SAP should remain in the core system wherever possible. Standardized interfaces and platform services can then provide the required context to Microsoft 365, Fabric, and AI applications. This allows new work and AI scenarios to be built outside the SAP core without unnecessarily moving the business logic embedded within it.

Start with specific use cases

The right architecture becomes clearer when organizations start with specific workflows. Particularly relevant are processes where moving between SAP and Microsoft already creates additional effort, or where AI applications lack the business context they need.

Typical starting points include:

  • Frequent switching between SAP and Microsoft applications
  • Manual exports and further processing of SAP data
  • Approvals and decisions that require switching between systems or channels
  • Recurring requests for information from SAP
  • AI applications that lack relevant SAP information or process context

Organizations can assess these use cases based on business value, technical feasibility, security requirements, and implementation effort. This helps identify where closer integration can deliver the greatest value and which technical foundations are required.

The prioritized use cases can then serve as the basis for developing a target architecture for SAP and Microsoft. The starting point is how people need to work: How should employees work in the future? What business context will they need? And which technical and organizational foundations need to be in place?

From switching systems to a connected work context

AI creates new opportunities to connect SAP and Microsoft more closely within everyday work. Organizations should use upcoming modernization initiatives to align their integration landscape with evolving ways of working and the AI scenarios they expect to support in the years ahead.

Specific use cases with clear business value provide the most effective starting point. They reveal where closer integration can add value and which foundations need to be established.

Campana & Schott addresses these questions through its SAP × Microsoft Discovery Workshop. Together with the client, the team identifies specific points of friction in day-to-day work, assesses relevant AI opportunities, develops and prioritizes use cases, and defines a target technical architecture. This provides the basis for cost estimates, the business case, and a roadmap.

More information about the SAP × Microsoft Discovery Workshop is available here:

SAP-Microsoft-Integration