SAP Build Bridges the Gap to External AI Assistants with New MCP Servers

SAP has officially expanded the horizon for its developer ecosystem with the release of Model Context Protocol (MCP) servers for SAP Build, a milestone designed to integrate SAP-specific development artifacts directly with popular third-party AI coding assistants. According to SAP release highlights, this release marks a strategic pivot from a purely walled-garden approach, allowing developers to leverage agentic capabilities in tools like GitHub Copilot and Cursor while maintaining context awareness of their SAP Cloud Application Programming Model (CAP), SAPUI5, and SAP Fiori projects.

The introduction of these MCP servers addresses a longstanding friction point: general-purpose AI assistants previously lacked deep knowledge of SAP’s proprietary data models and UI frameworks. By exposing project metadata via the standard Model Context Protocol, SAP now enables external AI agents to read, analyze, and generate code that is syntactically and contextually correct within the SAP environment.

Scenarios and Supported Workflows

The Q3 2025 update focuses on three core pillars of the SAP development stack: CAP, SAPUI5, and SAP Fiori. For CAP projects, the MCP server supports service scaffolding and entity modeling, enabling AI assistants to understand the nuances of Core Data Services (CDS) definitions. In SAP Fiori, the `@sap-ux/fiori-mcp-server` package allows agents to generate complete list report applications or add flexible column layouts using simple natural-language prompts such as “add a SAP Fiori elements list report app to my CAP project”.

Early evaluations highlight specific productivity gains in modernization tasks. For SAPUI5, the MCP tools support refactoring and linting by providing the AI with contextual documentation and migration guides, effectively serving as an intelligent pair programmer for legacy code updates. However, the integration has initial rough edges; independent tests found that while CAP and UI5 flows were robust, some Fiori generation flows encountered schema validation blocks that required manual intervention.

Integration Capabilities and Constraints

While the release opens the door to external tools like Cursor and GitHub Copilot, it also introduces new considerations for the development environment. The MCP servers are optimized for local development workflows where these third-party agents thrive, such as in Visual Studio Code. This has created a dichotomy for users of SAP’s cloud-native SAP Business Application Studio (BAS), where support for external MCP clients was initially more limited than in local setups.

Despite these environmental nuances, the standardized protocol enables broader ecosystem applications. Vendors such as K2view have already begun positioning their solutions as unified MCP servers that securely expose SAP data to AI agents, suggesting the architecture will support complex, multi-system governance beyond simple code generation.

What This Means for SAPinsiders

New MCP servers transform passive SAP architectures into active, agentic ecosystems. For technology executives, this shift means the developer toolchain is no longer a closed loop but a composable environment where AI agents can plan and execute complex tasks. Leaders should expect a reduction in routine coding cycles, but must prepare for a landscape where development increasingly involves orchestrating these autonomous agents rather than writing every line of syntax.

Real-world adoption is already accelerating, with partners demonstrating rapid time-to-value using these generative capabilities. In a recent “Hack2Build” initiative, 20 partner companies from 10 countries delivered functioning innovations using SAP Build Apps and Process Automation in just five working days, validating the drastic reduction in prototyping time. Similarly, firms like IBsolution have deployed cross-application workflow automations that integrate distinct business processes without deep programming requirements, showcasing the immediate ROI of AI-assisted builds.

As AI agents gain deeper access to system context, prioritize governance and data security when evaluating this technology. While productivity benefits are clear, the “local vs. cloud” tool limitation remains a challenge; organizations strictly bound to SAP Business Application Studio may face friction adopting external agents compared with those that allow local VS Code usage. Best practices include starting with non-critical modernization tasks, such as SAPUI5 refactoring, to validate the AI’s accuracy before allowing it to architect new core business logic.