AI in SAP LeanIX
As of 2026-09-25
What is AI in SAP LeanIX?
C124 maps SAP LeanIX's wider 2026 AI roadmap — discovery, the AI Agent Hub, governance. This card narrows to the mechanism: two distinct ways AI reaches LeanIX data, an in-workspace Enterprise Architecture Assistant grounded only in the customer's own fact sheets and connected sources, and an external MCP Server (with visual MCP Apps added June 2026) that lets Claude, Microsoft Copilot or another AI tool query and render LeanIX data without leaving that tool's own chat.
Two directions, not one AI layer
C124 covers SAP LeanIX's 2026 AI story broadly — landscape discovery feeding the inventory, GA milestones through Q1–Q2 2026, and the AI Agent Hub governing agents, LLMs and MCP servers across a landscape. This card narrows to a mechanism question that landscape-level view doesn't answer on its own: when someone asks an AI a question about the architecture, where does that conversation actually happen, and where does the data come from? SAP ships two genuinely different answers, and mixing them up is the most common source of confusion in a first LeanIX-AI conversation with a customer.
Why it matters
- A customer who asks 'could the Assistant leak data from another tenant' gets a structural answer, not a policy promise — it only ever draws on that tenant's own fact sheets, decision log and connected sources.
- A team that has used the MCP Server since 2025 may not know MCP Apps exist, and keeps pasting raw JSON into a chat when a rendered fact sheet or relationship map has been available since June 2026.
- Recommending a bespoke point-to-point integration instead of the governed MCP path reproduces exactly the three failure modes — schema drift, missing governance, no cross-platform standard — SAP built MCP to avoid.
Key points
- Two directions: an in-workspace Enterprise Architecture Assistant (SAP LeanIX itself), and an MCP Server letting external AI tools (Claude, Microsoft Copilot) reach LeanIX data.
- The Assistant is grounded exclusively in the tenant's own fact sheets, decision log and connected sources — explicitly not a generic public-trained model.
- The Assistant and its bundled Enterprise Content Research Agent were announced available on 23 June 2026.
- The MCP Server itself launched in 2025; MCP Apps (visual rendering of fact sheets, relationship maps, reports, dashboards inside chat) were added 3 June 2026.
- SAP frames MCP as solving three problems of ad hoc integrations: no standardized schema, no governance controls, no cross-platform standard.
- The AI Agent Hub (C124) is the governance layer above both directions, discovering and governing which MCP servers, LLMs and agents can touch a landscape.
- Choice of direction depends on where the human already is: inside LeanIX -> the Assistant; inside another AI chat tool -> the MCP Server.
- C124 covers the wider roadmap and governance story; this card covers the mechanism of how AI actually reaches LeanIX data.
Terms used on this page
- Enterprise Architecture Assistant
- SAP LeanIX's in-workspace chat, grounded exclusively in the tenant's own fact sheets, decision log and connected sources.
- Enterprise Content Research Agent
- A capability bundled with the Assistant, announced available 23 June 2026.
- MCP Server
- SAP LeanIX's 2025 Model Context Protocol server letting external AI tools query landscape data through a governed protocol.
- MCP Apps
- The 3 June 2026 addition rendering fact sheets, relationship maps, reports and dashboards as visual components inside a connected chat client.
- AI Agent Hub
- The governance layer (covered in C124) that discovers and governs which MCP servers, LLMs and agents may access a given landscape.
Sources
Full card available to members. What the full card adds: the full decision framework · the common pitfalls and their fix · the cheat sheet · the facts worth quoting.