Concept card

SAP AI Agent Hub — Vendor-Agnostic Agent Governance

As of 2026-10-10

Edited by Cédric Mary, sole editor and publication director — 27 years in the SAP ecosystem. How these pages are produced, and where automation is used.

What is SAP AI Agent Hub?

SAP AI Agent Hub, built on SAP LeanIX, is an inventory and governance command center for AI agents, MCP servers and AI models from SAP and other vendors: automated discovery, verification, business-value calculations, risk views, telemetry via SAP Cloud ALM and a self-service AI Agent Portal. It governs; it does not run or trace agents itself.

SAP AI Agent Hub is the governance layer of SAP's agent architecture. SAP Help defines it as "a vendor-agnostic command center that provides a single entry point for the discovery and governance of AI agents, large language models (LLMs), and Model Context Protocol (MCP) servers across the business landscape". At Sapphire 2026 SAP placed it as the third tier of the SAP Business AI Platform, next to the context layer and Joule Studio, "built on SAP LeanIX", to "discover, manage, and govern all AI agents, SAP and non-SAP".

The questions it answers

SAP frames the product around three questions: What agents do we have? (automated discovery into one inventory), Are they governed? (owner, risk classification, verification status), and Are they delivering value? (calculations and reports on cost savings, revenue and risk reduction). It is an architecture and governance tool. It does not execute agents, does not replace the runtime's guardrails (C162) and is not a prompt-level debugging or evaluation tool.

How it models agents

Agent Hub uses the SAP LeanIX meta model, where the unit of record is the fact sheet. Three AI-specific subtypes exist: AI agent (subtype of Application), MCP server (subtype of Interface) and AI model (subtype of IT component). They link to business capabilities, business contexts (processes), organisations and providers, so each agent can be traced to the capability it supports, the process it touches, the team that owns it and the model it runs on.

Why it matters

Agent sprawl across SAP, Microsoft, Google, AWS and ServiceNow is an audit problem before it is a productivity gain; Agent Hub is SAP's documented answer for inventory, ownership and verification.
Its business-value and AI-risk fields turn agent governance into a portfolio decision process rather than a spreadsheet.
Knowing what it does not do (execute, trace, filter) prevents selling it as a runtime control or observability tool.

Key points

Scope: AI agents, MCP servers and AI models as LeanIX fact-sheet subtypes (of Application, Interface, IT component).
Discovery: SAP agents, Microsoft Entra ID, Azure API Center, Google Agent Registry and Agent Engine, Amazon Bedrock, ServiceNow, SAP AI Core models, plus an upload API.
Governance: verification (Verified/Rejected/Unverified), AI Risk field, business-value calculations, reports and dashboard, AI Agent Portal.
Telemetry for monitored agents comes from SAP Cloud ALM; agent mining via SAP Signavio Process Intelligence.
Access: included in LeanIX Application Portfolio Management workspaces; standalone edition in closed Early Adopter Care.
Corrected launch history: Agent Hub launched inside SAP LeanIX APM on 4 November 2025 — not a May 2026 or 'Q3 2026 target' GA of a stand-alone product; Sapphire 2026 (12 May) repositioned it as a Business AI Platform pillar.
AI Governance Assistant (announced 22 September 2026) adds EU AI Act and NIST AI Risk Management Framework classification directly on agent, MCP server and model fact sheets.
The stand-alone edition's meta model is limited to seven fact sheet types: business capability, business context, organization, application, interface, IT component and provider.
Reference architecture (SAP Architecture Center, 1 Oct 2026): hub = registry and governance on SAP LeanIX; SAP Agent Gateway = runtime enforcement; SAP Cloud Identity Services = identity authority; telemetry to SAP Signavio and SAP Cloud ALM; SuccessFactors maps agents to units and roles.
SAP Connect 2026 (SAP News, 5–7 Oct 2026): SAP AI Agent Hub is generally available and positioned as governance for every agent, LLM and MCP server across SAP and non-SAP; SAP's AI governance architecture is ISO/IEC 42001-certified (SAP's own management system, not the customer's agents).

Terms used on this page

Fact sheet: Basic unit of record in SAP LeanIX and SAP AI Agent Hub, storing attributes, relationships and governance data of one entity.
AI agent discovery inbox: Queue where discovered agents, MCP servers and models are reviewed, linked to fact sheets, created or rejected.
Verification badge: Status on an AI fact sheet (Verified, Rejected, Unverified) recording a governance decision against defined criteria.
AI Agent Portal: Self-service catalogue of verified agents that business users can browse without signing in to the LeanIX workspace.
AI Governance Assistant: Capability inside Agent Hub, announced 22 September 2026, that classifies agents, MCP servers and models against the EU AI Act and the NIST AI Risk Management Framework directly on their fact sheets.
Business capability (fact sheet): LeanIX construct representing what the business can do; AI agent fact sheets link to the capability they support so agents can be traced to a business outcome.
Standalone Agent Hub meta model: The reduced object model used by the fully stand-alone Agent Hub edition, limited to seven fact sheet types rather than the full LeanIX enterprise-architecture meta model.

Sources

Full card available to members. What the full card adds: the full decision framework · the SAP vs Snowflake / Databricks / Fabric comparison · the common pitfalls and their fix · the cheat sheet · the architecture schemas · the code blocks · the facts worth quoting.

Open in the app →

SAP AI Agent Hub — Vendor-Agnostic Agent Governance

What is SAP AI Agent Hub?

The questions it answers

How it models agents

Why it matters

Key points

Terms used on this page

Sources

Related concepts

Academy modules on this

Deep studies

Watch