Architecture visual
Shown when a public-safe visual is available for this architecture; otherwise a neutral placeholder is used instead of a broken-image icon.
Private research prototype
CapAI Agent Hub
Enterprise teams need a governed way to configure, execute and observe AI agents around SAP work.
Who it serves and what it addresses
Agent catalogue, system and service mapping, evidence setup, execution trace, workflow, telemetry, health, diagnostics, audit and controlled-action boundaries.
- Intended users
- SAP architects, platform owners, operations and governance teams
- Data dependencies
- Approved SAP service metadata, scoped operational evidence and configured model-provider access where enabled.
- Last reviewed
- 2026-07-05
How the capability connects
Agent Hub: Primary governed intelligence and execution control plane.
Engineering Control Center: Validated and evolved through engineering controls, diagnostics, testing and release evidence.
Primary governed intelligence and execution control plane. Validated and evolved through engineering controls, diagnostics, testing and release evidence.
Control and confidentiality boundaries
Read-only or controlled-action operation, explicit evidence provenance, human accountability and no publication of restricted implementation data.
Architecture-level description only. No credentials, packages, source code or private builds are public.
- Publication boundary
- Architecture-level description only. No credentials, packages, source code or private builds are public.
- Source availability
- Not publicly available
- Future direction
- Stronger real-time traceability, controlled scheduling and policy-led operations.