Glossary
Governance Engine
The policy layer that controls and audits every action an agent or connector takes across the platform.
When an AI agent or an integration creates an invoice, updates a purchase order, or triggers a payment, an organization needs to know who authorized it, what policy permitted it, and exactly what happened. The Governance Engine is what makes autonomous integration safe for real business operations.
It enforces policies expressed across four dimensions — identity, resource (UDM Noun), action (UDM Verb), and condition — and applies them uniformly to every channel: direct connector calls, MCP tool invocations, and A2A delegations. Prohibited actions are blocked before execution, and actions above a configured threshold can require human approval.
Every action produces an immutable, tamper-evident audit record, enabling compliance reporting against frameworks such as SOC 2 and GDPR.