GOVERNED BY DESIGN · AGENTIC GOVERNANCE PROFILE BUILDER

Build the governance file for one AI agent.

Six named primitives. One agent. Mapped to EU AI Act Article 26 deployer obligations.

Works for agents you're building and agents already running in production. Three modes: pre-deployment, in-flight build, and retrofit on a deployed agent.

The six primitives

Six. There is no seventh.

• Decision Boundary Contract — what the agent may and may not decide
• Oversight Trigger Matrix — named human intervention, calibrated to risk
• Accountability Canvas — four named owners, before deployment
• Audit Schema (Handoff Receipt) — queryable fields, retention, receipt structure
• Gap Register — what is not yet documented, dated and owned
• Suspension Authority Ledger — who can suspend, on what trigger, who is notified

What you take away

A Governance Profile covering all six primitives, exportable as PDF, and a Gap Register exportable as CSV.

How it maps to EU AI Act Article 26

Article 26 sets out what a deployer of a high-risk AI system is expected to do: use the system per the provider's instructions, assign competent human oversight, monitor operation, retain logs, and inform affected people where required. The Builder maps each of those obligations to a named primitive and gives you a structured place to record how your organisation meets it — including what it does not yet meet.

The application date for standalone Annex III high-risk systems is 2 December 2027, and 2 August 2028 for systems embedded in regulated products, following the Digital Omnibus on AI (OJ L_202601744, in force 27 July 2026). The obligations did not change. Only the date did.

This is a documentation aid, not legal advice, and it does not determine your obligations or discharge them. For decisions about your institution, involve counsel.

No account. No payment. No sales follow-up. Built by AI Streamline Hub.