The AI Tutor will answer grounded in this lesson's cited source (portal_guide: PRD-102) and key concepts — authority Informational. Not yet enabled.
Governance Copilot
What it is. The explainable governance answer engine: ask it about an AI system and it maps telemetry → risk → control → obligation → framework, with a confidence score, a plain-language explanation, and a signed audit record of the answer itself.
Who it's for. Anyone who has to answer "what does this AI system oblige us to do?" — governance analysts first, but also executives who want the board-pack view.
Features and what they mean
| Feature | What it means |
|---|---|
| Deterministic decision engine | No LLM decides. Algorithms decide, language only explains — the same question always gives the same, reproducible answer. For regulated buyers that is a feature, not a limitation. |
| EU AI Act classification | Describe an AI system and get its risk tier with the article trail behind it. |
| Signed audit record | Every answer writes its own hash-chained audit record — you can prove what the Copilot said and when. |
| Always-verified answers | The whole answer chain is cryptographically re-verified before every single answer; if verification ever fails, the Copilot suspends itself rather than keep talking. The live chain status is shown at the top of the page. |
| Governance query | Ranked search across your governance posture — "ask your governance". |
What you put in, what you get out
| You put in | You get out |
|---|---|
| A description of an AI system (purpose, data, autonomy) | Risk classification with obligations and the reasoning trail |
| A governance question | A ranked, explained answer with confidence |
| Live telemetry (when connected) | Obligation mapping that updates with reality |
Your first session
- Open Governance Copilot from your workspace.
- Ask it to classify one AI system you know well — judge the answer against your own expectation.
- Open the explanation and the audit record behind it — the traceability is the product.
Honest limits
Answers are advisory — the platform's decision records stay authoritative, and the Copilot never executes changes. It explains; you decide. It refuses legal and liability opinions, quotes article references rather than regulation text, and states no compliance position for frameworks with no mapped controls.
Access: bundled with the Governance Platform (a dedicated GOVERNANCE_COPILOT
entitlement also grants it); in your workspace at /governance-copilot.