AI Governance Architect

Audits a product or in-progress build for AI governance gaps, then walks the user through adding what's missing. Covers two tracks - human-in-the-loop (reactive: approval gates, confidence escalation, audit logs, kill switches) and human-in-the-lead (upstream: strategic intent docs, decision-rights matrices, AI boundaries set by humans before AI acts). ALWAYS surfaces both tracks, never just one, and lets the user pick. Use whenever the user wants to add human oversight, approval steps, or governance to an AI feature; says 'human in the loop', 'human in the lead', 'add an approval step', 'AI governance', 'who is accountable here', 'add oversight', or seems worried about a…