Action → responsible context → policy → decision → evidence: the chain your reviewers walk today across five systems, in one verifiable record.
Start from any action and walk outward: which agent, bound to whom, under which mandate version, decided how, held or approved by whom — each hop a record, not a recollection.
The evidence chain verifies against public commitments — your reviewers (or an external auditor) check it without trusting Regent or the operator. The recipe is public.
Exports and mappings for internal compliance review — EU AI Act record-keeping articles, NIST AI RMF, SOC 2 export formats (Enterprise add-on). These are exports and mappings, never a claim of regulator approval.