Authority Metadata | Encoding Decision Mandate into Enterprise Knowledge | R.A.H.S.I. Framework™
An AI assistant retrieves an HR policy. The reviewer can open it and see “Confidential” and “Internal Policy.” Neither label establishes whether it governs this employee, in this jurisdiction, on this date, for this decision.
CLASSIFICATION ≠ AUTHORITY.
Consider a synthetic document tagged:
- Content Type = HR Policy
- Department = Human Resources
- Sensitivity = Confidential
- Retention = Seven Years
- Classification = Internal Policy
Those fields describe and govern aspects of the asset. They leave employee population, local scope, approval, effective period and supersession unresolved. Retrieval relevance and permission to access do not close that gap.
R.A.H.S.I. calls the missing layer Authority Metadata: an enterprise-designed representation of the mandate an asset is entitled to carry. Its conceptual dimensions include authority owner, mandate, source authority, decision class, scope, jurisdiction, context, validity, approval status, effective dates, supersession and escalation requirement. This is an architectural vocabulary, not a published Microsoft schema.
The distinctions matter.
Microsoft Purview Unified Catalog documents custom metadata for organization-specific business attributes (in preview), alongside governance domains, data products, glossary terms and access policies. Sensitivity labels classify and protect content; Microsoft says Copilot and agents recognize them. These are potential surfaces for representation and governance. They do not, by themselves, establish that a particular AI experience consumes or enforces an enterprise’s mandate assertions. That needs.
| R.A.H.S.I. Framework™
🛡️ Need implementation, not just insights?
Design HR-policy mandate controls around accountable ownership, validity and reconstructable reliance.
https://www.aakashrahsi.online/post/authority-metadata
Top comments (0)