Assessment (New AI, Existing AI, Architectural Context)
The full CALIBRE + ARA + Secure-RAG pipeline that turns a scenario into a scored initiative.
What it is
Governance Fabric's assessment pipeline: ARA produces a Scenario Brief and risk register, CALIBRE scores five factors and applies amplifiers, Secure-RAG resolves regulatory evidence, and the Decision Engine synthesises a verdict.
When to use
- New AI idea: pre-approval scoring for any AI project proposal.
- Existing AI product: reverse-engineer governance from a system already in production.
- Architectural context (/initiative/{id}/context): declare Architecture-Sensitive Amplifiers (Agentic, Physical, Cryptographic, Ledger, Biometric) that automatic inference cannot detect.
Walk-through
- 1Sidebar: New assessment.
- 2Step 1, Basics: name, owner, sponsor, business unit and description.
- 3Step 2, Risk profile: choose quick scoring or the full assessment, and tick any architecture amplifiers.
- 4Step 3, Review, then create. The initiative page opens with GII, band, findings, actions and decision.
- 5For an existing initiative: from its detail page, open More > Architectural context to add or edit amplifier flags at any time.
Common pitfalls
- Vague scenario descriptions produce vague ARA output. Aim for 3-6 sentences with concrete data types and affected parties.
- Skipping amplifiers when they clearly apply. A declared amplifier raises the band by one level until its sub-scores show no sub-score of 3 or more; missing it under-rates the system.