Recognises a decision not written back to its deliverable
The decision changes an earlier deliverable (a definition, a threshold, a decision right) and that deliverable does not cite the issue. The register and the decision drift apart.
Can AI do this?
AI does this partly Some of the tested AI models master this; others do not show it. 2 of 4 models in the current benchmark master it.
| Model | Tier | Probability of mastery | Status |
|---|---|---|---|
| gemini-4-argon | Flagship | 91% | Mastered |
| gemini-3.5-flash-lite | Low cost | 6% | Not shown |
| gpt-6-luna | Low cost | 89% | Not shown |
| claude-haiku-5-5 | Low cost | 90% | Mastered |
Mastered means a probability of at least 90% that the model has this knobit, estimated from its graded answers. “Not shown” means the answers do not settle it either way. How the models are tested.
What it looks like in practice
Expected result: The decision and the deliverable to update
How it is assessed: Flags the deficient case and passes the acceptable one. Deficient: GI-017 sets arrears completeness to 100 % for active loans; ST-006 / ST-007 in policies.yaml carry no source_issues: [GI-017] and still disagree Acceptable: The standard updated, with source_issues: [GI-017]
Also known as: Decision not written back
Related knobits
Builds on
All 20 knobits of Manage and escalate governance issues.
Source: Review checklist 05-issue-management, row 15 (DMBOK; 38505 monitor)