Keep the lesson for the next incident
Classify unresolved boundaries
Group results by missing source, conflicting observations, stale context, noisy measure, insufficient cohort, policy ambiguity, or genuinely non-discriminating evidence. Keep one alternative matrix and decision consequence for each class. Distinguish a well-formed uncertainty from a result that omitted basic checks.
Keep discriminator examples
Retain compact cases showing alternatives, predictions, evidence, stable invariants, next observation, and stopping rule. Include a result that should remain inconclusive and one resolved by a small comparison. Update examples when decision criteria or source boundaries change.
Review recurrence and ownership
Assign owners for evidence collection, policy choices, and result review as needed. Sample inconclusive outputs on a defined cadence and track vague or repeated boundaries. Reopen investigation for recurring missing discriminators, unowned follow-ups, or provisional conclusions reused beyond their limit.
What to carry forward
Close the learning record with unresolved-boundary classes, discriminator examples, owners, and a trigger for vague or repeatedly untestable results. Preserve honest uncertainty as a valid outcome. The durable improvement is clearer next evidence and better interim decisions, not pressure to produce certainty.
Keep the decision with the work.
Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.
Create an account See the product workflow