Decide where the work belongs
Rank decision risk
For each overconfident claim, record the action it supports, affected cohort, plausible alternatives, and consequence of accepting it as true. Prioritize claims that close a decision or suppress further review. Do not use strong wording alone as proof of high operational risk.
Assess support quality
Compare evidence coverage, source freshness, alternative tests, reproducibility, and known limitations. Classify the next move as narrow wording, collect a discriminator, investigate the confidence rule, request human judgment, or monitor. Keep labels comparable only when their evidence standards match.
Set queue and trigger
Choose one correction with owner, claim, evidence cutoff, revised wording standard, and acceptance trigger. Add a reopening condition for new unsupported certainty or changed source coverage. Do not recalibrate a broad workflow from one sentence without checking comparable cases.
What to carry forward
Return a ranked confidence queue linking each claim to decision risk, evidence quality, owner, correction, and trigger. The leading item should state the truthful certainty boundary. Mark interim wording provisional when a material alternative remains untested or the standard is still being defined.
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