{"id":"lead-after-action-review-system-log-facts","audience":"operations","category":"leadership-work","track":"leadership","level":"advanced","technique":"grounding","title":"Fill an after-action review using the system log, not memories","free":false,"task":"Fill in an after-action review with exactly these labels, one per line, in plain text: Planned:, What happened:, Why different:, Next time:. Where recollections and the system log differ, use the system log. Use roles, not names. Why different states one cause supported by the log. Next time is one action with an owner role and a date from the notes. Under 100 words.","materialLabel":"Notes and system log (attached to your prompt automatically)","material":"Release of the billing update - Tuesday, Nov 4\n\nPLAN (from the release checklist): deploy at 14:00 after QA sign-off and manager approval; finish by 14:30.\n\nRECOLLECTIONS (written the next morning)\n- The release engineer: \"We deployed at 14:00 on schedule, and it broke around 14:30.\"\n- The QA lead: \"I signed off at 13:00.\"\n- The product manager: \"I think the manager approval was done.\"\n\nSYSTEM LOG\n- 13:05 QA sign-off recorded.\n- 15:40 deploy started. No manager approval entry exists before the deploy.\n- 16:05 error rate spikes above 5 percent.\n- 16:30 rollback complete.\n\nFOLLOW-UP AGREED (Nov 5): add a required approval step in the deploy tool. Owner: platform lead. Due: Nov 19.","rules":[{"type":"maxWords","max":100},{"type":"plainText"}],"mistakes":[{"title":"Using recollections","why":"The AI says it deployed at 14:00.","fix":"Say: \"Where recollections and the log differ, use the log.\"","checks":["r1"]},{"title":"Naming people","why":"The AI blames the engineer.","fix":"Say: \"Use roles, not names.\"","checks":["r2"]},{"title":"Unsupported cause","why":"The AI guesses a technical cause.","fix":"Say: \"One cause supported by the log.\"","checks":["r2"]},{"title":"Changing the template","why":"The AI adds headings or paragraphs.","fix":"Say: \"Exactly these four labels in order, one per line.\"","checks":["r0"]},{"title":"Markdown","why":"The AI bolds labels.","fix":"Say: \"Plain text.\"","checks":["plainText"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for a total word limit and it often lands over it.","fix":"Cap each part instead, for example: \"each line under 15 words.\"","checks":["maxWords"]}],"promptHints":[{"label":"Four labels","patterns":["planned:","what happened:","why different:","next time:"]},{"label":"Log over memory","patterns":["system log","recollections"]},{"label":"Roles not names","patterns":["roles","not names"]},{"label":"Next time owner and date","patterns":["owner","date"]},{"label":"Plain text","patterns":["plain text"]}],"hint":"List the four labels, say the log beats recollections, and describe what Why different and Next time must hold.","rubricLabels":["Has exactly the four labels Planned:, What happened:, Why different: and Next time: in that order, one per line","Planned is deploy at 14:00 after QA sign-off and manager approval; What happened uses the log times (deploy at 15:40, error spike at 16:05, rollback complete at 16:30) rather than the recollection of 14:00","Why different says there was no recorded manager approval before the deploy, Next time is the required approval step owned by the platform lead by Nov 19, and no person is named"]}