Separate verified facts from guesswork in incident reports

Sort timeline events in post-incident logs into proven evidence versus speculative blame so your team addresses real root causes.

You getVerified Incident Timeline and speculative Assumptions and Blind Spots

Prompt

Prompt
Act as an impartial root-cause incident auditor. Your job is to separate verifiable, timestamped facts from human assumptions, speculation, and retrospective blame in the provided incident notes.

Rules:
- Categorize every major claim as either a Verified Fact, an Unverified Assumption, or an Attributed Blame statement.
- Verified facts require concrete, measurable logs, alerts, or audit trails.
- Flag whenever a human error is blamed instead of a system or process gap.
- Keep observations neutral and objective.

Output format:
1. Verified Incident Timeline (bullet list with timestamps, actions, and verified data)
2. Speculative Assumptions and Blind Spots (numbered list detailing unverified assertions)
3. Questions to Verify with System Telemetry

Here is the incident report:

Example output

Example for: production database outage during a scheduled migration

1. Verified Incident Timeline
  • 14:02 UTC: Migration script migrate_v4_users.sh was executed by deploy-runner-02 (confirmed in deployment pipeline logs).
  • 14:07 UTC: P99 database response latency spiked from 18ms to 4,200ms (confirmed in Datadog alert #4921).
  • 14:11 UTC: Connection pool saturation reached 100% on the primary read replica (system metrics log).
  • 14:24 UTC: Read traffic was manually rerouted to fallback cluster; error rates dropped below 0.1% (traffic routing audit trail).
2. Speculative Assumptions and Blind Spots
  1. Claim: The migration script was untested. (Unverified assumption: no staging logs were checked before concluding the script lacked review).
  2. Claim: The on-call engineer panicked and delayed failover. (Attributed blame: ignores whether alerting dashboards clearly surfaced the connection pool exhaustion).
  3. Claim: Network congestion from the cloud provider contributed to the delay. (Unverified speculation: provider status dashboard reported 100% normal uptime).
3. Questions to Verify with System Telemetry
  • Was the staging database sized to match production row locks during the migration test run?
  • Does the failover playbook contain automated health checks, or did it require manual database restart approval?
Written by AI from this prompt. Your results will vary.

Paste raw postmortem notes or Slack incident room transcripts directly into Claude, Gemini, or ChatGPT. Follow up by asking what automated monitoring could have detected the incident earlier.

Make it repeatable

Keep using this prompt without pasting it in every time.

Keep exploring

More in Critical Thinking & Decisions
Critical Thinking & Decisions

Separate luck from skill in past decisions

Break down past wins or losses to see what was sound judgment versus random chance. Use this during retrospectives to avoid rewarding reckless risks or penalizing disciplined bets.