Post-incident learning review
After a near-miss or incident (non-safeguarding), run a structured, blame-free learning review and produce follow-up actions.
When to use this project
Use this after a safeguarding concern, near-miss, complaint, or operational incident — once the immediate response is complete and it's time to learn, not react.
How to use AI in this project
AI structures a fair review: it reconstructs a timeline from your notes, applies a "5 Whys" systems lens (people-blaming is off-limits), and drafts actions with owners. Paste in an anonymised incident summary — never real names — and keep any safeguarding statutory reporting off the tool entirely.
AI tips for this project
- Never paste identifiable details about children, staff, or reporters into the AI.
- Focus the 5 Whys on system causes — training, rota gaps, unclear policy — not individuals.
- Cross-check the AI's timeline against your own notes before agreeing to it.
- Statutory reporting (LADO, police, regulator) is human-only. AI never touches it.
The 4 steps
- 01
Reconstruct the timeline
Facts first, no interpretation yet.
How to use AI here · Paste your anonymised incident log entries in time order. Ask AI to reconstruct a clean timeline with times, actions taken, decisions made. Cross-check every entry against your notes before agreeing.- No individuals named
- Gaps are marked, not filled in with guesses
- 02
5 Whys, systems focus
Move from symptom to system.
How to use AI here · Ask AI to run 5 Whys against the timeline, focused on system factors (training, rota, comms, policy clarity) — with an explicit instruction to avoid blaming named individuals.- Analysis stays with systems, not people
- Interventions are testable, not aspirational
- 03
Actions and owners
What we will actually do.
How to use AI here · Turn each root cause into 1–2 actions. Ask AI for owner (role, not name), a realistic due date, and how you'll know it's done. Reject vague verbs like "review" or "consider".- Owners are roles, not overloaded individuals
- Definition of done is observable
- 04
Team comms
Close the loop with the people involved.
How to use AI here · Draft a short team message: what happened at high level, what we're changing, what we're NOT changing (and why), and how people can raise concerns. Read it as if you were the most affected team member.- Tone is caring, not corporate
- No individual is singled out