The starting point
An incident is sorted or an attempt has failed. The messages exist, but the reasoning and the lessons are scattered across them.
The move
Write a short account within the week: what happened, what we understood, what we are changing.
Why it works
With no record, the same incident comes back and the team rediscovers it. With one page, the next person knows what was tried, what worked and what was changed.
A file stalled between two teams because each was waiting for a different confirmation. The account describes what each side saw, how the blockage was understood, and which handover agreement changed. Nobody is named as the culprit. The mechanism is fixed.
An AI can rebuild the timeline from the messages and requests in minutes. Keep your time for what it does not know: why people believed what they believed, and what is changing.
Share it with the people who need it: the team, whoever takes over, the teams next door. When the story can help others, publish it, leaving out what is not yours to share. Much of what people know about incidents comes from accounts other teams published.
Try this
Choose a recent incident or failure. Write one page:
What happened and what is still uncertain: …
What we thought at the time and the checks we made: …
The actions, their effects, and the limits we met: …
What we are changing, who handles it, and how we check it: …
Have the facts read by the people concerned. At the agreed moment, check that the action decided was carried out.
To discuss
Which recent incident left no written account?