Stamp and Repeat: Turn Every Debrief Into a Tool.

Every flight I flew ended in a debrief. What went wrong? What could we do better? What do we carry into the next flight? I have debriefed everything I have done since. It is not a talent. It is a habit, practiced until it runs on its own.

But a lesson that lives only in your memory fades. My father was a manufacturing engineer. He used to say it is easier to build a tool that makes something over and over again than to make it by hand each time. Stamp and repeat.

So every debrief should end with a tool. A checklist. A rule. A process. A question you will ask before every raise, every launch or every hire. When I debriefed Internet Venture Works, one tool that came out of it was a rule: treat easy money as a warning sign, not a compliment.

Run the debrief well

1. Do it soon, but not immediatelyGive people a few days to let the adrenaline settle. Wait much longer and memories rewrite themselves into more flattering versions.
2. Facts before opinionsBuild the timeline first: what happened, in what order, and what each person knew. Most debrief arguments are disagreements about facts.
3. Let rank go quietAsk the people with the least authority to speak first, so they are not echoing the boss.
4. End with an ownerOne person leaves the room responsible for turning the lesson into a tool, with a date when it will exist.

The evidence for the humble tool

Atul Gawande made the case in The Checklist Manifesto. In a study across eight hospitals around the world, his team’s surgical safety checklist cut deaths and complications by more than a third. Google’s engineers built the same habit into software. Their site reliability teams write blameless postmortems: find the contributing causes, assume everyone acted in good faith with what they knew, and change the system so the failure does not repeat.

This matters even more with AI. The temptation is to just use the tool and hope for the best. The leverage comes from building the framework first, the checklists and processes your experience has taught you, and then letting AI work inside it so it performs repeatably and reliably. A loss you have debriefed and turned into a tool becomes part of the machine. A loss you have only survived is just a scar.