The summary arrives a quarter of an hour after the meeting. Seven decisions, one line each, all in the same font. Decision four reads: “Integration testing moves to after the pilot.”
Nobody in the meeting objected. It was minute fifty-five, the proposal came from the most senior person in the room, and the person responsible for testing was on leave that day. Three weeks later the pilot stops on day one, on a bug integration testing would have caught in an hour.
The summary was flawless. It recorded exactly what was agreed.
Seven lines of equal weight
A tool that summarises a meeting does one thing very well: it records. It also does one thing you can’t see by looking at it: it levels.
“Next meeting moves to Tuesday” and “testing moves to after the pilot” get one line each. The first decision changes nothing. The second changes the level of risk across the whole project. On the list they look identical and are equally easy to tick off.
I have written before about a different weakness of automatic notes: they don’t show that an action item has an owner who won’t deliver it. Here the trouble is the opposite. This decision will be carried out efficiently and on time, and that is exactly the problem.
Why nobody objected
Bad decisions are rarely made because someone dug in. They are usually made in silence, for three reasons.
The first is tiredness. Towards the end of a meeting people agree to things they would have argued about an hour earlier.
The second is the position of the person proposing. Objecting to a director in front of ten witnesses costs more than objecting to a colleague.
The third is absence. The person who will bear the consequences is often not at the table, and the others don’t know their work well enough to spot the problem.
None of these three things is in the transcript. All it contains is the word “agreed”.
Three questions for the list of decisions
In my experience five minutes after the meeting are enough, and three questions put to the list of decisions before the summary goes any further.
Which of these decisions will be hardest to undo? Most arrangements can be reversed a week later at no cost. One or two close a road. Those I read twice.
Who will bear the consequences, and were they in the room? If they weren’t, the decision was made without the only person who could have said “this won’t work”.
What did we give up, and did anyone say it out loud? Every decision postpones or crosses out something. If nobody named it, nobody weighed it.
After these three questions there is usually one decision left to go back to. Sometimes none, and that is good news too.
How to go back to a decision the next day
Returning to something agreed yesterday feels awkward. In practice a short message to the person who made the decision is enough:
“Yesterday we agreed that integration testing goes after the pilot. Before we start, I would like to check this with the person responsible for testing, because they weren’t in the meeting. I’ll come back with an answer by tomorrow.”
There is no undermining of anyone’s decision in that message. There is a request for one day to check. Hardly anyone refuses, and if the decision holds, everyone is more confident in it. It also leaves a trace of how the decision was made, which matters later, when someone has to tell a good decision from a bad outcome.
A question asked the next day costs one conversation. The same question asked three weeks later costs the pilot.
What the tool has to do with it
Not much, and that is the honest answer. The tool is not to blame for the list looking even. Handwritten notes had the same flaw, only fewer people read them.
You can ask the tool to flag decisions with no owner or no date, and it will do that well. What it doesn’t know is who was missing from the room, it didn’t hear that the last “fine” was said in the tone of “let’s get this over with”, and it doesn’t know the project well enough to tell a detail from a decision that closes a road.
Judging which of the seven lines deserves pushback stays with the PM. The tool gives them more time for it, because they no longer have to take notes.
One question for this week
Take the summary from your last big meeting. Which decision on that list would be the most expensive if it turned out to be wrong, and has anyone checked it since?
All situations described in this article are based on real events, but contain no company names, no individual names, and no data from any specific project.
