Afterward I thought about the version of that meeting where my notes were private. I would have walked out with a wrong technical understanding, made decisions on it, and probably briefed other people on it. The error would have traveled, getting further from the source and more expensive to fix with every step. The only difference between those two outcomes was a shared screen.
Most note-taking records what was said. What it doesn't show anyone is how it was understood, and that's where the expensive mistakes live. An AI transcript doesn't solve this. It captures every word, but two people reading the same transcript of a complex technical discussion will often come away with different understandings. The interpretation still happens; it just happens after the meeting, when the people who could correct it are no longer in the room.
Making the notes visible is half of it. The other half is telling people they're invited to correct you, because the social cost of interrupting the facilitator can feel higher than the cost of letting an error slide. The first time someone takes you up on that invitation, and you thank them and fix the note in front of everyone, the next correction gets easier. Over time I found it built something I hadn't set out to build. When the group can watch the record take shape and change it, the record belongs to them, and that builds trust. This week's video covers what I say at the start of a meeting and how to try this without changing your tools.
As always, I'd enjoy hearing what gets corrected the first time you try it. Hit reply.