AI Transcription Did Not Change How I Take Notes. It Changed How I Listen.
The transcript was never the point. What changed when I stopped writing during conversations was more surprising than any feature.
I adopted transcription tools for the obvious reason: I wanted records of my meetings. A year in, the records are rarely opened, and yet the tool has genuinely changed my work. What it changed was not note-taking. It was attention.
Before transcription, meetings had a hidden second task running behind the conversation: me, half-listening, composing the summary in my head, deciding what was worth writing down. That filtering process felt productive and was actually a tax on comprehension. You cannot fully hear someone while simultaneously editing them.
With the transcript guaranteed, I stopped composing and started listening, and the difference showed up in the questions I asked. Follow-ups got more specific because I had actually heard the specifics. The person across from me noticed before I did; 'you actually seem present now' was not a compliment I expected from a piece of software's side effect.
The honest caveat: this only works if you trust the capture. The month I spent verifying transcription accuracy was the month the habit formed, and it was worth it. If you secretly doubt the record, you will keep a mental backup, which defeats the entire mechanism. Verify once, thoroughly, then let go.
The broader point is one I keep relearning about AI tools: the advertised benefit is often the small one. Transcription promised archives and delivered attention. The best features in my stack are all like this, quietly changing behavior in a room adjacent to the one on the label.