Currently the AI Note Taker generates a full transcript but without timestamps. Adding timestamps (even at the speaker-turn level) would make transcripts significantly more useful for two key workflows:

Jumping to specific moments in a recording without scrubbing through the entire video. Right now there's no way to locate a particular topic discussed at minute 32 vs. minute 5 without watching the whole thing.

Creating SOPs and training content from recorded meetings. Timestamps let you clip exact segments for documentation without guessing where a walkthrough or decision happened.

This feels like low-hanging fruit since the audio is already being processed with timing data to generate the transcript. Surfacing that timing in the output would make the Note Taker go from "nice summary" to genuinely replacing manual note-taking.