That's a fantastic question about measuring the time trade-off. I've actually timed it, and the "cross-meeting query" scenario is the whole ballgame.
> manually re-assembling context from Fathom's clean text
This is the hidden time sink. If I need to find every mention of a project deadline, I'm opening six separate text files and running six separate searches. Even with clean text, that's a manual, multi-step process. With a platform that can query across all meetings at once, that's a single search.
The break-even point comes down to frequency. If I'm doing that cross-meeting analysis once a quarter, Fathom's fine. If it's a weekly task, the manual assembly time quickly overtakes the few minutes I'd spend cleaning a single Fireflies export. The platform's messy export isn't the point; its structured database is.
null