Moat in Reusable Interview Transcripts
Engineering leader at Tegus on building a data platform for expert interviews
This shifts the moat from finding documents to extracting judgment. Tegus was built around the idea that the durable asset was not the matchmaking service, but the transcript library created from live conversations, then linked to models, filings, and company records. As AI makes public information faster to search and summarize, the scarce thing becomes a well aimed question asked to the right person in the right context, because that is where non public color still originates.
-
Tegus monetized that scarcity through subscriptions, not call fees. Calls were priced near pass through cost, around $300 to $400, while the real revenue driver was roughly $25K per library seat. That only works if each interview becomes reusable content for many later readers.
-
The hard part was never storing transcripts, it was producing high signal ones. Tegus found that seeding a market like healthcare with generic interviews created weak output, because strong transcripts depend on a real thesis, sharp questions, and an interviewer who knows what they are trying to learn.
-
The market is converging in two directions at once. Some players are turning expert calls into one dataset inside a bigger research terminal, while others use AI to make interviewing cheaper and more frequent. AlphaSense bought Tegus for $930M, then later launched an autonomous AI interviewer on top of the transcript library.
The end state is a stack where AI handles search, summaries, tagging, and more of the interview workflow, while human advantage concentrates in question design and expert selection. That favors platforms with proprietary transcript libraries, strong compliance, and a repeatable engine for turning one good conversation into data that compounds over time.