Qualitate bets AI interviewers can replace the expert network call
Qualitate is betting that an AI interviewer can run expert network calls at a scale and price no human service can match. Kadakia argues public data is now worth zero, so investing edge comes from proprietary knowledge in experts' heads and from how fast a fund can ask questions. Voice AI now performs at human level, which lets a product replace a manual, scheduling-bound service.
Takeaways
- The AI interviewer must match or beat humans. Kadakia says voice capture reached human parity around 2022 and 2023, which made an AI interviewer possible. Qualitate started in early 2024. People associate automated calls with frustrating retail support hotlines, so a weak product would never get adopted.
- Removing scheduling is the biggest unlock. Experts can take calls whenever they want, across time zones and at weekends. This draws in experts who would otherwise not participate. The marketplace model keeps the pool growing, calls cost about a third of traditional prices, and some clients run 50 to 70 calls in two weeks.
- Channel checks become persistent, longitudinal research. Funds repeat the same panel and questions monthly or quarterly. Each client gets its own AI moderator with persistent memory, so later checks build on earlier ones. A few dozen firms have moved substantial channel check work over, some running five to 10 checks per name across 15 or 20 companies.
- AI works as a co-pilot to humans. Investors still want to hear emotion and inflection on expensive decisions. A fund does a few calls itself, then hands the rest to the AI, which suggests follow-ups for later human calls. Projects launch from a natural language request in under 60 seconds, with results in days.
- AI screening tightens fraud and compliance controls. Experts reading answers from ChatGPT or Claude is a serious industry problem. The interviewer detects AI-generated speech and ends the call. Humans review every transcript for material non-public information and personal data. Fewer than 1% contain it, and compliance teams prefer calls that are not live.
- Public data is worth zero now. Kadakia argues alpha lies in proprietary information in people's heads and in how fast a fund generates questions and launches studies. Funds sharing the tool will not converge, because each asks different questions. Yipit, by contrast, gives everyone the same data.
- Qualitate is scaling headcount fast. Backed by Insight Partners in its Series A, the company grew from about 10 people earlier this year to nearly 50, targets 60 to 70 by year-end, and plans to double again next year. Kadakia says no one has left.
Does AI research turn alpha into beta