Human Judgment as a Service
Joe Kim, CEO of Office Hours, on the end of crowdwork
The key shift is that research buyers are starting to shop for access to human judgment, not for a specific format like a call or a survey. Once AI can run interviews, summarize answers, and route follow up questions automatically, the moat moves to who has the best pool of people, the best matching, and the lowest friction from question to answer. That is why expert networks, survey tools, and user research platforms are converging into one market.
-
Tegus showed one path for this convergence, turning one off calls into a reusable data asset. It moved beyond brokering introductions and sold a transcript library for roughly $25K per seat, because the lasting value was not the meeting itself, it was the structured insight created from many meetings.
-
Office Hours is pushing a different path, treating the core product as a marketplace for hard to reach expertise across diligence, B2B user research, mentorship, and AI training. In that model, a buyer might book a live call, launch an asynchronous AI interview, or run a targeted research project, but all three draw from the same underlying network of verified experts.
-
Prolific shows the same collapse from the survey side. Its platform serves academics, frontier labs, AI startups, and enterprise market research teams with the same matching, screening, verification, and API layer. That makes a survey less like a standalone product and more like another interface for querying a participant network.
Going forward, the winners in primary research will look less like agencies and more like liquidity layers for human insight. The strongest platforms will combine deep participant profiling, self serve workflows, AI interviewing, and reusable knowledge artifacts, then sell that stack across many budgets that used to sit in separate boxes like expert calls, surveys, and market research.