Expert Networks Becoming Knowledge Infrastructure
Joe Kim, CEO of Office Hours, on the end of crowdwork
The winning asset is no longer cheap labor, it is a trusted roster of specialists who can be deployed fast across many research and AI workflows. When buyers need a cardiologist, securities lawyer, or former CFO, the hard part is not finding any worker, it is proving expertise, clearing compliance, and getting that person on a call or task quickly. That makes expert aggregation a supply quality and liquidity game, not a generic marketplace game.
-
Traditional expert networks mostly optimized the demand side, with large ops teams manually sourcing experts for each project. Newer platforms are pushing harder on the supply side, building reusable pools of verified experts so repeat conversion is higher and time to match is shorter.
-
The same verified expert base can now serve several products. One hour calls for investors, user interviews for product teams, surveys, and AI post training work all draw from the same underlying inventory of people with real domain knowledge. That is why these categories are converging.
-
The closest analog is Mercor and Handshake AI, which turned credentialed talent networks into AI data businesses. Mercor was estimated at $50M ARR by end 2024, while Handshake AI reached $80M in annualized revenue by August 2025, showing how quickly expert supply can monetize once frontier labs start buying.
From here, the best expert networks are likely to look more like infrastructure for knowledge work. The platforms that win will own verification, matching, compliance, scheduling, and reusable expert data across finance, legal, healthcare, and adjacent fields, then layer multiple products on top of that same expert graph.