AI Interviews Transform User Research
Joe Kim, CEO of Office Hours, on the end of crowdwork
AI interviews are turning user research from a slow, manual service into software that can run continuously. Instead of writing a rigid survey or scheduling dozens of calls, a product team can send an AI interviewer to a targeted group of buyers, get spoken or typed answers with follow up questions, and receive a synthesized readout. That expands the market because it makes nuanced research cheap enough and fast enough to use far more often.
-
The practical shift is from static forms to adaptive conversations. Older B2B research often meant multiple choice surveys sent to a panel. AI interviews let the system dig into an answer, ask conditional follow ups, and capture the kind of detail teams use for pricing, messaging, feature design, and competitor readthroughs.
-
For Office Hours, the wedge is not the AI voice itself, because that layer is already easy to buy through APIs. The harder part is supplying the right people to interview, proving they actually fit the target persona, and making the experience flexible enough that experts can start, pause, and resume asynchronously.
-
This same workflow is showing up outside user research. Mercor uses AI video interviews and tests to screen large pools of doctors, lawyers, engineers, and other specialists for AI labs. That is the same core pattern, using software to run high volume interviews while reserving value for expert matching and credentialing.
The next step is that interviews, surveys, expert calls, and transcript search increasingly merge into one stack. Platforms that own expert supply, identity, scheduling, compliance, and synthesis will keep expanding, while the raw AI interviewer becomes a standard feature that every research and talent platform offers.