HireVue's core unit is a recorded video answer. The candidate sees a question, records up to two minutes, hits submit. Two minutes per question, six to eight questions, all sequential, all one-way. AI scoring runs on the recordings; humans review the result. This format exists because in 2014 it was the fastest way to get more candidates through the pipeline. It is no longer the best one.
Picked conducts a live interview. The AI asks the question, listens to the answer, asks a follow-up that depends on what the candidate just said, probes deeper where the answer is interesting, moves on where it is not. Same time budget (about 30 minutes), substantially more signal, a candidate experience that feels like a conversation rather than a stare-into-the-lens audition.
Recorded-video AI scoring has been challenged repeatedly on bias and validity grounds. We are not piling on; the research is public, and HireVue has responded by retiring facial analysis in 2021. The structural issue with the format remains: scoring a recorded monologue against a rubric is a fundamentally different signal than scoring a live conversation. We are building on the second.