Every Picked candidate has the right to request human review of a non-advancement decision. The right is surfaced to the candidate at the point of decision, not buried in a privacy policy.
How it works.
- The candidate clicks Request human review on the decision note. The window is seven days from the original decision.
- A human reviewer (a member of the hiring-company side, not Picked) is notified via email and Slack. The reviewer is the role owner by default, configurable per workspace.
- The reviewer reads the full transcript, the score sheet, and the original rationale. They have ten working days to respond.
- The response is structured: either an override of the original decision, or a confirmation with a brief explanation. The candidate receives the response in the same dashboard surface.
Why this is in the product, not the policy.
EU AI Act Article 14 (human oversight) and GDPR Article 22 (right not to be subject to a solely automated decision) both require this. We build it in rather than gating it behind a written-and-mailed request because that is the only way the right is practically exercised. About 6% of non-advanced candidates request review in beta; about 12% of those reviews change the decision.
Customers in regulated industries (financial services, healthcare) can route human-review responses through a compliance reviewer before they return to the candidate; configure this on Settings, Compliance, Human review.
What to do next: confirm the human-review reviewer assignment in the workspace settings before any role goes live.