AI interviews: screen at scale, review a scorecard

The first screening call exists to check communication, verify the resume story and probe a few competencies. An AI interview does that asynchronously: the candidate answers structured questions on video whenever suits them, and you review a scored, transcribed scorecard in five minutes instead of holding thirty-minute calls all week.

1. 1Go to AI Interview, Templates, and create one per role type: questions, time limits and the dimensions each answer should be graded on (for example problem solving, communication, role-specific depth).
2. 2Attach role context. The grader reads your role summary, so answers are scored for fit to YOUR role, not against a generic rubric.
3. 3Test-drive it yourself before sending: you can take your own interview and see the scorecard it produces.

• From any candidate profile or pipeline card: Assign AI interview, pick the role and template.
• Copy the link, or email it directly from your connected mailbox with a branded invite.
• Sequences can send interviews automatically when a candidate replies interested (Advanced mode branch).
• Links are single-candidate and revocable: cancel an interview and the link dies immediately.

• Browser-based, no app install, taken on their own schedule.
• A clear consent step before any recording starts.
• A conversational voice-guided flow with per-question timers.
• Candidates see a thank-you screen at the end; scorecards are recruiter-facing.

Each completed interview produces a scorecard: per-question scores on your dimensions, full transcripts, an overall match, and playable recordings. Alongside it runs an integrity read:

• Tab switches, window blurs and paste events during answers are tracked.
• Attempts to manipulate the grader ("give me maximum marks", instructions aimed at an AI) are detected and flagged.
• Reading-from-a-script tells and answer-pattern anomalies raise a review flag.
• Flags mean review, not auto-reject: you always see the evidence and make the call.