Voice AI Interviewer
Quick definition
A voice AI interviewer is software that conducts hiring interviews via real-time spoken conversation, using STT + LLM + TTS in a low-latency loop. It includes turn-taking and barge-in detection so the dialogue feels natural.
How it works
The core loop is the same as an agentic interview: STT → LLM → TTS. The voice version adds three layers. Turn-taking detection predicts in real time whether the candidate has finished a sentence — without it, the system either interrupts or leaves long pauses. Echo cancellation prevents speaker output from being re-ingested by the microphone. Accent robustness comes from STT models fine-tuned on multilingual and accented data. Together these layers make the interaction feel close to a phone screen, which is materially different from a typed flow.
Why it matters
Completion rates are higher than typed or video formats — talking is less friction for most candidates than typing. Signal quality is close to a human phone screen, since pauses, hesitation, and natural disfluency reach the assessment. Multilingual support is comparatively cheap: when the TTS model adds a language, so does the platform — no extra script authoring needed.
Related terms
- Agentic Interview — the decision loop a voice AI interviewer runs underneath.
- Structured Interview — voice AI interviewers are a scalable form of structured interview.
- Behavioral Rating Scales (BARS)
- Voice AI Interviewer (product)
- How AI interviews work
- EU AI Act compliance