Products › Voice Interview
No camera. No app.
Just a conversation.
A spoken round, transcribed as it happens
A voice round any candidate can take on the phone they already have. No camera, no download, barely any bandwidth — transcribed live and scored against the same rubric as every other round.
Works on a phone, on a weak connection, in a shared room.
assets/voice-interviewer-demo.mp4 — 16:9, 1920×1080, H.264. Autoplays muted on loop; keep it under roughly 8 MB and 30 seconds.
A spoken round, transcribed live, scored against the rubric.
The camera is the
part that loses you people.
Some candidates have no quiet room, no good light and no confidence on camera — and none of that tells you whether they can do the job. A voice round removes the barrier and keeps the assessment.
- The lowest barrier of the three rounds. No camera, no video file to upload, and far less bandwidth than a recorded round — so a candidate on mobile data can actually finish it.
- For roles where speaking is the work. Sales, support, hospitality, collections, anything client-facing, and any role with a real language-fluency requirement. A voice round tests the skill rather than a proxy for it.
- A speaker-attributed transcript with timestamps. Which is what makes a spoken answer reviewable at all — read it in two minutes instead of listening for twenty, and jump to any moment.
- Tone and confidence, from the voice. Reported as one input among several, never as a verdict, and never as a reason to pass on somebody by itself.
- The same rubric as video and chat. A voice score and a video score sit on one axis, so mixing modalities across a pipeline does not break the comparison.
Transcript — customer success, round 1
A phone, five minutes, a scored round.
Five stages, and only one of them needs a person. The invitation sends itself and the transcript is written as the candidate talks.
What the Voice Interview does
Why a voice round reaches more candidates+
The transcript, and what it is actually for+
Question sets, and follow-ups that were not scripted+
Scoring, with the excerpt attached+
Tone and confidence, handled with care+
Pipelines, leaderboards and evaluation+
A recording of somebody’s
career
Audio and transcripts are the most sensitive data you will hold about a person who does not yet work for you.
A human makes the call
It ranks and evidences. It does not reject anyone. The hire, the pass and the pipeline stage are decisions a named person takes, on the record.
One AI gateway, no loose ends
Every model call passes through a single choke point with schema validation, caching, per-tenant rate limits and spend caps. There is no second path to a provider.
Media behind signed links
Audio sits in private storage reached only through short-lived signed URLs, with lifecycle rules that expire recordings on a schedule rather than keeping them indefinitely.
And plainly. A voice round still needs a microphone and a connection, which is why the check runs first — and the Chat round exists for candidates who have neither, or who assess better in writing. Tone and confidence signal is an input and should never on its own be the reason somebody is passed over; a strong accent and a weak line are the two easiest things in this field to mistake for a weak candidate. We publish no accuracy figure, because the honest measure is how often a human changed a score in your own pipeline. And there is no telephone dial-out — the round runs in a browser, not over a phone call.
Frequently asked questions
Do candidates need a camera, an app or a strong internet connection to take the interview?+
How are spoken answers scored, and can I see why a score was given?+
Does the AI judge candidates on their tone of voice or accent?+
Can I use voice interviews alongside video or chat interviews in the same hiring pipeline?+
When is a voice interview the wrong format for a candidate?+
How accurate is the AI scoring in the voice interview?+
Published 9 September 2026 · Updated 9 September 2026 · Written by the TalbotIQ team