Duration and depth
Duration sets the candidate's clock. From it the platform works out how many questions fit and how deeply each skill can be examined, and it will tell you when your skill list is too long for the time you have allowed — six skills in fifteen minutes cannot be more than a sentence each, and it says so rather than producing a shallow interview quietly.
Follow-up questions
Adaptive follow-ups let the interviewer probe an answer that was interesting or incomplete. They cost time from the same budget, so a flow with follow-ups on asks fewer planned questions.
Three settings: off, allowed where there is time, or required for every main question.
Participation mode
- AI only — nobody else is in the room.
- AI with watchers — colleagues observe without the candidate seeing them.
- Hybrid — the AI opens and a human can take the floor.
- AI-assisted — a human leads with the AI alongside.
Modes with human participants create a live room and need somebody to actually join. Scheduling one and not attending leaves the candidate with the AI, which is a worse experience than having chosen AI-only.
Recording
Recording is on by default and produces the replay attached to the report. Turning it off means the transcript is all you keep.
Language
The interview's language decides what questions are asked and spoken in. It is not your interface language: a recruiter reading English can run a Tamil interview, and the questions, the spoken prompts and the candidate's chrome all follow the interview.