What happens, start to finish.

You set it up once per role. Your candidates take it when it suits them. You read what comes back.

Before: you set it up

Describe the role or paste the job ad. We draft a job description and interview questions.
Edit, reorder or delete any question. Nothing goes to a candidate until you approve it.
Choose the competencies each answer is scored against.
Invite one applicant, or upload a CSV of everyone who applied.

One link, by email or however you usually reach them. It opens in their phone's browser. No app, no account, no password.
A short welcome that says what's about to happen and roughly how long it takes.
An interviewer that speaks each question aloud and listens to the answer.
A follow-up question when an answer is thin or interesting, built from what they just said.
If an answer runs long, a gentle nudge to wrap up, so the interview keeps to time.
They choose when: before work, after a shift, on a Sunday night.

Example interview

Interviewer
Tell me about a time a shift got busier than you expected.
Candidate
Saturday at the café, two people called in sick, so I ran the coffee machine and the register at once.
Interviewer
Running both at once is a lot. When the queue built up, how did you decide what to do first?

The second question came from her answer. That's the difference.

A full transcript of every answer.
The recording, which your team can replay.
A score against each competency, with the reasoning.
A model recommendation, labelled as AI output for context, not a hiring decision.
Export to CSV.