Explainable AI interview coaching

The interview coach that shows its work.

Run an interview, then see the exact words behind every score. Disagree with one? Contest it and watch Criterion reconsider in the open.

Every score linked to evidence Contestable, in the open Scored on what you say, never how you look
REC 02:14Product Designer
Interviewer
Walk me through a product decision you made where the data and your instinct disagreed. What did you do?
You
We saw retention climb after the redesign, but qualitative sessions told a different story. I ran a five person study before shipping wider, and it turned out the metric was masking a drop off I would have missed if I trusted the dashboard alone.
A time I changed my mind was...The tradeoff came down to...
Type your answer
Live readProvisional
Communication0.0
Product sense0.0
Craft & rigor0.0
0.0Tracking well
reconfirmed at the end against the full transcript

How it works

Four steps. Nothing hidden at any of them.

See the rubric first

Pick a role and the exact criteria appear before a single question. Tune the weights if you want. That is the whole standard you are measured by.

Answer a real interview

Questions adapt to what you say, by voice, video, or chat. A live read scores you as you go, so there is no black box waiting at the end.

Trace every score

Each number links to the exact phrases that earned it, highlighted right in your transcript. No vague feedback, no generic advice.

The payoff

Every score traces to the words you said.

This is the real scorecard component. Open the evidence on a score and see exactly what earned it.

Product sense
Frames the problem, reasons about users and tradeoffs
8.1/10

Why this score: You led with the user problem before any solution, and named a concrete tradeoff you accepted on purpose. You lost a point by not sizing the impact.

Supports · your words“I started with what the user was actually trying to do, not the feature request.”
Supports · your words“I accepted slower onboarding so the core loop stayed simple.”
Detracts · your words“It probably helped a lot of people.” No size, no number.
Craft & rigor
The score most worth arguing with
6.8/10

Think it is wrong? Say so. Criterion re evaluates against your transcript, out loud, and the number can move up, hold, or go to human review.

Why it is different

Same answer. Same 6.2. Only one tells you why.

Most coaches hand you a number and move on. Criterion hands you the same number, then shows the exact words behind it and lets you push back.

Most AI coaches
6.2 /10

Good structure. Try to be more specific next time.

No evidence. No way to ask why.
Criterion
6.2 /10

Why this score: you named the user problem before the fix, but never sized the impact.

“I started with what the user was trying to do, not the feature request.”
Linked to your wordsChallenge it

Ready when you are

Run one interview and see your own words scored.

Contact

Try it, then tell me what breaks.

This is