Product

What a good AI interview report looks like

Kevin Minn, Founder and CEO, Motium
· 5 min read
Key points

Most interview tools end with a number. "Candidate fit: 4.2." It looks precise, and it tells you almost nothing: not what the candidate said, not what was never asked, and not whether you should believe it.

How AI interview candidates should be scored

A score is only useful if you can check it. So the test for any interview report is simple: for every judgment, can you see the evidence, the standard it was held to, and the moment in the recording? If not, you are trusting a black box with a hiring decision.

This is how Motium's report is built, and why each part is there.

There is no overall score, on purpose

An average hides the shape of a candidate. A designer who is excellent at the craft and has never managed a deadline can average out the same as someone who is fine at both. Those are different hires. Motium scores each of the position's competencies separately, from 1 to 5, and never adds them up. The position marks which competencies matter most for that job, so you read the important ones first.

Every score shows its evidence

Each competency score comes with one line of reason and the moments it rests on: the candidate's own words, linked to the second in the recording where they said them. One click and you hear it yourself. If you disagree with the score, you have what you need to disagree with.

Each score also shows the standard it was held to. A "3" means "meets requirements: gives relevant evidence they can do the work." You can open any score and see what a 3, a 4 or a 5 looks like for that competency. The full scale is on how candidates are scored.

"We did not find out" is an answer

If a competency never got a fair question, maybe time ran short or the conversation went elsewhere, the report says "insufficient evidence." It does not give a low score. A low score says the candidate is weak. Insufficient evidence says we do not know yet. Mixing those two up is how good candidates get rejected for questions they were never asked.

Must-haves are checked one by one

Most positions have a few things that are not negotiable: a license, a tool, a kind of experience. The report lists each must-have from the job description and marks how far the interview got: discussed, only mentioned, or not covered. Discussed is not the same as met. Only a person on your team can mark a must-have as verified, for example after checking a certificate.

When something is still open, you can ask the candidate for more. One button drafts a short follow-up request: a work sample, a link, or a written answer. The candidate answers on a simple page, and the answer joins their report.

A review written like a colleague would write it

At the top sits Wren's note, ready the moment the interview ends. It is short bullet points: why she would or would not recommend the next round, the main concern measured against the posting, and what to ask in the next conversation. Every point carries an example from the interview. A point she cannot back with an example is left out.

Session checks, kept in their place

The report also notes facts about the session: if the candidate left the interview page, pasted text, or if another person appeared on camera. These sit in a collapsed section with their times. They never change a score. A person decides what they mean. Why we built them this way is in spotting interview fraud without spying on candidates.

The decision is a separate step

The last thing on the report is the decision, and it is yours. The approve button names the next round, so it is clear what you are agreeing to. Every advance or decline records a reason, who made it and when. The report informs the decision. It does not make it.

Questions people ask

Why does Motium not give an overall score?

An average hides the shape of a candidate. Two very different people can average the same. Motium scores each of the position's competencies separately and never adds them up.

What does insufficient evidence mean?

The competency never got a fair question, so there is nothing to score. It is not a low score.

See a full Motium interview in under three minutes.Watch the demo

Keep reading

AI in hiring

How Motium uses AI in hiring, and where it stops

An AI hiring tool where a human makes the decision: what Motium's AI interviewer does, what it never does, what candidates are told, and the signed record.

October 31, 2026 · 5 min read
Product

Using Motium with Lever: an AI interviewer inside your Lever flow

An AI interviewer that integrates with Lever: move a candidate to your interview stage, Motium invites them, and Wren's review comes back as a note.

November 7, 2026 · 4 min read