A copilot for the interviewer. Never for the candidate, never in their ear.

An interviewer copilot is not a cheating tool.

Search for "interview copilot" and most results are tools that whisper answers to candidates. Pyro's copilot is the opposite: it works for the interviewer, the candidate agrees to the session being transcribed, and it never feeds anyone an answer.

It is also more than a note-taker. A note-taker writes up the interview afterwards. The copilot works while the interview is happening, so the follow-up you needed arrives while you can still ask it.

Interview guide, note-taker or copilot.

Tools that help interviewers fall into three groups. They are often sold under the same name, so it is worth knowing which is which.

• An interview guide gives the interviewer a fixed list of questions and a rating scale before the call.
• A note-taker records the interview and writes notes, a summary or a scorecard draft afterwards.
• A live copilot works during the interview: it tracks what has been asked, drafts follow-ups from the last answer and shows which criteria are still open. Pyro’s copilot does that, and produces the draft scorecard afterwards too.

The human round starts where the first round ended.

Most interview intelligence tools start cold: they hear the human round with no idea what came before. In Pyro, the copilot’s questions are built from the gaps screening found, the CV, the recruiters’ notes and the questions earlier rounds already asked, including ARIA’s first round.

And it ends in the same record: the draft scorecard, the transcript and the decision sit on the same timeline as the screening and the AI interview.

What changes for the person asking the questions.

• You arrive with questions built for this candidate, not a generic list.
• You stay in the conversation instead of writing notes.
• When an answer is vague, a follow-up is waiting for you to use or ignore.
• You can see, live, which criteria you have covered and which are still open.
• You leave with a draft scorecard full of quotes, and the ratings are yours to make.

Better interviews from people who do not interview every day.

Most interviews are run by people whose main job is something else. The copilot gives every interviewer the same structure, so the candidate is judged on the criteria, not on how practised the interviewer is.

The benefits of a structured interview, without the overhead.

Structured interviews, where every candidate is asked about the same criteria and rated on the same scale, have long been preferred in hiring research because the answers can be compared. Their weakness is effort: someone has to prepare the questions, keep track during the interview and write it all up afterwards.

The copilot takes that effort away. The questions are prepared for you, coverage is tracked while you talk, and the draft scorecard is written from the transcript. You keep the judgement.

What happens after the call.

• The draft scorecard arrives with quotes and timestamps for each criterion.
• The interviewer confirms or changes each rating and adds their own notes.
• The scorecard joins the same record as screening and ARIA’s first round.
• The person who moves the candidate to the next stage is named on the record.

Built so the human stays in charge.

• The candidate agrees to the session being transcribed before it starts.
• Only the interviewer sees the copilot. It never speaks to the candidate.
• The copilot suggests questions; the interviewer decides what to ask.
• There is no AI score on the scorecard. The ratings are the interviewer's.
• Every stage move after the interview names the person who made it.

Four lines the copilot does not cross.

An interview is a conversation between people. The copilot is built to support the person asking the questions, never to take their place.

• It never speaks to the candidate, and the candidate never sees it.
• It never scores the candidate. The ratings are the interviewer’s.
• It never decides. A person moves every candidate, and is named on the record.
• It never runs without the candidate’s consent to transcription.

Help that follows the conversation.

A script breaks the moment a candidate says something unexpected. The copilot follows what is actually said: when a planned question is answered early, it is marked as covered; when an answer opens something new, a follow-up appears for it.

The interviewer stays in control of the order, the tone and the questions. The copilot just makes sure nothing that matters is missed.