AI Interview Assistant
Every round scored. Including the human ones.
Your AI rounds already produce structured, comparable evidence. Then the candidate reaches the round that decides the hire — and the record goes blank. The Interview Assistant joins your human rounds, visibly, and closes that gap: recording, transcript, and a cited evaluation on the same scale as everything else.
Final round · Priya N. — Account Executive
Interviewer: VP Sales · 41:32
Click a transcript line — the recording jumps to that moment.
Transcript
AI evaluation
- Objection handling9/10
- Discovery8/10
- Forecast rigor8/10
Confirms Quota and territory match her AI phone screen.
Follow up Ask for the win/loss split behind the 14 deals.
Advance Strong close; verify pipeline mix next round.
Trusted by HR teams at 200+ companies
The problem
The round that decides the hire is the one with no record.
Screens, assessments, and AI interviews all leave structured evidence behind. The final human round — the one with the most weight — usually leaves a calendar entry and a hazy memory.
Feedback arrives late, or never
Chasing panelists for written feedback is a job in itself. Most of it lands days later, thinner than the interview deserved — and some of it never lands at all.
Notes replace attention
The interviewer spends the call typing instead of listening — and still walks out with half a record of what was actually said.
Decisions rest on memory
Debriefs get settled by whoever remembers loudest. There is no transcript to check, so nobody checks. The most important round produces the least evidence.
How it works
One toggle. That’s the setup.
01
Schedule
Book the interview from Goodfit — it lands on your own calendar with the meeting link. Flip one toggle to record. That is the whole setup.
02
It joins, visibly
"GoodFit Interview Assistant" appears in the Meet or Teams call as a named participant. It listens and records. It never speaks.
03
Evidence lands
Recording, speaker-tagged transcript, and a fully cited AI evaluation — on the candidate’s page in minutes, next to every other round they’ve done.
01 · The recording
It joins the call like any other participant.
No browser extension, no “invite this email to your meeting” ritual, and nothing covert. You flip a toggle when you schedule; the assistant does the rest.
Visible by design: it joins the participant list under its own name. Candidates and panelists always know it’s there — nothing about the recording is covert.
One toggle per interview: opt in when you schedule, from the same flow that books the calendar event. Works with Google Meet and Microsoft Teams links.
Reschedules handled: move the interview and the assistant moves with it — cancelled and re-dispatched automatically, with a hard guarantee of at most one assistant per interview.
Live status, always: watch it move from Scheduled to Joining to Recording to Completed on the candidate’s page, so you never wonder whether the round was captured.
Final round · Interview Assistant live status
02 · The transcript
Every word, tagged and timestamped.
The interview becomes a record you can search, quote, and check — not a summary you have to take on faith.
Transcript candidate identified · high confidence
Click a line → the recording jumps to that second.
Speaker-tagged and timestamped: every turn carries who said it and when — a record you can quote, not a summary you have to trust.
Click to the moment: select any transcript line and the recording jumps to that exact second. Settle any debrief disagreement in one click.
Knows who the candidate is: the candidate’s voice is identified automatically — with an explicit confidence level, and a banner when it isn’t sure. It never silently guesses.
The full recording, kept: audio is stored alongside the transcript, so any claim in any evaluation can be checked against what was actually said.
03 · The evaluation
A structured evaluation, not a meeting summary.
Minutes after the call ends, the candidate’s page carries the same quality of evidence a Goodfit AI round produces — scored, cited, and ready for the debrief.
Scored like every other round: a per-skill scorecard on the same 0–10 scale recruiters use across Goodfit — so the human round finally compares cleanly with assessments and AI interviews.
Every score cites its source: justifications reference the specific transcript turns they came from. No black box — click through and listen to the moment behind the number.
Red flags surfaced, with receipts: concerns and red flags each carry their own citations, so nothing serious hides in minute 43 of a recording nobody rewatches.
Briefed on the role: the evaluation reads the candidate’s resume and the job’s skills first — it judges against this role, not a generic rubric.
Sets up the next round: a hire recommendation with written reasoning, plus suggested follow-up questions so the next interviewer starts where this round left off.
AI evaluation sample
Red flag Q2 numbers not broken down when asked twice — cites t2, t9.
Recommend Advance — strong objection handling; verify pipeline mix.
04 · The consistency check
Did their story hold up between rounds?
Every evaluation cross-references what the candidate told the AI interview weeks earlier — and returns a verdict: confirms, contradicts, or mixed. A standalone notetaker can’t do this, because it never ran the earlier round.
ConfirmsQuota attainment and territory match what she told the AI phone screen three weeks ago.
MixedDescribed her role as “leading the pod” here; the screen recorded an individual-contributor scope. Worth one question.
05 · The panel
Humans stay in charge. On purpose.
The assistant does the paperwork; people make the call. Two design decisions keep it that way.
Blind by design: the feedback form deliberately hides the AI’s verdict until the panelist has submitted their own — human judgment is never anchored by the machine’s.
No account needed: every interviewer — including external panelists who’ve never touched Goodfit — gets a personal link to submit their verdict, skill ratings, and notes.
One skill list, three views: panel ratings, the AI evaluation, and the recruiter’s scorecard all rate the same skills — so they line up side by side instead of talking past each other.
One-click scorecard draft: take the AI’s evaluation as a pre-filled draft of your own scorecard, then edit and submit it as yours. Ten minutes of admin becomes one click plus judgment.
Your verdict on Priya N.
Submitted before the AI’s evaluation is shown to you.
Why it’s different
Citations, not vibes.
Notetakers summarize meetings. Goodfit runs an evaluation.
Every score is sourced
Skill scores cite the exact transcript turns they came from. Click a citation, land on that second of the recording.
Cross-round consistency
Did their story hold up since the AI screen? Only Goodfit ran that round — only Goodfit can check.
Blind panel verdicts
Panelists vote before seeing the AI’s take. External interviewers need no account — just a link.
You make the call
One click turns the AI’s evaluation into your scorecard draft. It never advances or rejects anyone.
On the record
It never speaks.
It never decides.
It’s never hidden.
The assistant is an evaluator, not a participant. It never asks a question, never interrupts, and never advances or rejects a candidate — evaluations are advisory, and every decision stays with your team.
FAQ