Which meeting tools does the note-taker work with?

Zoom, Meet and Teams. The bot joins the scheduled external call on the invite you already sent, so you do not move your interviews anywhere.

Do the people on the call know it is being recorded?

Yes. Participants are notified that the call is being recorded, so nothing is captured silently. Notice and consent rules differ by country and by state, so check the compliance guides for where you hire before you switch it on.

2 Hire Credits each time the bot joins an external call, on Starter and up. A human-led round hosted inside Expert Hire costs 2 credits and covers the copilot, the notes and the transcript. Reports are never metered.

Can I score an interview I already ran?

Yes. Upload or pick the recording, choose the job, and you get the full report without redoing the interview. Every score comes from the transcript of the call that happened.

Is the score the hiring decision?

No. A score is never an automated pass or fail. Every score traces back to the transcript, so a human reviewer can check the claim instead of trusting the number. Score cutoffs exist as a Workflow setting on the round, and they are off until you turn them on.

How is this different from a generic meeting note-taker?

A generic note-taker gives you a loose summary. This one files a structured intake summary under fixed headings, Client & Role Context, Hiring Requirement, Ideal Candidate Profile, Must-Have Criteria, Good-to-Have Criteria and Budget, Compensation & Commercials, and marks anything that was not discussed as Not discussed rather than inventing it. On a candidate interview it scores the answers against the requirements you set on the job and the JD, then files the report on the candidate.

What does the intake summary look like?

It opens on a Summary tab, with Details beside it, and runs under the same six headings on every call. Bullets sit under each heading, and where a topic never came up the line reads Not discussed instead of a guess. The recording and a searchable transcript sit in the right rail, and clicking a transcript line jumps the recording to it. You can edit the summary before you share it.

What about conversations that were never on the calendar?

Instant Meeting rooms cover those. You open a room in one click, and the call is recorded and auto-classified into a titled report.

AI note-taker

Every live interview captured, scored and searchable.

A bot joins the calls you already booked on Zoom, Meet or Teams. It records the call and transcribes it. On a client call it files a structured intake summary under fixed headings, and marks anything that was not discussed as Not discussed instead of inventing it. On a candidate interview it scores the answers against the skills you picked for the job, the weightage you set on each one, and the job description. You get a full report instead of notes typed from memory an hour later.

Book A Demo

30-day free trial, nothing charged for 30 days. The calendar note-taker is available on Starter and up.

Zoom, Meet and Teams
Participants are notified of recording
2 credits per external call

An Expert Hire interview scheduled with automated meeting capture — Capture the conversation where it already happens.

The problem

Interview notes get written from memory, an hour later.

The interview happens. The write-up does not. What reaches the debrief is a few lines of recall, and no two interviewers write the same few lines.

Two people can interview the same candidate on the same day and file notes that cannot be compared at all.

Notes thin out as the day goes on

The first interview gets a paragraph. The fourth gets three bullets, typed between calls, from whatever is still remembered.

The candidate you saw last is the one you know least about.

Everyone captures something different

One interviewer writes down the technical answer. Another writes down an impression. The two notes have nothing in common.

The debrief runs on opinion instead of evidence.

Nothing traces back to what was said

A rating with no quote behind it cannot be checked. The call it came from ends and nothing is kept.

Nobody can explain the decision six weeks later.

A panel cannot be stacked

Five interviewers, five formats, five private scales. Building a shortlist out of that means guessing what each score meant.

The loudest write-up wins the argument.

Here is what happens instead.

How the bot joins

Four steps. None of them are yours.

Nobody changes how they run interviews. The call happens where it always happened. The report shows up on the candidate afterwards.

Point it at your calendar.

The bot works from the interviews you already schedule. It picks up scheduled external calls. Nobody has to start a recording or move a meeting somewhere new.

It joins the call you booked.

Zoom, Meet and Teams. The bot joins on the invite you already sent and stays for the whole call. The interview runs where it was always going to run.

Everyone is told it is recording.

Participants are notified that the call is being recorded. Nothing is captured silently, and nothing about the bot is hidden from the candidate.

It records, transcribes and scores.

The call is recorded and transcribed. The answers are then scored against the skills and weightage you set on the job, and against the job description. The report is ready before the interviewer closes the tab.

A human-led round hosted inside Expert Hire needs no bot. It is recorded, transcribed and scored against the same skills and JD as an AI round, with a copilot alongside the interviewer. Instant Meeting rooms cover the conversations that never made it onto a calendar.

What you get back

Two kinds of call, two kinds of report.

On a client call you get a structured intake summary: Client & Role Context, Hiring Requirement, Ideal Candidate Profile, Must-Have Criteria, Good-to-Have Criteria, and Budget, Compensation & Commercials. On a candidate interview you get a scored report against the skills you picked. Both come with the recording and a searchable transcript.

Dashboard › Meetings › Instant meeting

Stripe <> Northwind

Rachel Whitmore · 24 Jul 2026, 15:09

Completed

SummaryDetails

Summary Edit

Client & Role Context

Client: Series B logistics platform, 240 people, hiring into the platform team
No reporting manager named on the call
Two openings confirmed, a third pending budget sign-off
Engagement covers backend only, not the mobile roles

Hiring Requirement

Senior Backend Engineer, Austin TX, hybrid three days on site
Replacement plus growth, target start within six weeks

Ideal Candidate Profile

Not discussed

Must-Have Criteria

Python and distributed systems at production scale
Owned an on-call rotation

Good-to-Have Criteria

Not discussed

Budget, Compensation & Commercials

Range not shared. Client to confirm band before the first submission.

Transcript0:00 / 35:33

RACHEL WHITMORE0:00:04

Before we scope the role, can you walk me through what the platform team owns today?

CLIENT0:00:19

Sure. They own ingestion and the routing service. The person we lost was the only one carrying on-call for both.

RACHEL WHITMORE0:00:41

Got it. So on-call ownership is a must-have, not a nice-to-have.

What you are looking at: the intake summary from a client call, with the recording and an 82 line transcript beside it. Click a transcript line and the recording jumps to it. Note the italic lines. Where something was not covered, the summary says so instead of filling the gap. Then a scored candidate report, the Meetings list every call files into, and the auto-record policy that decides which calls the bot joins. Sample data.

See a sample report

It tells you what you forgot to ask.

If budget never came up, the summary says Not discussed under Budget. It does not guess a range. You leave the call knowing exactly which gaps to close before the first submission.

A score on every skill you picked.

Each skill you set on the job gets its own score, drawn from the transcript and any code the candidate wrote. You can see which answers moved the number.

The recording and a searchable transcript.

The full recording plays back next to a speaker-separated transcript. Search it, or filter it to just the candidate or just the interviewer. Malpractice status sits on the same card.

Share it or download the PDF.

Send the report to a hiring manager or a client, or export it as a PDF. Reports and insights cost 0 credits and are never metered.

It is the same report an AI voice round produces. See the wider interview platform.

Calls you already ran

Score a past call without redoing the interview.

Some of the calls that matter already happened. The panel met the candidate last week, the recording is sitting somewhere, and the only written record is a Slack message.

Upload or pick that recording, choose the job, and you get the full report. Nobody is asked back for a second interview. Every score comes from the transcript of the call that actually happened.

Score a past call

01Upload or select the recording.
Bring a recording of a call you already ran, or pick one the platform has already captured.
02Point it at the job.
Pick the job. The answers are scored against its skills, the weightage on each one and its JD.
03Get the same report.
You get the report every other round produces, with every score backed by the transcript.

A backlog of recorded calls becomes comparable, and no candidate interviews twice.

Comparing a panel

Five interviewers, one set of skills.

Panels are hard to read because everyone measures something different. Here the measure is fixed before the round runs. You pick the skills for the job and set a weightage on each one. You set up the JD, and answers are scored against that too. Every round on the job uses those same inputs.

The same skills across AI and human rounds.

In an AI round the follow-up questions adapt to what the candidate says. In a human round you ask. Either way the answers are scored against the same skills, the same weightage and the same JD, and land on the same report.

Set skills and weightage

Every score points at evidence.

Scores come from the transcript and any code the candidate wrote, not from an interviewer's recollection. A hiring manager who disagrees can open the transcript and check the line.

Read the scoring methodology

A person still makes the call.

The report is input for a human reviewer. A flag is never an automatic verdict, an integrity flag never rejects anyone, and there is no automated pass or fail.

See the recruiter workflow

Consent and data handling

Recording is a legal question, so we treat it as one.

Everyone on the call is told it is being recorded. The recording sits under the same controls as the rest of the platform. The rules where you hire are yours to check.

Nothing is recorded silently.

Participants are notified that the call is being recorded. A candidate is never captured without being told, and neither is the interviewer.

Recording law is local.

Notice and consent rules differ by country and by state. Some places have specific rules for AI and video interviews. Read the guides for where you hire before you switch the bot on. We publish them, but we are not your legal team.

AI hiring compliance guides

Recordings sit under the same controls.

Recordings, transcripts and reports sit under the same controls as the rest of the platform: encryption in transit and at rest, role-based access control, audit logs on every action, no shared candidate pool, and export whenever you want.

Read the privacy policy

Jurisdiction-specific guides cover the Illinois AI Video Interview Act, NYC Local Law 144 and the EU AI Act. SOC 2 and ISO certification are in progress. SSO and SAML are on the roadmap.

What it costs

2 credits per call joined.

Every action draws from one shared monthly pool of Hire Credits. The bot costs 2 credits each time it joins an external call. It unlocks on Starter and up.

Team members are never charged per seat. Unused plan credits roll over for one month on paid plans. The report costs nothing on top.

See plans and the credit meter

2
Calendar notetaker bot
Per external call joined. Available on Starter and up.
2
Human interview with AI copilot
A recruiter-led round hosted inside Expert Hire. Covers the copilot, notes and transcript.
5
AI interview
A live AI voice round with Ethan, any round type.
0
Reports and insights
Always included, never metered.

Where it fits

One piece of a larger record.

The note-taker puts a live human call into the same record as everything else the candidate did. Here is what sits around it.

Instant Meeting rooms

For a call that has no invite. Open a room in one click. It is recorded and filed as a titled report.

Set skills and weightage

How scoring is set up. You pick the skills, set a weightage on each one, and the JD is scored against too.

The interview platform

Five round types in one pipeline: resume screening, AI interview, human-led interview, prompt assessment and coding test.

For recruiters

Where a captured call lands day to day: the workspace, the review queue and the shortlist.

Integrations

Greenhouse is available today in early access, with more ATS connections coming. Mia is reachable from Claude over MCP, and ChatGPT support is in beta.

Scoring methodology

How a score is built, what it is built from, and where the human review sits.

Pricing and credits

The full credit meter, plan by plan, including which tier unlocks the note-taker bot.

FAQ

Questions, answered.

What teams ask before they let a bot into an interview.

Still have a question? Book a demo and we will walk you through it.

Stop writing up interviews from memory.

Let the bot join the calls you already booked. You get a scored report backed by the transcript on every candidate, instead of a paragraph typed between meetings.

30-day free trial with 75 Hire Credits, nothing charged for 30 days. The calendar note-taker unlocks on Starter and up.