What Briefing Room is

The problem it solves

Hiring loses information at every handoff. A recruiter's sense of what the role needs never quite reaches the interviewer. A résumé is skimmed and the reasoning evaporates. Four interviewers ask overlapping questions and none of them cover the one requirement that actually decides the hire. By the time a decision is made, nobody can reconstruct what it was based on.

Briefing Room holds that information in one place, from the moment a role is defined to the moment an offer is signed. Every stage produces evidence that the next stage can read: a rubric with weighted criteria, a résumé score with a quote behind each criterion, an interview transcript scored against the same rubric, and a running count of how much of the rubric is still unevidenced.

New here? Getting started walks through setting up a company and publishing the first role.

The shape of the product

Nine things happen to a candidacy, in this order. Each has its own page in these docs.

1. Define the role. A conversation or a pasted job description becomes a weighted rubric the team can argue with and edit.
2. Get candidates in. People apply on your hosted careers page, or you upload résumés for people who did not. See Getting candidates in.
3. Read every résumé. Extract, check the text is readable, redact identifying details, score against the rubric with a quote per criterion.
4. Interview. An AI voice screen, or a round led by your own interviewers, with every transcript scored the same way.
5. Coach the round. A brief before it saying what is still uncovered, a co-pilot during it, and a critique of the interviewer afterwards.
6. Decide. Two numbers per job, advance and reject, with coverage recorded beside every score.
7. Move the pipeline. A board whose every move records its reason.
8. Talk to the candidate. One email thread per application, with an agent that classifies replies and escalates anything it should not answer.
9. Stand behind it. The notice, consent, EEO, human-review and retention steps that automated hiring decisions require.

Five rules that hold everywhere

These hold on every screen, and knowing them explains most of the product's behaviour.

A criterion nobody asked about is left unscored, not scored zero. An interview that never mentioned mentoring says nothing about mentoring, good or bad. So the interview scorer gives it no score, and the share of the rubric left unscored is reported separately as coverage. (The résumé scorer treats silence the other way on purpose; see Reading résumés.)

Every claim carries a quote. Résumé scores, interview scores, interviewer critiques and compliance findings are all dropped if the model cannot quote the source word for word. A judgement with no evidence cannot be checked, so it is not kept.

A rubric edit adds a new version and keeps the old one. An assessment is always read against the rubric version it actually measured, so a score from March still means what it meant in March.

Scoring happens in the background. A résumé score appears shortly after the application does, and the board shows Scoring… until it lands. The same is true of interview scores and of every email we send.

Your data is walled off. No other company can see your candidates, and interviewers see only the jobs they have been given access to. The database enforces both, so a mistake on one screen cannot expose them.

What it deliberately does not do

• It does not write back to your ATS. ATS connections bring applicants in and change nothing in the ATS. See ATS integrations.
• It does not decide the close calls, or anything until you say so. Automatic decisions are off for a new company, so every decision waits for a person in Inbox → Needs you until an admin turns them on. Even then, a candidate between a job's two numbers (80 to advance, below 65 to reject on a new job) waits for a person.
• It does not source candidates. There is no outbound sourcing and no candidate search across other companies. Your pool is yours alone.
• It does not run background checks, payroll, or onboarding. The product ends at a signed offer.
• It is built for United States hiring law. GDPR, the EU AI Act and India's DPDP Act are out of scope. See Compliance.

Who does what

Each member has one of five roles: owner, admin, recruiter, hiring manager or interviewer. Owners and admins set up the company: verification, integrations, branding and automatic decisions. Recruiters and hiring managers work the pipeline. Interviewers see only the jobs they have been given access to. Some connections belong to each person rather than the company: their own calendar, their agreement to the meeting note taker, their Granola, and their AI assistant. See Integrations.

Where the reasoning lives

Every automatic decision is written to the candidacy's timeline, marked as taken by the system rather than a person, with the score and the threshold it was measured against. The timeline is the record of what happened to a candidacy and why.