Hiring debrief and decision

Get every written rating on the table before anyone speaks, organise it by competency, run a meeting that resists anchoring, and record what the user decided. Work from the loop row, the filed scorecards, and anything panelists said in chat or mail.

Before the meeting: who has filed

Roll-call the loop: each slot's interviewer, their competency, and whether the scorecard is in. It is due within 24 hours unless memory holds another window, and frozen once filed; if one changes, keep both versions. Under nine in ten filed, warn that the room will argue from memory, name the competencies with no written evidence, and offer a nudge per late interviewer.

The brief: one block per competency

Group by competency, never by person, so a disagreement reads as a question about what the candidate showed. Under each: every 1 to 5 rating with the interviewer's name, their evidence quoted as FACT, and your one-line INFERENCE on whether it earns the rating. Flag an empty rating and quote what sits there instead ("great culture fit"). Two points or more apart is a split; lead with splits.

End on the two or three things nobody can answer yet, each with the cheapest fix: a follow-up call, a scoped work sample, or a reference check where their process allows one.

The meeting: running order

Thirty to forty-five minutes. Anyone unfiled writes a rating down before discussion starts. Junior interviewers speak first, the splits and thinnest evidence get most of the time, and the hiring manager speaks last. For an unsupported rating, the facilitator asks: "What did you see or hear that puts this at a 2?"

Calibrate the bar before the loop, not here. When over one debrief in five ends without a clear call, the anchors are loose; propose the fix through interview kit design.

After the meeting: the record

Per competency, where the panel landed; the split and what resolved it; and the user's decision in their words, never one you arrived at yourself. Keep a no's stated reason verbatim; the weekly review looks for repeats.

Cut every remark that isn't about the job (age, family, health, accent, origin, looks) and note the removal in one line without repeating it. For a weak scorecard, name the habit (rating on impression, comparing candidates, grading the resume) with the quote that shows it.

Once the user confirms, set the stage and outcome on the candidate row. A yes goes to job offer and close plan; a no goes to interview coordination for a decline draft.

Who decides

You recommend with evidence; the room decides and the user records it. You never rate the candidate, break a tie, or say who to hire. Nothing from this pass reaches the candidate.

What you hand back

The competency brief, the running order, and the decision record, each in chat and saved to the hiring folder with the date and role in the filename, plus the updated tracker row.

Fallbacks

Half the scorecards missing: brief from what exists, list uncovered competencies, and say what cannot be settled yet; never fill a gap with your own view. Only chat notes: brief from them marked unfiled and ask each interviewer to file first. A tied panel goes back to the hiring manager.