Synthesize Interview Feedback
Synthesize interview transcripts, notes, and panel feedback against an accepted scorecard.
Turn fragmented interview evidence into a draft assessment a qualified human can review. Do not collapse uncertainty into a score or recommendation.
1. Establish the review frame
Find the accepted job description or internal role profile, interview plan, scorecard, competency definitions, interviewer assignments, candidate materials, and available transcripts or notes. Confirm the candidate, role, interview set, decision stage, and intended reviewers when ambiguous.
If no accepted scorecard exists, state that limitation and propose a review frame for human approval before judging fit. Do not retrofit criteria from the candidate's performance or introduce protected or irrelevant characteristics.
2. Preserve provenance
For each material point, distinguish:
- observed interview evidence or work-sample behavior;
- a claim made by the candidate;
- an interviewer's interpretation or judgment;
- the companion's inference;
- a contradiction between sources; and
- an unknown or untested criterion.
Keep links or references to the originating transcript, note, scorecard field, or work sample where available. Never invent quotes, scores, interviewer views, or evidence. Treat automated transcription as fallible and flag consequential ambiguity.
3. Map evidence to the scorecard
Organize evidence by the accepted competency or outcome. For each criterion, show supporting evidence, counter-evidence, confidence, and what remains unknown. Respect any defined scoring anchors, but do not manufacture numerical precision when the evidence is qualitative or inconsistent.
Identify interviewer coverage gaps, duplicated questioning, unsubmitted feedback, untested must-haves, and disagreements that need discussion. Keep post-interview impressions separate from evidence recorded during the interview where that distinction is available.
4. Prepare the draft assessment
Deliver:
- a concise candidate and interview-set summary;
- a criterion-by-criterion evidence table;
- contradictions, missing feedback, and material unknowns;
- candidate claims that still need verification;
- candidate-experience or process risks; and
- a draft assessment framed for qualified human review.
When useful, give understandable overall fit guidance using the same non-dispositive language as earlier review stages:
- strong potential fit --- the interview record supports most critical criteria and shows no material conflict;
- potential fit with questions --- relevant evidence exists, but important points still need discussion or verification;
- limited fit based on current evidence --- important criteria lack support or have material counter-evidence; or
- unable to assess --- the interview record is too incomplete, ambiguous, or inconsistent.
Explain the evidence, counter-evidence, confidence, and gaps behind the label. It is not a comparative ranking or a recommendation to reject or advance. The skill must not hire, disposition, or communicate a decision; make the human judgments still required explicit.
5. Hand off safely
Offer focused follow-up questions or an additional interview only when they address a material gap and remain proportionate. Prepare ATS notes or a panel brief as a separate draft; require explicit review before writing external records or sharing sensitive feedback.
Use @prepare-candidate-interviews when the next step is another structured interview. Use @review-recruiting-pipeline when missing feedback or stalled decisions affect the wider pipeline. External candidate communication belongs in @send-candidate-outreach and requires exact send approval.
If the panel accepts the synthesis format, offer to save the scorecard mapping, evidence labels, destination, and review point as a custom skill. Use a Routine only for a stable, review-first workflow; never automate the hiring judgment.
