Human ResourcesInterviewing & Selection

AI Prompts for Interview Scorecards

A scorecard turns an interview into evidence that can be compared. It lists the competencies the interviewer was responsible for, asks for the evidence (what the candidate said they did), and then for a rating against anchors — in that order, so ratings follow evidence rather than the other way round. Scorecards without an evidence rule produce numbers that reflect the interviewer more than the candidate.

These prompts design the scorecard, convert raw notes into completed ones, and audit a set of completed scorecards for the consistency problems that show up in debriefs. The decision is made in the debrief by people; the scorecard is the evidence they use.

Before you use these

Have these ready to replace the highlighted [variables]:

Competencies per stage and the anchors
Interview notes (verbatim where possible)
Completed scorecards from a recent loop for the audit
Your rating scale and any rules on confidence or 'insufficient evidence'

These prompts structure evidence against stated, job-related criteria and draft documents. They do not make hiring decisions. Do not ask the model to rank, reject or recommend candidates on protected characteristics or their proxies (name, address, school, gaps, age indicators), and do not treat its output as a decision: a person accountable for the hire reviews every score and makes the call. Check local law before using any automated or AI-assisted screening.

The prompts

1. Design the scorecard with an evidence-required rule
2. Convert interview notes into a scored scorecard
3. Audit completed scorecards for consistency

1. Design the scorecard with an evidence-required rule

Best forA scorecard that makes evidence precede rating and records confidence.

Inputs needed

Competencies per interviewer
Anchors
Rating scale

How to use itGive the loop design. The model produces the scorecard template per stage with the evidence field first, anchored ratings, confidence, and the overall recommendation rule.

Expected outputScorecard template per stage, field definitions, the evidence-required rule, the overall recommendation options, and the interviewer instructions.

Act as a talent operations designer creating interview scorecards for [role].

Loop: [stages, competencies per stage, interviewer per stage]
Anchors: [per competency, 1–4 with evidence descriptions]

1. Scorecard template per stage: for each competency — evidence (what the candidate did/said, quoted), rating against the anchor, confidence (high/medium/low), 'insufficient evidence' option; then overall recommendation (strong hire / hire / no hire / strong no hire) with the one-sentence reason; then 'what I could not assess' and 'questions for the next stage'.
2. Evidence-required rule: no rating without evidence in the field above it; ratings without evidence are excluded from the debrief.
3. Field definitions: what belongs in evidence (actions and results) and what does not (impressions, style, personal circumstances).
4. Timing rule: complete within an hour of the interview, before speaking to other interviewers.
5. Instructions (five lines) printed on the card.
6. Data handling: where scorecards are stored, who sees them, retention per policy — to verify with HR.

Design so a reader who was not in the room understands why the rating was given.

2. Convert interview notes into a scored scorecard

Best forTurning raw notes into evidence-backed ratings, for the interviewer to confirm.

Inputs needed

Interview notes
Scorecard template and anchors

How to use itPaste the notes. The model extracts evidence per competency, proposes a rating against the anchors, marks confidence, and lists what was not assessed — as a draft for the interviewer.

Expected outputDraft scorecard with quoted evidence, proposed ratings, confidence, gaps, and the items the interviewer must confirm or change.

You are drafting an interview scorecard from notes for the interviewer to confirm.

Notes: [paste verbatim notes]
Scorecard: [competencies for this stage, anchors 1–4]

1. For each competency: the evidence from the notes (quote the candidate's actions and results), the proposed rating with the anchor it matches, and the confidence based on how much evidence there is.
2. Where evidence is thin or absent: mark 'insufficient evidence' rather than guessing a rating; note the follow-up question that would have produced it.
3. Exclusions: any note content that is not evidence (impressions, style, personal details) — listed separately and not used.
4. Overall recommendation: proposed, with the one-sentence reason tied to the competency evidence.
5. Questions for the next stage.
6. Items the interviewer must confirm: every proposed rating; flag any where the notes could support two ratings.

This is a draft. The interviewer's judgment and confirmation are required before it enters the debrief.

3. Audit completed scorecards for consistency

Best forFinding rating drift, missing evidence and interviewer patterns across a loop or a period.

Inputs needed

Completed scorecards for a loop or period
Interviewer identities (for pattern analysis)

How to use itPaste the scorecards. The model checks evidence completeness, rating-to-evidence alignment, and interviewer patterns — reported at the process level.

Expected outputAudit findings: evidence completeness, misaligned ratings, interviewer leniency/severity patterns, competencies rarely assessed, and the fixes.

Act as a talent operations analyst auditing interview scorecards for [role / period].

Scorecards: [paste: interviewer, stage, competency evidence, ratings, confidence, recommendation]

1. Evidence completeness: share of ratings with no or thin evidence, by stage and interviewer.
2. Alignment: ratings that do not match the evidence recorded (high rating, weak evidence; strong evidence, low rating) — quote and explain.
3. Interviewer patterns: consistent leniency or severity, narrow use of the scale, overall recommendations that ignore competency ratings.
4. Coverage: competencies frequently marked insufficient — is the question, the stage or the interviewer the cause?
5. Consistency across candidates for the same role: was the same evidence rated the same way?
6. Fixes: anchor rewrites, interviewer calibration, loop changes; and what to raise with individual interviewers privately.

Report at the process level. Do not compare candidates against each other here — that belongs in the debrief with the hiring team.

A completed scorecard entry

One competency from a scorecard after the notes-to-scorecard prompt, before the interviewer confirms it.

Competency: Structuring ambiguous problems (anchor 1–4) Evidence (quoted from notes): "We had three teams reporting the same KPI differently. I got the three owners in a room, wrote the three definitions on the board, and asked which decision each one was actually used for. Two collapsed into one. I documented it and the ops review used the single definition from the next month." Probe answer: the candidate identified the decision behind each metric unprompted; described their own actions, not the team's. Proposed rating: 3 — "structures a problem with several stakeholders and reaches a working resolution" (anchor 3). Anchor 4 would require evidence of doing this at function scale or designing the approach others then used — not shown. Confidence: medium (one example; strong detail; no second instance asked for) Excluded from evidence: "very articulate", "seemed senior" (style, impression) Interviewer to confirm: rating 3 vs 4; whether a second example was available

The excluded line is the point. Fluency is not the competency.

Related prompts

Logical next step

After this, most HR teams move on to Structured Interview Design.

Get the free HR & Recruiting AI Starter Kit → Nine of these prompts as an intake → assess → offer hiring workflow with an intake worksheet, delivered by email. See what's inside

All Human Resources prompts · Search the full library