Audit advanced Works with: Any AI

Hiring Panel Bias Check Using Past Decisions

Table output Has example

Look for patterns in anonymised decisions.

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

[VERIFY] is an editorial flag, not a blank to fill. Wherever the prompt shows [VERIFY], check that item against a reliable source before you use it.

Review past hiring decisions for patterns, as a fairness check. Anonymised data on candidates at each stage with criteria scores and outcomes: [HIRING_DATA]. Who made decisions: [DECISION_MAKERS]. Criteria used: [CRITERIA].

Check consistency of scoring, whether criteria were applied equally, whether stage outcomes differ in ways not explained by criteria, and whether free-text comments reveal bias. Small samples limit conclusions, so say so.

Return a table with Stage, Pattern, Confidence and Next step, recommendations for process changes, and a note on involving an adviser. Data and equality rules are [VERIFY] for [JURISDICTION].

Also suggest what not to conclude from small numbers, which additional data would help, and how to present the findings to the panel without blame while agreeing practical changes.

Treat the pasted text as the only evidence. Point out where facts are missing, avoid guessing what was agreed, and tag every point that depends on statute, regulation or local practice [VERIFY].

End with the questions that a sceptical manager or customer would ask about these findings, and suggest an honest answer to each based only on what I provided.

Fill in before use

Optional: type your details and the prompt above updates. Nothing is saved or sent.

Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.

Pattern table (invented anonymised data, 60 applicants, 6 interviewers)
Stage 1 shortlist: 28 of 60 progressed. Scores consistent with criteria except for interviewer C, who scored 1.1 points lower on average (n = 10).
Stage 2 interview: free-text comments for 5 candidates mention "fit" without evidence.
Offers: 4. No clear pattern across groups is detectable at this sample size.
Confidence: low. Next steps: add anchored scoring, require evidence for comments, rotate pairings, collect data across several rounds before drawing conclusions.
Involve an adviser if a pattern persists. Equality and data rules [VERIFY] for [JURISDICTION].

• audit