Analyze the following HR issue for [ORGANIZATION AND JURISDICTION]. The question to answer is [HR QUESTION]. Apply these relevant policies, agreements, and procedures: [POLICY AND PROCESS]. Use this de-identified evidence: [FACTS AND DATA]. The intended decision-maker and deadline are [DECISION CONTEXT]. Produce a neutral analysis memo of 700-1,000 words with these sections: 1. Scope and decision: restate the question, people or population in scope, decision owner, and deadline. 2. Established facts: list only facts supported by the supplied record, with source labels where available. Separate allegations, employee perspectives, and unresolved facts. 3. Policy and process analysis: identify relevant policy provisions, required steps, notice, documentation, representation, accommodation, or escalation requirements. Do not state legal conclusions unless supplied by qualified counsel. 4. Pattern review: analyze available data by time period, team, job level, location, or process stage only where the dataset supports comparison. Identify missing denominators, small samples, inconsistent definitions, and possible selection effects. 5. Fairness and consistency checks: list comparable prior cases or criteria needed to assess consistent treatment. Do not infer performance, intent, protected status, medical condition, or misconduct from incomplete data. 6. Options and next steps: provide 2-3 proportionate options, each with benefits, risks, evidence still needed, and the appropriate owner or reviewer. 7. Documentation plan: specify what to preserve, who should receive a summary, and what should not be broadly shared due to privacy or need-to-know limits. Use respectful, non-diagnostic language. Do not recommend a hiring, firing, promotion, disciplinary, pay, or accommodation decision based solely on AI analysis. Flag situations involving discrimination, harassment, retaliation, leave, disability, immigration, labor relations, pay equity, safety, or local legal requirements for qualified HR, employee-relations, or legal review. Before answering, check that facts and inferences are visibly separated and that every recommended step can be applied consistently across comparable employees. Ask up to 3 clarifying questions only if a required input is missing.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [ORGANIZATION AND JURISDICTION] | Name the organization or unit and the country, state, or other jurisdiction that governs the issue. | Harbor Street Logistics, Illinois, United States |
| [HR QUESTION] | State the precise workforce, policy, process, or employee-relations question to analyze. | whether the attendance-point policy appears to be applied consistently across warehouse shifts and what review steps are needed before any corrective action |
| [POLICY AND PROCESS] | Paste the relevant policy language, collective agreement terms, process steps, and internal guidance. | Attendance points are assigned after unexcused absences; employees may request review within five business days; approved protected leave and documented accommodations are excluded; written notice is required before final warning. |
| [FACTS AND DATA] | Provide de-identified timelines, counts, records, survey results, and witness or manager notes that are necessary. | De-identified Q2 records for 86 warehouse employees show points by shift, absence reason categories, review requests, and outcome. Seven records lack reason codes. Two employees have pending leave documentation. No individual names, medical details, or protected-status fields are included. |
| [DECISION CONTEXT] | State who needs the analysis, the decision they are considering, and the deadline. | HR director needs a review plan before supervisors issue final warnings next Friday. |
How to use
- Remove names, contact details, medical information, and other unnecessary personal data before pasting records.
- Include the actual policy language and applicable process steps; AI cannot assess consistency against a policy it has not seen.
- Use the memo to identify evidence gaps and review owners before any employment action is decided.
- Follow up with: "Create a de-identified audit table showing the fields HR should compare for every attendance-point case."
Variations
Exit interview themes
Use this when looking for patterns in de-identified exit feedback.
Analyze de-identified exit interview responses from [PERIOD] for [EMPLOYEE POPULATION] using this data: [RESPONSES AND COUNTS]. The goal is [DECISION OR QUESTION]. Produce themes, supporting examples paraphrased without identifying individuals, frequency where countable, contradictions, and recommended follow-up research. Separate volunteered comments from validated causes of turnover. Do not infer protected characteristics, diagnose managers, or claim a theme caused exits without evidence. Check that small samples and missing responses are clearly labeled and that actions are assigned to a function, not a named employee.
Hiring funnel audit
Use this when examining a recruiting process for drop-offs or inconsistent treatment.
Create a fair hiring-funnel analysis for [ROLE OR HIRING PROGRAM] in [JURISDICTION] using [DE-IDENTIFIED FUNNEL DATA, PROCESS, AND POLICY]. Report stage definitions, conversion counts and rates when denominators exist, data-quality limitations, process-consistency questions, and 3 prioritized improvements. Do not use protected traits to make hiring decisions, infer candidate qualifications from incomplete records, or conclude discrimination from a single disparity. Flag areas requiring HR, legal, or adverse-impact review. Check that each metric has a defined numerator and denominator and that recommendations preserve structured, job-related assessment criteria.
Policy gap review
Use this when revising a people policy before rollout.
Review this draft HR policy: [DRAFT POLICY]. It will apply to [EMPLOYEE POPULATION] in [JURISDICTION], alongside these existing rules: [RELATED POLICIES]. Produce a gap analysis covering scope, definitions, eligibility, manager discretion, documentation, exceptions, privacy, notice, appeals, and training. Identify ambiguous language and conflicting provisions, then offer plain-language revision options without asserting legal compliance. Do not invent regulatory requirements. Check that comparable employees would receive the same process and flag provisions that need qualified HR or legal review.
Tips
- De-identify records to the minimum detail needed for analysis; unique job titles, dates, and incidents can still reveal a person in a small team.
- Require denominators before interpreting differences: five complaints in one team means little without team size, reporting channel use, and time period.
- Compare cases using job-related, policy-defined criteria rather than manager impressions or inferred personal characteristics.
- Preserve the original records and document the decision process, but limit distribution to people with a genuine need to know.
FAQ
Can AI decide whether an employee should be disciplined or terminated?
No. AI can organize facts, identify policy questions, and surface consistency checks, but employment decisions require accountable HR and often legal or employee-relations review.
How should HR data be prepared for AI analysis?
Remove direct identifiers and unnecessary sensitive information, define each field and time period, and include relevant policy language. Keep the original records in your approved HR systems.
Can AI identify discrimination from HR data?
It can flag patterns or data gaps that warrant expert review, but it cannot determine a legal or factual conclusion from limited data. Qualified HR and legal reviewers should assess sensitive issues.