Create a copy-paste prompt pack for my HR team based on the company context below. Produce exactly 8 standalone prompts: job-description review, interview scorecard, candidate-rejection email, onboarding plan, performance-conversation preparation, manager coaching, employee-relations fact summary, and policy communication. Company context: [COMPANY CONTEXT] HR policies and constraints: [POLICIES AND CONSTRAINTS] Workforce and role details: [WORKFORCE DETAILS] Tone and culture: [TONE AND CULTURE]
Current HR priorities: [HR PRIORITIES]
For each prompt, write it as an HR practitioner speaking to ChatGPT, Claude, or Gemini. Include only the inputs that prompt genuinely needs, using square-bracket placeholders. Make each prompt require a usable output format, such as a scorecard table, a 30-60-90 day plan, a conversation agenda, or an email with subject line.
Build in the controls appropriate to the task. Hiring prompts must separate essential from preferred criteria, use job-related evidence, and avoid protected-characteristic proxies. Performance prompts must distinguish observed behavior, impact, expectations, and documented examples. Employee-relations prompts must organize allegations, sources, chronology, missing facts, and next actions without deciding credibility prematurely. Policy communications must state the effective date, affected audience, required action, and where questions go.
Do not draft legal conclusions, make termination decisions, diagnose medical issues, or invent policy language, employee facts, or investigation findings. State when local employment counsel, HR leadership, or a trained investigator should review the result.
Before answering, check that every prompt has a clear output, only necessary placeholders, and no unsupported legal or factual claim. If a required input is missing, ask up to 3 clarifying questions; otherwise provide the 8 prompts only.Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [COMPANY CONTEXT] | Describe the company size, industry, locations, and working model. | 280-person B2B software company; US and Canada; hybrid product, sales, and support teams. |
| [POLICIES AND CONSTRAINTS] | List relevant internal policies, collective agreements, legal jurisdictions, and approval rules. | Use the employee handbook; California, New York, Ontario rules apply; HR director approves policy changes. |
| [WORKFORCE DETAILS] | Describe the main employee groups, job families, and manager experience level. | Most managers lead 4-10 people; hourly support team and salaried technical staff. |
| [TONE AND CULTURE] | Describe the communication style and values the HR materials should reflect. | Direct, respectful, low-jargon, and candid about decisions. |
| [HR PRIORITIES] | List the current people issues or programs the prompt pack should emphasize. | Reduce inconsistent interviews, improve manager feedback, and standardize onboarding. |
How to use
- Paste the prompt and replace the five company-level inputs with your operating context.
- Save the resulting standalone prompts in your HR knowledge base or approved prompt library.
- For any employee-specific use, add only verified facts and remove details the model does not need.
- Follow up with: “Adapt the manager-coaching prompt for first-time managers handling missed expectations.”
Variations
Interview scorecards
Use this when a hiring team needs a consistent evaluation tool for one open role.
Create a structured interview scorecard for [ROLE] using the job requirements in [JOB POSTING]. Define 4-6 job-related competencies, behavioral questions for each, strong and weak evidence indicators, and a 1-4 rating scale with anchors. Separate must-have from trainable skills. Do not include personality-fit criteria or protected-characteristic proxies. Output a one-page interviewer guide and a scorecard table. Ask up to 3 questions only if a required input is missing. Check that every question maps to a stated requirement.
Policy announcement
Use this when employees need a clear explanation of a changed HR policy.
Draft an employee announcement about [POLICY CHANGE] for [AUDIENCE], effective [EFFECTIVE DATE]. Produce a subject line, 180-250 word email, five-question FAQ, and manager talking points. Explain what changes, what does not change, actions employees must take, deadlines, and the contact for exceptions. Do not invent legal rationale or benefits details. Flag statements that need counsel or benefits-administrator confirmation. Ask up to 3 questions only if a required input is missing.
Investigation plan
Use this when HR needs to organize an intake before a workplace investigation begins.
Turn these reported workplace concerns into a neutral investigation intake plan: [ALLEGATION SUMMARY], [KNOWN FACTS], and [APPLICABLE POLICY]. Output: issues to assess, a chronology with fact gaps marked, evidence-preservation steps, witness categories, neutral interview topics, confidentiality reminders, and escalation points. Do not decide whether allegations are true, promise confidentiality, or recommend discipline. Flag where trained investigators, HR leadership, or employment counsel should lead. Ask up to 3 questions only if a required input is missing.
Tips
- Use AI to structure a fact record, not to decide whose account is credible or what discipline is warranted.
- For interviews, require evidence-based ratings before the debrief so the loudest interviewer does not set the conclusion.
- Keep medical, accommodation, immigration, and investigation details out of broad prompts unless they are necessary and handled in an approved system.
- Have the policy owner verify every effective date, eligibility rule, and contact route before distribution.
FAQ
Can AI write employee-relations documentation?
It can turn verified notes into a neutral chronology or meeting summary. HR should check it for omitted context, confidentiality, and unsupported conclusions.
Can I use AI to choose between candidates?
Use it to create job-related scorecards and summarize documented evidence, not to make or justify a hiring decision.
Should HR paste employee medical information into a public AI tool?
No. Follow your privacy, security, and approved-tool rules, and minimize personal information in any prompt.