AI Prompts for HR — Complete Framework for Job Descriptions, Interviews and Performance Reviews 2026

Human resources sits at the intersection of two realities: the work is deeply human, and yet a significant portion of the day-to-day output is writing. Job descriptions, interview scorecards, performance review narratives, onboarding documents, offer letters, and policy FAQs all require time, consistency, and careful language. That combination makes HR one of the best early adopters of AI tools.

A 2024 SHRM survey found that HR professionals spend an average of 14 hours per week on administrative writing tasks. AI does not replace HR judgment — it eliminates the blank page. This guide provides a practical prompt framework for every major HR writing task, along with the legal and ethical guardrails that matter most.

Why HR Is One of the Best Early AI Adopters

The tasks most suited to AI assistance share three characteristics: they are high-volume, they follow a repeatable structure, and the cost of a first draft being imperfect is low because a human reviews the output before it is used. HR writing hits all three criteria.

Job descriptions at a 200-person company might need to be refreshed 40 or 50 times per year. Each performance review cycle might require a manager to write 8 to 12 individual narratives under time pressure. Behavioural interview question banks get stale and inconsistent across hiring managers. AI solves the volume and consistency problem. The HR professional still exercises judgment on what goes into the prompt, what comes out, and what gets published.

Part 1: Job Descriptions

Prompt Framework for Writing Job Descriptions

A strong job description prompt needs five inputs: role title, seniority level, must-have skills, company culture signals, and the output format you want. Vague prompts produce generic output. Specific prompts produce usable first drafts.

The framework:

Write a job description for a [ROLE TITLE] at [SENIORITY LEVEL] level.

Company context: [2-3 sentences about company stage, industry, and culture]

Must-have skills and experience: [bullet list of 5-7 hard requirements]

Nice-to-have skills: [bullet list of 2-3 preferred but not required]

Format: Include a company overview paragraph, a responsibilities section (8 bullet points), a requirements section (split into "Required" and "Preferred"), and a benefits paragraph. Keep total length under 600 words. Use inclusive language throughout.

Example: Senior Product Manager Job Description

Write a job description for a Senior Product Manager at a mid-level seniority (5-8 years experience).

Company context: Series B SaaS company building workflow automation software for professional services firms. 80 employees. Culture is async-first, outcome-focused, and direct.

Must-have skills:
- 5+ years of product management experience in B2B SaaS
- Proven experience owning a product roadmap from discovery to launch
- Strong data analysis skills (SQL is a plus but not required)
- Experience working directly with enterprise customers
- Track record of working cross-functionally with engineering and design

Nice-to-have:
- Experience in workflow automation, legal tech, or professional services
- Familiarity with Jobs-to-be-Done framework

Format: Company overview (60 words), responsibilities (8 bullets), requirements split into Required and Preferred sections, benefits paragraph. Under 600 words. Use inclusive language throughout.

ChatGPT, Claude, or Gemini will produce a complete, formatted draft from this prompt in under 30 seconds. The output will need review — verify that responsibilities reflect your actual team's work, not a generic PM description — but you have a working draft rather than a blank page.

How to Avoid Bias in AI-Generated Job Descriptions

AI models are trained on historical text, which means they can reproduce historical biases. The most common issues are gendered language (words like "aggressive," "dominant," or "rockstar" have been shown to deter female applicants), unnecessary credential requirements (requiring a degree when the role does not functionally need one), and physical requirements that may discriminate against disabled candidates.

Counter this with explicit prompt instructions:

Review the job description above and flag any language that:
1. Uses gendered adjectives (e.g., "aggressive," "nurturing," "dominant")
2. Requires credentials that are not strictly necessary for the role
3. Could deter applicants from underrepresented groups
4. Uses jargon that excludes non-native English speakers

Rewrite any flagged sections using neutral, inclusive language.

Tools like Textio and Applied also run automated bias checks on job descriptions if you want a dedicated workflow.

Compliance Note

AI has no access to your jurisdiction's employment law. It does not know whether your state requires salary ranges in job postings (California, New York, Colorado, and Washington all do as of 2025), whether certain background check disclosures are required, or whether specific language is needed for OFCCP compliance. Always route AI-generated job descriptions through legal review before posting.

Part 2: Interview Questions

Prompt to Generate Behavioural Interview Questions (STAR Format)

Generate 8 behavioural interview questions for a Senior Product Manager role.

Each question should follow the STAR framework (Situation, Task, Action, Result) and probe for a specific competency. Include these competencies: stakeholder management, prioritisation under constraints, handling failure, data-driven decision making, cross-functional leadership, customer empathy, technical communication, and roadmap strategy.

Format each question with: the competency name in bold, the question, and 2 follow-up probing questions.

Prompt to Generate Technical Screening Questions

Generate a technical phone screen question set for a Senior DevOps Engineer role.

The screen is 45 minutes. Include:
- 3 conceptual questions (e.g., how would you explain X)
- 3 scenario-based questions (e.g., walk me through how you would handle X)
- 2 tool-specific questions covering Kubernetes and Terraform
- 1 question about incident response process

For each question, include what a strong answer looks like in 2-3 bullet points.

Prompt to Create a Structured Interview Scorecard

Create a structured interview scorecard for a Senior Product Manager role.

Include 6 competencies relevant to this role. For each competency:
- Define what it means in the context of this role
- Provide a 1-5 rating scale with behavioural anchors for scores 1, 3, and 5
- Include one interview question that best assesses this competency

Format as a table with columns: Competency | Definition | Interview Question | Score (1-5) | Evidence/Notes

Part 3: Performance Reviews

Prompt to Draft a Performance Review from Bullet Points

Write a performance review narrative for [EMPLOYEE NAME], [ROLE TITLE], covering the period [DATE RANGE].

Here are my bullet-point notes from this period:
[paste your bullet points]

The narrative should:
- Be written in third person
- Be approximately 300-400 words
- Cover achievements, areas of strength, and 1-2 development areas
- Use specific examples from the notes
- Be professional but warm in tone
- Avoid vague filler phrases like "hard worker" or "team player" without evidence

Prompt to Turn Quantitative Data into Narrative Text

Turn the following performance metrics into a 150-word narrative paragraph for a sales manager's performance review. Focus on what the numbers mean, not just what they are.

Metrics:
- Closed $1.4M in ARR against a $1.1M target (127% attainment)
- Managed a team of 6 AEs, 2 of whom were promoted
- Reduced average sales cycle from 87 days to 61 days
- Net Promoter Score from his accounts: 72 (company average: 58)

Tone: specific, evidence-based, professional. Do not use superlatives.

Prompt to Generate Development Goals from Review Themes

Based on these recurring themes from a performance review, generate 3 SMART development goals for the next 6 months.

Review themes:
- Strong technical execution but struggles to communicate strategy to senior leadership
- High output individually but has not yet developed direct reports' capabilities
- Takes feedback well but is slow to act on it without follow-up

For each goal, provide: goal statement, success metric, timeline, and 2 specific actions the employee can take.

Part 4: Onboarding Materials

Prompt to Create a 30-60-90 Day Plan Template

Create a 30-60-90 day onboarding plan template for a new Senior Product Manager joining a Series B SaaS company.

For each phase (Days 1-30, Days 31-60, Days 61-90), include:
- Primary focus for the phase (1 sentence)
- 4-5 specific goals
- Key relationships to build
- Deliverable or milestone that marks the end of the phase

Format as a structured document. Tone: practical and welcoming, not bureaucratic.

Prompt to Write FAQ Documents for New Hires

Write a New Hire FAQ document for the People Operations team at a 120-person remote-first tech company.

Include answers to the following questions in plain, friendly language:
- How do I set up my equipment?
- How does payroll work and when do I get paid?
- How do I book time off?
- What is the process for expensing work-related costs?
- Who do I contact if I have a benefits question?
- How does performance review work here?
- What are the core working hours, and how flexible is the schedule?

Keep answers under 100 words each. Use plain language. Avoid HR jargon.

HR AI Prompt Library

Task	Prompt Template Input	Output Format
Job description	Role + level + skills + culture + format instructions	Structured JD with sections
Bias review	Paste JD + ask for flagged language	Annotated rewrite
Behavioural questions	Role + competency list + STAR instruction	Question bank with follow-ups
Technical screen	Role + duration + topic areas	Q&A with strong answer criteria
Interview scorecard	Role + competencies	Table with rating scale
Performance review narrative	Bullet notes + employee context	300-400 word narrative
Metrics to narrative	Data points + tone instruction	150-word paragraph
Development goals	Review themes + timeframe	3 SMART goals with actions
30-60-90 day plan	Role + company stage + format	Phased plan with milestones
New hire FAQ	Question list + company context	Plain-language Q&A document

Legal and Ethical Guardrails

These are the hard lines for AI in HR:

AI must never make final hiring decisions. AI can score, rank, or summarise — but a human must make the final call on who is hired, promoted, or terminated. This is not just an ethical position; it is increasingly a legal one. The EEOC's 2023 technical assistance document on AI and employment makes clear that employers are responsible for discriminatory outcomes from AI tools regardless of whether a vendor built the tool.

AI must never evaluate protected characteristics. Do not feed AI a CV with a name, photo, graduation year, or address and ask it to assess "fit." These proxies correlate with race, gender, age, and national origin. Blind screening means removing identifying information before AI touches the document.

AI-generated content must be reviewed before use. Every prompt output in this guide is a first draft. The HR professional owns the final document. If a JD, review narrative, or scorecard contains something inaccurate, discriminatory, or legally non-compliant, it is the employer's liability.

Do not input personal employee data into public AI tools. If you are using ChatGPT, Claude.ai, or any AI tool outside your company's approved stack, do not paste real employee names, salaries, health information, or performance ratings. Use placeholders and anonymise before prompting. Check whether your company has an enterprise agreement with an AI vendor that includes data processing protections.

AI makes HR faster and more consistent. It does not make it risk-free. The professionals who get the most value from these tools are the ones who understand exactly where the human judgment must stay.