HR is full of writing tasks that are important but repetitive, and that is exactly where ChatGPT shines. This guide gives you 12 prompts you can paste in today, grouped by job: hiring, onboarding, policy, feedback, and the awkward emails. The non-negotiable is privacy: strip out real names, salaries, and anything that identifies a person before it goes into a public tool. Treat ChatGPT as a fast, tireless first-drafter, never as the final word on a people decision.
Here is the honest truth about HR work: a huge share of it is writing. Job adverts, offer letters, onboarding checklists, policy updates, the careful email to a manager who keeps skipping one-to-ones. None of it is hard, exactly. It is just constant, and it eats the hours you would rather spend on the actual people.
That is the gap ChatGPT fills well. Not the judgement calls, not the empathy, but the first draft. According to SHRM, 39% of organisations have already adopted AI somewhere in HR, and 92% of CHROs say they want more of it [1][2]. The wider picture backs this up: McKinsey found 88% of organisations now use AI in at least one business function [3]. HR is squarely in that wave.
Below are 12 prompts you can paste in today, grouped by the job you are actually doing. They are written so a first-time ChatGPT user can use them without Googling anything else. But read the next section first, because there is one rule you cannot skip.
Read this first: the one rule that keeps you out of trouble
HR data is some of the most sensitive data in any company. Salaries, performance issues, medical accommodations, disciplinary records. Consumer AI tools can use what you type to improve their models unless you turn that setting off, and either way you are still sending personal data to a third party. That is a risk you do not need to take.
So the rule is simple: never paste real, identifiable employee information into a public chatbot. Strip out names, swap in “the employee” or “Manager A,” remove actual salary figures, and describe situations generically. ChatGPT does not need to know it is Priya in finance to write a good performance-review template. It just needs the shape of the task.
Anonymise before you paste. Replace real names with roles, remove exact salaries and dates of birth, and never upload a document containing personal data into a free consumer tool. If your organisation has an enterprise AI tool with a data-protection agreement, use that for anything sensitive. When in doubt, leave it out.
The simple structure behind every good HR prompt
Every prompt below follows the same shape, and once you see it you can write your own. Tell the AI who to be, what to do, and how you want it back. Role, task, format. If you want the full breakdown, we wrote a whole guide on the 4-part prompt formula, but role-task-format gets you 90% of the way for HR work.
The other trick: give it your tone. “Write in a warm but professional voice” produces something completely different from “write in a formal, policy-style voice.” Tell it, or it will guess.
Recruiting and job descriptions
Job descriptions are the perfect starter task. They are formulaic, they take ages to get right, and a good prompt gets you 80% there in seconds.
Act as an experienced recruiter. Write a job description for a [job title] at a [company size and industry] company. Include a short role summary, 6 key responsibilities, 5 must-have requirements, and 3 nice-to-haves. Keep the tone warm and inclusive, avoid jargon and gendered language, and keep it under 500 words.
Review the following job advert for language that might discourage strong candidates from applying, especially women, career changers, and people from non-traditional backgrounds. Suggest specific rewrites. Here is the advert: [paste your anonymised advert].
I am hiring for a [role]. Based on these three key responsibilities [list them], write 8 interview questions: 4 behavioural, 2 situational, and 2 that test for [specific skill]. For each, add one sentence on what a strong answer would include.
A quick warning on the hiring side: if your company uses AI to screen or rank candidates, that is a different and more regulated activity than drafting questions. We covered the implications in our piece on AI resume screening. Drafting is low-risk. Automated decisions about people are not.
Onboarding and the first 90 days
New hires form much of their early impression of your company in the first couple of weeks. A structured onboarding plan is one of the highest-leverage things HR produces, and it is tedious to build from scratch every time.
Create a 30-60-90 day onboarding plan for a new [role] joining a [team/department]. For each phase, list goals, key people to meet, and what success looks like. Format it as a simple table the new hire and their manager can both follow.
Write a warm welcome email to a new starter joining as a [role] next Monday. Include what to expect on day one, who will meet them, and one line that makes them feel genuinely glad they accepted. Keep it under 200 words and friendly, not corporate.
Policies and internal communications
Policy writing is where ChatGPT saves the most time and where you must check the hardest. It will produce a clean, readable draft. It will not know your local employment law. Use it for structure and clarity, then have the right person verify the substance.
Draft a clear, plain-English [topic, e.g. remote work] policy for a [country] company with about [number] employees. Cover purpose, who it applies to, the main rules, and what to do if there is a question. Avoid legalese. Flag anywhere I should get legal or compliance sign-off.
Rewrite the following policy so a brand-new employee can understand it in one read. Keep every rule intact, but use short sentences, plain words, and a friendly tone. Add a two-line summary at the top. Here is the policy: [paste].
ChatGPT does not know your jurisdiction, your union agreements, or last month’s legal update. Treat every policy draft as a starting point that a qualified person reviews. The time you save is in the writing, not the responsibility.
Performance and feedback
Feedback is hard to write well, especially the constructive kind. ChatGPT is a surprisingly good sparring partner for getting the tone right, as long as you keep the actual person anonymous.
I need to give an employee feedback about [generic issue, e.g. missing deadlines]. Help me phrase it so it is specific, kind, and focused on behaviour rather than character. Give me an opening line, the core message, and a forward-looking close. I will add the real details myself.
Create a performance-review template for a [role] that covers achievements, areas for growth, goals for next quarter, and support needed. Make it strengths-first and easy to fill in. No personal data, just the structure.
Tricky conversations and the emails nobody enjoys
Half of HR is wording difficult things carefully. ChatGPT will not have the conversation for you, but it will help you walk in prepared.
I have a difficult conversation coming up with a team member about [generic topic]. Act as an experienced HR coach. Give me a simple structure for the conversation, three things to say to keep it constructive, and two phrases to avoid because they sound accusatory.
Rewrite this email so it stays firm and clear but lands warmer and less abrupt. Keep it brief. Here is my draft: [paste your anonymised draft].
If your inbox is the real bottleneck, the same anonymise-then-prompt approach works for triage too. Our inbox triage workflow pairs nicely with these prompts.
Learning and development
Outline a 60-minute internal training session on [topic] for non-managers. Include three learning objectives, a rough timing breakdown, one interactive exercise, and three questions to check understanding at the end. Keep it practical, not theoretical.
That last one scales. Once you have the outline, you can ask ChatGPT to turn each section into slides, a handout, or a follow-up email. One prompt becomes a whole session.
What ChatGPT should not do in HR
Let us be honest about the limits, because this is where people get burned. ChatGPT should not make decisions about real people. It should not be the sole tool screening candidates. It should not be fed sensitive personal data in a consumer account. And it should not be trusted on legal or regulatory specifics without a human expert checking.
Used inside those lines, it is a brilliant drafter, summariser, and thinking partner. Used outside them, it becomes a compliance risk and a fairness problem. The skill is knowing the difference, and now you do.
Your first 20 minutes this week
Do not try to roll AI into every HR process at once. Pick one task you wrote last week, a job description, a welcome email, a policy summary, and run the matching prompt. Anonymise your inputs, paste, edit, done. Time yourself.
You will almost certainly notice two things: the draft is better than a blank page, and you finished in a fraction of the time. That is the proof you need. Build from there, one prompt at a time, and keep a human firmly in the loop on anything that touches a real person.
Frequently asked questions
What are the best ChatGPT prompts for HR?
The most useful HR prompts follow a role-task-format structure and target writing-heavy tasks: job descriptions, 30-60-90 onboarding plans, plain-English policy drafts, performance-review templates, and reframing difficult feedback. The 12 prompts in this guide cover each of those, and you can adapt them by filling in your role, company tone, and constraints.
Is it safe to use ChatGPT for HR work?
It is safe for drafting and structuring, but not for sensitive personal data in a public tool. Anonymise everything first: replace real names with roles, remove salaries and identifying details, and never upload employee records into a free consumer account. For anything sensitive, use an enterprise tool that has a data-protection agreement in place.
Can ChatGPT replace an HR manager?
No. ChatGPT is a fast first-drafter and thinking partner, but it cannot make fair decisions about real people, handle the human side of difficult conversations, or be trusted on local employment law. It saves time on the writing so HR professionals can spend more time on the judgement calls that actually need a person.
Can I use ChatGPT to screen or rank job candidates?
Be very careful here. Drafting interview questions or job adverts is low-risk, but using AI to automatically screen, rank, or reject candidates is a regulated activity in many regions and carries real fairness and legal risk. Keep automated decisions about people under human review and check your local rules before relying on AI for screening.
How do I write my own HR prompt?
Use the role-task-format structure: tell ChatGPT who to act as (an experienced recruiter, an HR coach), what to produce (a policy draft, a review template), and how you want it back (a table, under 300 words, warm tone). Add your context in brackets, keep it anonymised, and treat the result as a draft you finish yourself.
About this guide
This is a practical prompt library for HR and People Ops professionals using ChatGPT, written from training 2,000+ non-technical professionals to use AI at work. It includes 12 copy-paste prompts and a clear data-privacy rule. Adoption figures come from SHRM’s State of AI in HR 2026 and McKinsey’s State of AI 2025.
Related reading
Sana is an AI educator and learning designer specialising in making complex ideas stick for non-technical professionals. She has trained 2,000+ learners across corporate teams, bootcamps, and keynote stages. Future Factors offers AI Bootcamps, Corporate Workshops, and Speaking & Consulting for businesses ready to adopt AI without the overwhelm.
More about Sana →- [1] SHRM. The State of AI in HR 2026. 2026.
- [2] SHRM. The State of AI in HR in 2026: 5 Critical Insights for CHROs. 2026.
- [3] McKinsey. The State of AI in 2025. 2025.
- [4] OpenAI. How People Are Using ChatGPT. 2025.
- [5] SHRM. AI in HR 2026: From Hype to Measured, Human-Centered Impact. 2026.
