HR teams spend an outsized chunk of their week on tasks that don't require human judgment: formatting job descriptions, drafting offer letters, answering the same onboarding questions for the fifteenth time this month. AI doesn't replace the human element of HR — it removes the clerical burden so you can actually do the human element. Here's what's working right now.
Most job descriptions are copy-pasted from the last hire, or worse, written by committee and bloated with jargon. AI is fast at producing first drafts that are structured, inclusive, and scannable — but it needs specifics to do the job well. The key is giving it a role brief, not just a title.
Prompt: "Write a job description for a [Job Title] at a [Company size/type] in the [Industry] industry. The role reports to [Manager title]. Day-to-day: [3-5 bullet points of actual tasks]. Must-haves: [3 non-negotiable skills or experience]. Nice-to-haves: [2-3 bonus qualifications]. Tone: [e.g. direct, warm, technical]. Avoid buzzwords like 'ninja', 'rockstar', 'fast-paced environment'. Keep it under 400 words."
Run it, then ask a follow-up: "Rewrite the requirements section to reduce unnecessary gatekeeping — flag any requirements that could exclude qualified candidates from underrepresented groups without being truly necessary for the role." This is the kind of quick audit that used to require a DEI consultant review.
Building structured interview guides for each role from scratch is tedious. AI can generate a full question bank — behavioral, situational, and role-specific technical questions — in under two minutes if you give it the job description and what competencies you're testing for.
Prompt: "Based on this job description: [paste JD], create a structured interview guide with: - 4 behavioral questions (STAR format) testing for [competency 1] and [competency 2] - 3 situational questions testing judgment in [relevant scenario] - 2 role-specific technical questions - 1 culture-fit question that isn't 'what's your greatest weakness' For each question, add a one-sentence note on what a strong answer looks like."
The "what a strong answer looks like" note is the part that usually takes hiring managers the most time. Getting that scaffolded up front means interviewers go in calibrated, not winging it.
New hire questions follow a predictable pattern: where do I find my benefits info, how do I submit expenses, who do I contact for IT issues, what's the PTO policy. Building an AI-powered FAQ bot for onboarding — even a simple one using Make.com to route Slack messages to a knowledge base — can cut the volume of repetitive questions to HR by 40–60%.
If a full chatbot feels like too much, start simpler: use AI to consolidate your onboarding docs into a single, well-structured "New Hire Quick Reference" document. Feed your existing scattered docs into Claude and ask it to synthesize them into a scannable, logically ordered reference with a table of contents. What used to be five separate PDFs becomes one document new hires actually read.
Prompt: "Here are our onboarding materials: [paste content]. Consolidate these into a single 'New Hire Quick Reference' doc. Organize by: First Day, First Week, First 30 Days, Ongoing Resources. Use plain language. Flag any contradictions or outdated information you notice between the source docs."
Performance review season is where HR time goes to die. Managers submit reviews late, write vague feedback that's either too glowing or too vague to be useful, and the calibration meetings drag on because nobody has a clear frame for comparison.
AI can't write someone's actual performance review — it doesn't know what the employee did. But it can turn bullet-point manager notes into structured, professional review language, and help managers who struggle to articulate constructive feedback turn vague feelings into specific, actionable language.
Prompt (for a manager to use): "I have these rough notes about my direct report's year: [paste notes]. Help me write a performance review that: - Leads with their strongest contribution with specific impact - Addresses the development area without being vague or punitive - Ties their performance to team/company goals where possible - Uses concrete language, not filler like 'great team player' Keep it to 300-400 words, professional but not stiff."
This is a prompt HR teams can package up and send to managers before review season — dramatically improving first-draft quality before it ever lands in your queue for feedback.
HR policy docs are written for legal defensibility, not human comprehension. AI is excellent at translating dense policy language into clear, plain-English summaries employees will actually absorb. Use Jasper AI or Claude to create "TL;DR" summaries of your updated policies — a two-paragraph plain-language version followed by a link to the full document.
For internal announcements (new benefit, policy change, return-to-office update), the same approach applies: give AI the facts and the policy context, and ask for a 200-word email draft that explains the change, the reason, and the next step. You review and approve — but you're not starting from a blank page at 4pm on a Friday.
AI in HR has real limits. Never use it to make employment decisions — candidate screening, performance ratings, termination — without clear human review and accountability. Bias in AI outputs is a documented problem, particularly in hiring contexts. Use it for drafting, organizing, and reducing clerical load. Keep the judgment calls where they belong: with the humans.
The goal isn't to automate HR. It's to give HR professionals back the hours they're currently spending on tasks that a well-prompted language model can handle in 30 seconds — so they can spend more time on the conversations, decisions, and relationships that actually require them.
💡 Building an AI-assisted HR workflow? Start with the right tools. Browse the full AI toolkit →
One weekly email with real-world AI workflows, prompts that actually work, and tool recommendations. No fluff.