Use AI for HR and Recruitment

What This Guide Covers

HR involves a lot of repetitive writing: job descriptions, CV screening, offer letters, onboarding checklists, training materials. AI can draft all of these in seconds — not to replace human judgement, but to handle the first draft so you can focus on the decisions that matter.

Important: AI should assist HR decisions, never make them. Always have a human review AI-generated content before it reaches candidates or employees. And never upload personal data (CVs, addresses, phone numbers) to public AI tools without explicit consent — for sensitive HR work, use local AI or approved enterprise tools.

• •Write job descriptions that actually attract the right candidates
• •Screen CVs against job requirements — AI-assisted, human-decided
• •Generate onboarding documents and welcome materials
• •Create training materials and quick-reference guides
• •Understand the legal and ethical boundaries

Step 1. Write Job Descriptions

A good job description does two things: it attracts the right people and filters out the wrong ones. Most job descriptions fail at both — they're either too vague or a wall of bullet points. AI can structure them properly.

Job Description Prompt Template

After the AI generates a draft, review it for accuracy (especially salary and location details), add any role-specific requirements the AI missed, and ensure the tone matches your company culture.

Step 2. Screen CVs — AI-Assisted, Not AI-Decided

AI can do a fast first pass on CVs, comparing them against your job description and flagging strong matches, gaps, and potential red flags. But AI should never be the final decision-maker — it misses context, can amplify bias, and may hallucinate qualifications.

CV Screening Prompt Template

Critical safety rules for CV screening with AI:

• •Never upload CVs containing personal data (names, addresses, phone numbers) to public AI tools like ChatGPT or Claude unless you have the candidate's explicit consent. For sensitive screening, use local AI via Ollama or enterprise-approved tools.
• •AI can introduce or amplify bias — it may favour certain universities, career paths, or phrasing patterns. Use it for organisation, not judgement.
• •Always have a human review the AI's summary before shortlisting or rejecting anyone.

Step 3. Generate Interview Questions

Tailor interview questions to the role and the specific candidate. AI can generate a structured set that covers technical skills, behavioural fit, and culture alignment.

Step 4. Create Onboarding Documents

Onboarding is mostly writing: welcome emails, first-week schedules, IT setup checklists, handbook digests. AI can generate all of these from your existing materials.

Onboarding Prompt Template

For the employee handbook digest: upload your full handbook as a PDF to ChatGPT or Claude and ask it to produce a one-page summary covering only what a new starter needs in their first week.

Step 5. Create Training Materials

Turn policy documents, process guides, and training manuals into digestible formats. AI excels at taking dense documentation and producing summaries, quizzes, and quick-reference cards.

Training Prompt Examples

• •Slide outline: "Turn this policy document into a 10-slide training deck outline. Each slide should have a title and 3-5 bullet points."
• •Quiz questions: "Create 10 multiple-choice quiz questions based on this training material. Include the correct answer and a brief explanation for each."
• •Quick-reference guide: "Summarise this 20-page process document into a one-page quick-reference guide. Use bullet points and include only what someone needs to do the job — not the background or rationale."
• •Role-play scenarios: "Create 5 customer service role-play scenarios based on our policies. Each should describe a situation, what the employee should do, and what not to do."

Step 6. Legal and Ethical Boundaries

Using AI in HR comes with real responsibilities. Here are the lines you should never cross:

• •Never let AI make the final hiring decision. AI can score, rank, and suggest — but a human must decide. Automated decision-making in recruitment may be illegal in some jurisdictions without transparency and appeal processes.
• •Don't upload personal data to public AI tools. CVs contain names, addresses, phone numbers, and employment history. Under GDPR, uploading this to ChatGPT or Claude without consent may be a data breach. For sensitive screening, use local AI or tools with data processing agreements in place.
• •Watch for bias. AI models are trained on historical data that reflects existing biases in hiring. They may unfairly favour certain universities, career paths, or demographic patterns. Always review AI suggestions through a fairness lens.
• •Be transparent. If you use AI to screen CVs or generate interview questions, your AI usage policy should say so. Candidates have a right to know how their application is being processed.
• •Verify everything. AI will confidently hallucinate qualifications, invent experience, and make assumptions not supported by the CV. Every AI-generated summary must be checked against the source document.

A simple rule: AI handles the first draft and the admin. Humans handle the decisions and the communication. Never outsource judgement to a machine.

Quick Tips

• •Build a library of HR prompt templates in a shared document. Everyone on the team uses the same templates — consistency across all hiring managers.
• •For recurring hires (same role, multiple times), create a ChatGPT Project with the job description, company values, and screening criteria pre-loaded.
• •Strip identifying information from CVs before uploading to AI tools — or better yet, use local AI so the data never leaves your machine.
• •Keep your AI usage policy updated. If you start using AI for screening, your policy should say so. See: Team AI Policy Guide
• •Start with non-critical tasks (drafting job descriptions, summarising policies) before moving to higher-stakes use (CV screening, interview questions). Build trust in the tool gradually.