Skip to content

ChatGPT prompts for recruiters: a practical, review-first playbook

Recruiting work is full of first-draft tasks: sourcing briefs, outreach notes, job ads, interview questions, reference requests, offer letters and follow-ups.

A well-built prompt turns those into usable starting points instead of generic filler. This playbook gives you 20+ recruiter prompts grouped by hiring stage, a seven-part prompt structure, a worked example, and a review checklist so every output is checked before it reaches a candidate or hiring manager.

What makes a recruiting prompt work

A prompt is a brief. The more clearly you describe the role, the task, the context, the limits and the shape of the answer, the less editing you do afterwards. The table below is the field map this page uses throughout. Treat it as a checklist before you paste any prompt.

The seven-part prompt field map

Use this map to diagnose a weak draft. If the output is vague, the missing field is usually Context or Constraints. If the output is plausible but wrong, the missing field is Review step.

Sourcing and outreach prompts

Sourcing prompts work best when you name the role, the seniority, the must-have skills and the channels you plan to search. Outreach prompts work best when you supply real candidate details and a word limit. The sourcing brief below asks for must-have skills, nice-to-have skills, related job titles, target industries, likely current companies, Boolean search strings, LinkedIn search keywords, communities to check and screening questions, with assumptions flagged separately. The outreach prompt asks for a first-touch message under 120 words, conversational and low-pressure, ending with one simple question. A Boolean search prompt asks for a LinkedIn string built from a content writer profile and highlights the skills needed to select the right candidate.

Worked example: turning a sourcing brief into a shortlist

Suppose you are hiring a senior backend engineer with distributed systems experience at a SaaS company. Start with the sourcing brief prompt. The output should give you a structured list of must-have and nice-to-have skills, adjacent job titles, target industries, likely current employers, Boolean strings, LinkedIn keywords, communities and screening questions. Before you search, do three things. First, read the flagged assumptions and mark each one as confirmed, rejected or unknown. Second, run the Boolean string in LinkedIn and check the result count; if it is too broad or too narrow, edit the string rather than the prompt. Third, use the screening questions as a phone-screen script, not as a scoring rubric. Only after a human has reviewed the shortlist should you move to the outreach prompt, where the candidate's current role, recent post or project, relevant skill, role title, company and reason for fit become the personalisation fields. This sequence keeps the model in a drafting role and keeps the judgement with you.

Job ads, social posts and internal promotion

Job description prompts ask for a candidate-focused draft that highlights impact, growth and benefits. A follow-up prompt can turn the same role into a short LinkedIn post with a clear call to action and no hype. Internal promotion prompts ask for an all-employee email that explains the opportunity and the growth benefits. In each case, verify salary details, compliance language and any claims about perks or culture before publishing.

Interview questions, scorecards and assessments

Interview prompts can produce a short set of questions that test adaptability, decision-making and problem-solving, with notes on what strong and weak answers look like. A follow-up prompt can build a structured scorecard for collaboration, communication and expertise, with a bias-free rating for each level. Assessment prompts ask for a challenging task with scoring criteria. Use these as drafts for a human panel to review, not as final hiring decisions.

Reference checks and compliance

Reference prompts can draft open-ended questions that reveal technical ability, teamwork, adaptability and areas for improvement, or a short email asking a former manager for a written reference. Compliance prompts can draft a high-level checklist covering consent, data privacy, documentation, anti-discrimination practices, interview records, background checks and offer communication, with a note that local legal review is required before implementation. Always verify reference notes and compliance language before acting on them.

Offers, onboarding and exits

Offer prompts can draft a personalised job offer email that goes beyond salary to cover technical challenges, career growth and differentiators, or suggest negotiation options that balance fairness, budget and competitiveness. Onboarding prompts can draft a warm welcome email with start date, schedule, tool access and culture tips. Exit interview prompts can produce questions about reasons for leaving, leadership and team culture, plus a follow-up thank-you email. Verify salary and offer details before sending.

Hiring manager collaboration

Intake prompts can suggest questions that clarify requirements, team culture, deliverables and priorities, including questions that surface what a hiring manager may not say upfront. A follow-up prompt can draft a concise email summarising the agreed profile, terms, success metrics, timelines, interview process and next steps. These drafts save time but should be checked against what was actually agreed in the meeting.

Rejection emails and talent engagement

Rejection prompts can draft an empathetic email for interviewed candidates who were not selected, encouraging them to stay connected. Talent engagement prompts can suggest a strategy for reconnecting with candidates who were not selected or did not apply, for future opportunities. Keep these messages factual and respectful, and check any claims about future roles.

Review checklist before you send

Run every draft through these checks. Does the output sound like your organisation, or like a generic template? Does it include role-specific details, or only broad statements? Does it follow the constraints you set, such as length, tone and format? Would you need to rewrite more than a sentence or two? If the answer to the last question is yes, the prompt needs more context, not a different model. Separately, verify salary figures, compliance language, candidate assessments, reference notes and market research before sending or acting on the output. Remove sensitive candidate details unless your team has approved the tool, privacy settings and data handling process. Use the model for drafting, summaries and structured outputs, not for final hiring decisions.

Weak prompt or weak output? A quick distinction

A weak prompt produces generic text that ignores the role, the seniority or the format you asked for. The fix is more context: add the missing field from the seven-part map. A weak output can look specific but contain an unverified salary, a biased assumption or a compliance claim that does not apply in your region. The fix is review, not rewriting the prompt. Keep the two problems separate so you do not over-engineer prompts to compensate for outputs that simply need a human check.

Prompt element and what to include
Prompt elementWhat to include
RoleTell ChatGPT who to act as, such as recruiter, sourcer, or hiring manager
TaskState the exact output needed, such as email, scorecard, job ad, or Boolean string
ContextAdd role, industry, seniority, location, hiring goal, and candidate background
ConstraintsMention tone, length, must-have details, exclusions, and compliance notes
Output formatAsk for bullets, table, email draft, checklist, scorecard, or template
Review stepAsk ChatGPT to flag assumptions, missing details, or claims to verify

This table is the field map used throughout the page. It is a prompt-design aid, not a guarantee of output quality.

Frequently asked questions

How much context do I need to give to ChatGPT for it to work well in recruitment tasks?

Treat the model like a new team member: give it the role, task, context, tone and output format before asking for a draft. For a job description, that means naming the role, the industry or company type, the key responsibilities or must-have skills, the tone you prefer and any specific expectations. You do not need to write an essay, but you do need enough detail for the model to avoid generic output.

How do I know if my prompt is relevant?

A strong prompt produces output that sounds natural, includes role-specific details, follows your instructions and needs only light editing. If the draft feels generic or you would need to rewrite most of it, that is a signal to add context rather than to abandon the prompt. The best way to improve is to brief the model the way you would brief a new recruiter on your team.

What is prompt engineering in recruitment?

It is the practice of giving the model enough context to act as a sourcing assistant, content writer, interview planner or hiring workflow helper. A well-engineered recruiting prompt usually includes the specific task, context about the role, seniority or industry, the tone you want, and any constraints such as length or format. Generic prompts tend to produce generic answers.

How much can I rely on the decisions made by ChatGPT?

Use it to support recruitment work, not to make final hiring decisions. It can draft candidate summaries, interview questions, scorecards, outreach messages and follow-up emails, but every output needs review. It can miss context, make assumptions or reflect bias when the input is incomplete or unclear. Treat it as a helper that sorts information and produces an opening draft; the call on whether a candidate matches the role, holds the right skills or should be hired belongs to a person who has weighed the evidence.