ChatGPT Prompts for Recruiters (Sourcing, Outreach, and Screening)

Most "ChatGPT prompts for recruiters" lists are 50+ prompts sorted by category — sourcing, then screening, then offers — with no sense that they all belong to the same requisition. A real req doesn't jump around: you take the intake, you source, you reach out, you screen, you coordinate interviews, and you close, in that order, for the same role, over and over. This is the set that follows a single requisition end to end, built around one setup prompt so you're not re-explaining the role to ChatGPT at every stage.

The setup prompt (paste once per requisition)

You're helping me recruit for [job title] at [company/client], reporting to [hiring manager/team]. Level is [junior/mid/senior/exec], location/remote policy is [X], and the role is [one-line description]. My tone with candidates is [warm and consultative / direct and efficient / high-touch executive search]. Never invent a salary range, start date, or benefit detail I haven't given you — leave a [BRACKET] instead. Keep candidate-facing messages under 150 words unless I ask for more.

1. Turning a vague intake into a real requisition

The single biggest time-sink in recruiting is chasing a hiring manager who said "find me someone like Sarah but more senior" and calling that a job description.

Turn these hiring manager intake notes into a structured requisition: [paste raw notes/transcript]. Give me: (1) a clean job description with must-have vs. nice-to-have requirements clearly separated, (2) 5 clarifying questions I should still ask the hiring manager before I start sourcing, (3) a one-paragraph "ideal candidate" summary I can use to brief myself and any sourcers. Flag anything that sounds like an unrealistic combination (5 years of experience in a 2-year-old technology, for example).

The leveling-check version: "Compare this draft job description against [similar role/level elsewhere in the company, or a market benchmark I paste]. Flag anything where the requirements, title, or listed range don't match the actual level of the role."

2. Sourcing strings that actually surface candidates

Write 5 Boolean search strings for LinkedIn Recruiter/X-ray search to find candidates for this role: [paste the requisition]. Vary the strings — one broad, one narrow to must-haves only, one that swaps in adjacent job titles, one that targets a specific industry background, one that excludes obvious mismatches (recruiters, students, wrong seniority). Explain in one line what each string is optimized to find.

The adjacent-titles version: "This role is titled [title], but people who could do it well might carry a different title at other companies. Give me 8 adjacent or non-obvious job titles to search for, ranked by how likely they are to actually match."

3. Outreach that gets replies, not silence

Write a cold outreach message to a candidate on [LinkedIn/email] for this role: [paste requisition summary]. Reference something specific and plausible about their likely background (not an invented fact — use [paste a real detail from their profile]), state the role and one concrete reason it's a step up, and ask a low-commitment question (15-minute call, not "are you interested in new opportunities"). Under 100 words, no generic "I came across your profile" opener.

The referral outreach version: "Write a message to a candidate referred by [name]. Mention the referral naturally in the first line, note one specific thing [name] said about them, and ask for 15 minutes — warmer and less templated than cold outreach."

The re-engagement version: "Write a follow-up to a candidate who opened my last message but didn't reply, sent [X days] later. Reference the original message briefly, add one new piece of information (about the role or timeline) rather than just repeating myself, and make it easy to say 'not right now' without ending the relationship."

4. Screening questions that predict performance, not just check boxes

Write 8 phone-screen questions for this role: [paste requisition]. Mix competency-based questions (past behavior, STAR-style if useful for follow-up probing) with 2 questions that surface motivation and reason for looking — phrased so they don't lead the candidate toward the "right" answer. Include one question most screeners forget that catches a mismatch early (availability, compensation range, or logistics that would kill the process later).

The resume-gap version: "Here's a candidate's resume: [paste, redact contact info]. Give me 3 non-accusatory questions to understand a gap or short stint I see at [company/dates], phrased as genuine curiosity, not an interrogation."

5. Interview coordination that doesn't drop the ball

Write a scheduling email inviting [candidate name] to interview for [role] with [interviewer name(s)/panel]. Include the format (call/video/onsite), expected length, who they'll meet and each person's role, and one thing to prepare (a portfolio, a case study, nothing). Warm and organized, under 130 words.

The interviewer-brief version: "Write a one-page prep brief for [interviewer name] ahead of their interview with [candidate name]. Include: the candidate's background in 2 sentences, what this interviewer should specifically probe for (not overlapping with other panelists' focus areas: [list]), and one flag from earlier stages worth following up on, if any."

6. Rejections and offers that protect the relationship

Write a rejection email for a candidate who reached [stage — phone screen/onsite/final round] for [role]. Tone: respectful and warm, no generic "we had many qualified candidates" filler, one honest but kind line on the general reason if [I want to share one], and an invitation to apply again for future roles if genuinely true. Under 100 words.

The offer-extension version: "Write an offer call talking-point outline for [candidate name] — role: [title], the number is [$X, or 'BRACKET for me to fill in']. Structure: lead with the offer, not the process; state comp and start date clearly; leave room for their questions; one line on next steps and deadline to respond." If they counter, our salary negotiation email templates works from the employer side too — the same gratitude-value-data-number structure, just answering instead of asking.

The competing-offer response: "A candidate says they have a competing offer at [$X or 'higher, no number given']. Help me write a response that doesn't panic-match blindly — asks one clarifying question about what matters most to them (comp, role, growth, team), and if I do have room to move, frames it around fit, not just a bidding war."

What not to hand to ChatGPT

• The actual hiring decision. It can draft questions and summarize notes; it should never rank or recommend candidates for you to rubber-stamp — bias creeps in from imbalanced training data the same way it creeps into a tired human reviewer, and the legal accountability for the decision stays with you either way.
• Candidate PII at scale. Paste role details and de-identified summaries into prompts, not full resumes with names, contact details, and background-check results — treat any consumer AI tool the way you'd treat a public channel.
• Adverse-action and compliance language. Background-check adverse-action notices, accommodation conversations, and anything referencing a protected class need your legal/compliance team's exact wording, not a fluent-sounding AI draft — verify before anything with legal weight goes out.

Make it repeatable

Save your setup prompt per requisition, plus your go-to sourcing strings and interview-brief template, in a ChatGPT Project so every new req starts from full context instead of a blank page. Once a candidate is deep in your process, they're doing their own prep — see what they're likely rehearsing with in our ChatGPT prompts for interview prep, which uses the same one-setup-prompt structure from the candidate's side.

FAQ

Will candidates notice if my outreach or screening questions come from ChatGPT? Not if you personalize with a real detail from their background before sending — candidates notice generic, could-be-anyone messages, not whether AI helped draft the structure. The outreach prompts above are built around one real, specific detail for exactly this reason.

Can ChatGPT replace parts of an ATS? No — it drafts writing (outreach, screening questions, rejection emails), not tracking, workflow, or reporting. Keep candidate records, stage tracking, and compliance documentation in your actual ATS; use ChatGPT for the words that go around it.

Is it safe to paste a candidate's resume into ChatGPT? Redact names, contact details, and anything identifying first, the same discipline healthcare workers use — see ChatGPT prompts for nurses for the same de-identify-first habit applied to a stricter regulatory setting. Check your company's AI policy before pasting any real candidate data, even redacted.

How is this different from the generic "100 ChatGPT prompts for recruiters" lists? Most of those group prompts by category with no sense of sequence — 20 sourcing prompts, then 20 screening prompts, disconnected from any one requisition. This set follows a single req in the order it actually runs: intake, sourcing, outreach, screening, interview coordination, close — so each prompt assumes the context the previous stage already built.

Does this work for in-house recruiters, not just agency? Yes — in-house recruiters will lean harder on the intake and interviewer-brief prompts since they're working closely with the same hiring managers repeatedly; agency recruiters will get more mileage from the sourcing and outreach prompts across a wider set of clients. The structure holds either way.