ChatGPT for Recruiters: Can It Prepare a Counter-Offer Call?

Yes, ChatGPT can help a recruiter prepare for a counter-offer conversation with a candidate and with the hiring client. It drafts questions, scripts, objection responses and call notes from a de-identified brief you write. It cannot supply salary data, read the candidate's mood or decide for anyone, and personal details stay out of the chat.

Most of the value arrives before the counter-offer. The work that protects a placement happens at offer stage, when you find out why the candidate is leaving and what their employer could change. Chat GPT for recruiters is most useful at that stage, because the work is structured thinking from your own notes. The call itself, and the judgement on it, stay yours. All the salary figures below are invented for illustration.

Follow me on Instagram@sagnikteaches

Where Chat GPT for recruiters helps on a counter-offer call

A counter-offer call is short, emotional and time-pressured. The candidate has just handed in notice, their manager has reacted, and the client may be hearing about it from you in the same hour. Recruiters who improvise tend to do one of two things: argue the money, or panic. Preparation gives you a structure to fall back on. It is also a fairness matter. A candidate deserves an honest conversation about their own reasons, not a pitch built to save your fee.

Connect on LinkedInSagnik Bhattacharya

A preparation routine needs four documents, each short enough to write in five to ten minutes: a case brief, a candidate call plan, a client call plan and a log note. The steps below build each one in order from a single running example.

Subscribe on YouTube@sagnikteaches

Step 1: a de-identified case brief

The brief is the only thing ChatGPT sees, so make it complete without making it identifiable. Use labels, round figures that you have been told directly, and no employer names. Here is a filled-in illustration for a sales manager role at a wholesale distributor client.

Case brief (de-identified)
Role: sales manager, wholesale distributor, team of 6, reports to the managing director.
Candidate A. Told me: current base 70,000, typical bonus about 6,000, notice 4 weeks.
Reasons for leaving, in their words: no route to a team lead role; commute; "I stopped learning in the second year".
Client offer: base 78,000, bonus target 8,000, start in 6 weeks, client's approved base range 75,000 to 82,000.
Status: offer accepted verbally on Monday; resignation handed in Wednesday; current employer has hinted at a counter.
Agency fee: 20% of first-year base, payable on start.
My aims: keep the candidate's decision honest and informed; give the client early warning; no pressure tactics.

Check what is absent: no name, no employer, no phone number, no email. Nothing in the brief tells a stranger who this is. If you cannot write the brief without a name, write it with the name and keep it in your ATS, and give the chat only the version above.

Step 2: prepare the candidate call

The goal of this call is to understand, not to win. Ask the tool for questions grouped by what you are trying to learn, then pick the five you would genuinely ask.

Using the case brief, prepare a 15-minute call plan for a recruiter speaking to Candidate A after their employer hints at a counter-offer.
1. Eight open questions that explore their reasons for leaving, in order, with the reason behind each.
2. Three questions that test whether a counter-offer would change those reasons.
3. A neutral summary I can read back to confirm I have understood.
Rules: no pressure, no statistics, no claims about what usually happens to people who accept counter-offers, no salary figures other than the ones in the brief.

A useful answer might open with "What was it that first made you look?" and "If your employer matched the base salary tomorrow, which of your three reasons would still be true?" It should include a read-back such as "So the money is not the issue; the missing route to team lead and the commute are." Compare that with the brief. If it has added a reason Candidate A never gave, delete the question. A tool is good at generating questions and poor at knowing which of them fit this person, so read each one aloud.

The sums that frame the conversation

Put numbers beside the story. Candidate A's current package is $70,000 plus about $6,000, so $76,000. The client's offer is $78,000 plus a target of $8,000, so $86,000, a difference of $10,000. Suppose the employer counters at a $76,000 base. The counter package, with the same bonus, is $82,000, which is $4,000 below the offer, or 4.7 per cent of $86,000. Your fee is 20 per cent of the offer base, $15,600. The gap in pay is smaller than most candidates expect, and the fee at stake is much bigger than the gap, which is a reminder that your interest is not neutral. Say so to the candidate, and keep the focus on their own reasons. Ask ChatGPT to check the arithmetic if you want, and verify it on a calculator, because a script built on a wrong figure is worse than no script.

Step 3: raise the counter-offer before it happens

The best counter-offer call is one in which nothing surprising happens. At offer stage, ask the candidate what they would say if their employer matched the salary and made a promise. Many recruiters leave that question until the resignation has been handed in, when it is harder. Ask ChatGPT for three phrasings that sound like curiosity, not suspicion:

• "Before you hand in notice, what do you think your manager will say, and how will you feel hearing it?"
• "If they offered the same money, which of the reasons you gave me would still be there?"
• "Is there anything your manager could say that would make you stay, and is it something they could do next month, or only promise?"

Record the answers in the ATS. If the candidate says "nothing would change my mind", that is useful to know and to tell the client, in the candidate's own words and only with their permission. If they say "if they gave me the team lead role I would stay", you have a very different case from one where the reason was commute.

Step 4: prepare the client call

Clients have different worries. A hiring manager wants to know whether the start date is safe, whether to offer more, and whether to keep other candidates warm. Your job is to give early, honest information, not to push them into raising the offer to protect your fee. Here is a prompt for the call plan.

Prepare a call plan for a recruiter speaking to the hiring client about a candidate who has resigned and may receive a counter-offer.
Include: (a) a three-sentence opening that gives the facts without drama; (b) four questions to ask the client, including whether they have flexibility inside their approved range of 75,000 to 82,000 base and what non-salary changes they could make; (c) how to describe the risk honestly without exaggerating; (d) what to say if the client asks me to push the candidate; (e) two things I must not say.
Use only the figures in the brief.

A good plan reminds you of the line between informing and pressing. For the client the facts might be: "Candidate A handed in notice on Wednesday. The current employer has hinted at a counter. The candidate's reasons were mostly about progression and commute, not pay. I will update you by Friday after speaking to them again." The questions might include "Is there a way to describe the progression path in writing?" The "must not say" list should include any promise that the candidate will definitely start, and any statement about the candidate's current employer's finances or intentions. If the client has a range, the plan can note that the offer sits at $78,000 in a $75,000 to $82,000 range, so the room is $4,000 and any move needs the client's approval. Keep that number from the client, not from the model.

Step 5: objection handling you can rehearse

Ask for objections as a table, then rehearse the answers aloud in your own words. The table below is an illustration of the shape, and each answer must be adjusted to your candidate and client.

None of these responses states a statistic or hints that the candidate will regret staying. The aim is to make the decision clearer for the candidate, whichever way it goes. If a candidate decides to stay, you thank them, tell the client at once, and move to the next candidate on the shortlist. A recruiter who keeps the relationship honest is the one the candidate calls next time.

A sample output, and the line you would cut

To see how review works in practice, take an illustrative reply to the client-call prompt above. The first three lines are usable. The fourth is the kind of line to delete.

Opening: "Candidate A resigned on Wednesday and their employer has signalled a counter-offer. I will update you by Friday."
Question 1: "Is there flexibility inside your approved range on base, or on the bonus target?"
Question 2: "Could you put the progression path in writing for the candidate?"
Line 4: "Market data suggests this role typically pays 85,000 to 90,000, so you may want to move up."

Line 4 has no source and no basis in the brief, and it would hand the client a reason to blame you if the number is wrong. Replace it with a question: "Do you want me to check your range against a salary survey before the next call?" That is honest, it is within your control, and it keeps the client's decision with the client. The rule from the brief applies throughout: the model may suggest what to ask, but any number must come from a named source.

A second case: a promotion, not a pay rise

Not every counter-offer is about money. Suppose Candidate B is an operations manager moving to a small manufacturer. Their employer replies to the resignation with a promotion to head of operations and a base rise from $64,000 to $72,000, against the new client's offer of $70,000. The counter is $2,000 higher, so the money argument has reversed and a prepared script about pay would sound strange. The call plan has to change: ask what the promotion involves, who confirmed it, whether the title and pay are in writing, and whether the reasons for leaving, such as a poor relationship with the director, have changed. If they have not, the extra $2,000 may not matter. If they have, you may be speaking to someone who ought to stay, and an honest recruiter says so. Brief the model with the changed facts and ask for questions about written confirmation and what the role really includes. You will recognise this as the same method with different inputs, which is the point of keeping the brief short.

When the client asks you to push

Sometimes the client says "just get them to confirm today". Here a script is no substitute for a position. You can say, truthfully, that you will ask the candidate for a time they will give a firm answer, and that pressing for an immediate answer is likely to produce a worse result. Ask the tool for three ways of saying that to a client who is under pressure from their own manager, and choose the one that sounds like you. A reply such as "I'll ask for a decision by noon tomorrow, and I'll tell you at once if I hear anything sooner" gives the client a time and respects the candidate. Avoid any line that promises the candidate will start. If your terms of business address withdrawal or a candidate not starting, check them before the call, since they set out what the client can expect, and neither a chatbot nor this tutorial can interpret them for you.

A four-week notice period, day by day

A notice period gives you time to keep the placement steady. For a four-week period, an illustrative plan looks like this. Ask the tool to draft the messages from your brief, and send them in your own words.

• Day of resignation: call the candidate, then the client. Log both.
• Days 2 to 3: a short check-in. Ask how the conversation with the manager went and whether anything has been offered in writing.
• End of week 1: if the candidate is staying firm, send the client a brief update and confirm the start date and any paperwork.
• Weeks 2 and 3: one call each week, focused on the new role rather than the old one: introductions, first-day logistics, a meeting with the team if the client agrees.
• Final week: confirm the start details, and be ready for a late counter. Ask in advance whether they have had a second conversation with their employer.

The tutorial on automating candidate follow-up shows how to schedule the routine messages, so your own calls can be about the conversation.

After the outcome: a five-minute debrief

Whether the candidate accepted the counter or started the new job, take five minutes to learn from it. Write down the reasons the candidate gave at offer stage, what the employer offered, what they decided, and what you would do differently. Over time, look for patterns in your own cases: how often a salary match changed the outcome, and how often a promise without a date did. These are your own figures, and you can quote them to a client without any outside statistic. If you want a second pair of eyes, give ChatGPT a batch of anonymised case notes with the instruction "list the three most common reasons given and the number of cases for each", then count them yourself. The recruiter who builds this record ends up with real evidence, instead of borrowed claims.

What never to ask ChatGPT for

Salary data. Ask it "what does a sales manager earn?" and you will get a fluent number with no source. Seeing such a figure in a script is the clearest sign to stop and verify. Use the client's band, the candidate's own statement, and a licensed survey if you subscribe to one.

Counter-offer statistics. Many articles state how often counter-offers fail. I have not seen a source I can verify for those claims, so I do not use them, and I would not put one in a script. If a candidate asks, say you do not have reliable figures, and talk about their own reasons.

Pressure lines. If a draft includes urgency you cannot justify ("the client will withdraw tomorrow"), delete it. It is a risk to your reputation and, if untrue, to your standing with the candidate.

Anything that treats people differently by who they are. If a script changes the pitch or the offer advice by age, sex, ethnicity, family situation or another personal characteristic, remove it. Offers should rest on the role, the client's range and the candidate's experience.

Step 6: log the outcome in the ATS

Your ATS is the system of record, and a good note is short, factual and written in your own words. A template helps. This is an illustrative filled-in note for the example, written from the recruiter's own notes and tidied by ChatGPT from the de-identified version.

Date: Thursday. Contacts: candidate (call 20 min), client hiring manager (call 10 min).
Candidate: resigned Wednesday; employer hinted at a counter. Reasons given: progression, commute. Said a salary match alone would not change this. Agreed to speak again Friday 10:00.
Client: informed of resignation and possible counter. Approved base range unchanged (75,000 to 82,000). No change to offer. Will keep second-choice candidate warm.
Next step: call candidate Friday; update client by 12:00.
Outcome (add later): accepted counter / started / withdrew.

Paste the final text into the ATS yourself, adding the identifying details there and nowhere else. Fill in the outcome when it is known. After a dozen cases you will have a record of which reasons for leaving were genuinely about money and which were not, and that is worth more than any script. The related tutorial on candidate summaries that clients read covers the document that precedes all this, and AI note takers and candidate consent is the one to read before recording any call.

Privacy: what stays out of the chat

Keep these out of ChatGPT: names, current and previous employers, phone numbers, emails, home areas, health or family details, salary history tied to a named person, and anything the candidate asked you to keep in confidence. Business plans such as ChatGPT Business describe not training on business content by default, at $20 per user per month billed annually or $25 monthly with a two-seat minimum. That is helpful for data handling, but it does not change what you owe candidates. Your agency's data-protection policy and your clients' terms of business come first, and the questions in questions to ask before buying AI that touches client data work for candidates too.

A two-minute test before you send or say anything

1. Every figure in the script is one the candidate or client gave you, or one you calculated and checked.
2. No sentence suggests a consequence you cannot support.
3. The candidate's reasons are in their words, not yours.
4. The client has been told what you know, not what you hope.
5. The note in the ATS says what happened, not what you wish had happened.

The short version: use ChatGPT to prepare the structure, and keep the numbers, the judgement and the conversation yourself. If you can explain why each line in your script is there, you are ready for the call.

Questions recruiters ask about AI and counter-offers

Can ChatGPT tell me what salary to offer or counter with?

It cannot know your market, and a figure from it is a guess that sounds confident. Use the client's approved band, what the candidate told you about their current package, and any salary survey data you license. Ask the model to calculate gaps and total packages from numbers you supply, and to draft questions, never to supply the numbers themselves.

Is it safe to paste candidate details into ChatGPT?

Keep names, current employer, contact details and anything that identifies the person out. Use a label such as Candidate A and describe the role generically. A business plan that does not train on your content by default helps, but it does not remove your obligations to candidates, so the safest habit is to give the tool only what it needs for the task.

Should I quote counter-offer statistics to candidates?

Avoid them unless you can name the source. Claims about how many people who accept a counter-offer leave within a year circulate widely and are rarely backed by a source you can check. A candidate who senses a sales line will trust you less. Ask them what changed in their current job after they handed in notice instead.

Where should the outcome be recorded?

In your applicant tracking system, as a short structured note: the date, who was spoken to, what was offered, the candidate's stated reasons, the outcome, and the next step. Write it yourself or from your own notes. If you use ChatGPT to tidy the wording, give it the anonymised version and paste the result into the ATS yourself.

Further reads

• How to Evaluate the AI Features in Your Recruitment Software — Check what your ATS already offers before adding another tool.
• Is AI CV Screening Fair? Bias Checks Every Recruiter Should Run — Fairness checks that belong in every AI-assisted hiring step.
• AI CV Screening for Recruitment Agencies: Setup and Safeguards — The earlier stage of the same pipeline: screening safely.
• AI for Recruitment Business Development: Winning New Clients — The client-side work that comes before any placement.
• How Much Time Does AI Save a Recruitment Agency per Role? — A realistic view of time saved per vacancy.
• Questions to Ask a Recruitment Agency About Its AI Screening — What clients ask agencies about AI, and how to answer.
• ChatGPT Prompts for Recruiters: Intake, Outreach and Shortlists — Eight ChatGPT prompts for job intake, outreach messages and shortlist write-ups, with redaction rules that keep candidate data out of the chat.
• AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.

Sources: OpenAI's ChatGPT Business pricing page (seat prices, two-user minimum, no training on business data by default); all salary and package figures in the examples are invented for illustration; checked 4 November 2026.