Most candidates start negotiating when the offer lands. That is already too late. The work that wins money happens days earlier: pinning a target range to verified data, setting a walk-away you will actually hold, and rehearsing the recruiter screen so your number sounds rehearsed by your career, not by a panic.
This matters because the upside is real. A 2025 Pew Research survey found roughly two-thirds of US candidates who negotiated their starting pay got a higher offer, and analyses of recent negotiation studies put the average raise from a successful ask near 18.8%. People who skip the conversation leave real money on the table — and because every future raise compounds off that starting number, the lifetime cost runs well into six figures. Twelve prep prompts below turn an AI assistant into a research analyst and sparring partner so you walk into the call already decided.
TL;DR
- Negotiation is won in prep, not in the call. Build three numbers (anchor, target, walk-away) before any recruiter conversation.
- Anchor to verified data: levels.fyi “actuals” (offers signed in the last ~6 months), the posted range now legally required in 16 US states plus DC, then 2-3 peer or recruiter data points.
- Use a thinking-tier model for the math and a fast model for script drafts. Claude Opus 4.7, GPT-5.5 Thinking, and Gemini 3.1 Pro all handle the reasoning; any free tier handles the rewrites.
- Never let the model invent your track record or market numbers. Feed it your real data, then make it audit your draft for apology language and hedging.
Who this is for
Candidates preparing for offer conversations, career switchers calibrating expectations, and anyone who has accepted a first offer without a single counter. If you have negotiated before, jump to the counter-offer and multi-offer scripts.
When these prompts will not help
- You have no real data. AI cannot conjure a credible range from nothing. Pull comps first, then prompt.
- You have no alternative and a hard deadline. Leverage drives outcomes; a weak BATNA means anchoring lower, not louder.
- You ask the model to “be honest about my chances” without your actual record. It will confabulate a confidence level you have not earned.
Which AI tool to use
You do not need a paid plan for any of this, but the model tier changes the quality of the reasoning steps.
| Task | Recommended model (June 2026) | Why |
|---|---|---|
| Range math, BATNA analysis, level mapping | Claude Opus 4.7, GPT-5.5 Thinking, or Gemini 3.1 Pro | Multi-step reasoning, holds your full data set in context |
| Script drafts, tone audits, rewrites | Any free tier (GPT-5.5, Sonnet 4.6, Gemini 3.1 Pro) | Fast, low-stakes drafting |
| Long offer letters or comp doc dumps | Claude (1M-token context) or Gemini 3.1 Pro (1M) | Big paste-ins without truncation |
Free tiers (ChatGPT Free, Claude Free, Google AI Pro at $19.99/mo for heavier use) cover a single negotiation comfortably. Save the reasoning steps for the thinking modes; everything else runs anywhere.
Prompt anatomy
Every salary-prep prompt should carry six elements. Skip any one and the output drifts generic.
- Role: candidate, hiring manager, or recruiter — name the persona the AI plays.
- Context: target role, company, level, your background, geography.
- Goal: one deliverable — an analysis, a script, an answer, a plan.
- Constraints: word count, banned phrases, must-include facts.
- Tone: confident, curious, measured — pick 2-3 anchors.
- Examples: paste 1-2 of your past answers so it copies your voice.
What this set covers
- Range research before you apply
- Target / anchor / walk-away math
- Recruiter screen scripts
- Counter-offer and multi-offer scripts
- “We don’t negotiate” rebuttals and internal-promotion cases
12 copy-ready prompt templates
Replace each [placeholder] with your own value before sending.
1. Three-number target
Role: [role]. Geography: [city/metro]. Years experience: [years].
Build my 3-number framework: (a) aspirational anchor (top of the
plausible range), (b) target (most likely outcome), (c) walk-away
(minimum I'd accept). For each, state the data it rests on. I'll
paste comps next; flag where you're guessing vs. citing my data.2. Range research synthesis
Here are my data points: [paste levels.fyi actuals, posted range,
peer numbers]. Synthesise one base / equity / bonus range. Flag
outliers. Tell me which level (IC4 / IC5 / Staff) my profile maps
to and why. Don't take the median blindly — weight recent, verified
offers higher.3. BATNA articulation
My best alternative if this offer falls through: [batna]. Pressure-
test it: (1) Is it a concrete offer or a hope? (2) What specifically
does it give me in cash, equity, and timeline? (3) How much deadline
pressure am I under? Be blunt — a weak BATNA means I anchor lower.4. Recruiter “what are you looking for?” script
A recruiter asks: "What are you targeting?" Write a 3-sentence reply
that (a) gives a range anchored at the high end of market, (b) ties
to my expertise, not "I deserve", and (c) leaves room to flex. Do
not commit to a single number. My context: [role, level, comps].5. Anchor justification
I want to anchor at [anchor]. Write a 2-sentence justification built
on my track record plus market data — never "I researched salaries".
Confident, not entitled. My strongest proof points: [2-3 results].6. Counter-offer script
Offer: [offer]. My target: [target]. Write a written counter that
(1) thanks them, (2) reiterates genuine interest, (3) presents the
counter with a one-line rationale, (4) leaves room for back-and-forth.
Six sentences or fewer. Warm, specific, not apologetic.7. Multi-offer leverage
I hold offer A: [offerA] and competing offer B: [offerB]. Write a
script to raise B without sounding mercenary: a factual mention, a
direct ask whether A can match or improve, and a clear decision
deadline. Keep it under 5 sentences.8. Non-cash levers
Beyond base salary, what should I negotiate? List 8 non-cash levers
(sign-on bonus, equity refresh, vacation, remote allowance, learning
budget, title, start date, role scope). For each: one line on how to
ask and what's realistic to get at a [company size/stage].9. “We don’t negotiate” rebuttal
The recruiter says "we have a fixed salary band". Script a reply that
(1) accepts the band exists, (2) explores non-band levers (sign-on,
start date, review timing), and (3) verifies the band actually maps
to my level. Don't accept the first "no" — stay warm.10. Internal promotion negotiation
I'm negotiating an internal promotion to [target level]. Build a case
doc: (1) impact stories with metrics, (2) market peer comp for the new
level, (3) what I'll own next, (4) the cost to the team if I leave.
This is case-building, not a hard ask — keep the tone collaborative.11. Tone audit
My draft counter: [draft]. Audit it for (a) apology language ("I hate
to ask"), (b) hedging ("if possible", "just"), (c) over-explaining.
Rewrite it tighter — confident but warm — and show me the before/after
on each fix.12. Decision-day checklist
I have an offer due in 48 hours: [offer details]. Walk me through
(1) sleeping on it before deciding, (2) comparing to my BATNA, (3)
comparing to my walk-away, (4) any open data I need before signing,
(5) one final ask to make before I accept. Output as a checklist.Where to anchor your numbers in 2026
The prompts are only as good as the data you feed them. Three sources, weighted in this order:
- levels.fyi actuals. It surfaces verified offers signed in the last ~6 months for a specific level at a specific company, with city-level filtering. Treat this as your primary anchor for tech roles.
- The legally posted range. As of June 2026, 16 US states plus Washington, DC require employers to post a good-faith pay range, either in the listing or on request (California, Colorado, New York, Washington, Illinois, Massachusetts, Minnesota, New Jersey, Maryland, Hawaii, Maine, Vermont, and more). That posted range is your floor, not your target.
- 2-3 peer or recruiter data points. Cross-check the above against people in the role. Glassdoor and self-reported aggregates skew low and stale — use them only as a sanity check, never the anchor.
Feed all three into prompt 2 and let the model reconcile them. Then validate anything it cites: if it can’t tell you the source, treat the number as a guess.
Common mistakes
- Prompting without specific context (company, role, level, geo) — the output is generic and useless in the room.
- Asking the model to “be honest” about your odds without giving it your real record. It invents a verdict.
- Reusing one answer across companies. Recruiters in the same market compare notes.
- No tone anchor, so the script lands flat or, worse, sounds entitled.
- Trusting cited numbers without checking. AI fabricates levels, dates, and comp figures with total confidence.
- Shipping the first draft. First drafts read AI-flavoured; the third does not.
Push the results further
- Paste two real examples of your own writing so the model copies your voice, not a template.
- Make the AI play the recruiter first. Weak answers expose themselves fast under follow-up questions.
- Write three drafts and send the third. Read each one aloud — anything you stumble over, cut.
- Keep a personal bank of phrasings that worked. Reuse beats re-inventing under pressure.
- Have someone currently in the role review your range before the call.
FAQ
- Can recruiters tell an answer was AI-written?: Yes, when it is generic. Specific stories, real numbers, and your own cadence are the antidote. Use AI to draft and audit, not to think for you.
- How much research is enough?: 60-90 minutes of focused comp research for an important offer. Beyond that, returns diminish fast — the data does not get more accurate, you just get more anxious.
- When should I start salary research?: Before you apply. A negotiation that begins after the offer is already on the back foot.
- Should I trust levels.fyi and Glassdoor numbers?: Trust levels.fyi actuals (recent, verified, level-specific) as your anchor. Treat Glassdoor and similar aggregates as a low-confidence sanity check only — they skew old and low.
- What if the company says they have a fixed band?: Use prompt 9. Verify the band maps to your level, then negotiate the non-band levers — sign-on, start date, equity refresh, review timing. The first “no” is rarely the last word.
- Will an AI invent salary numbers?: It will, confidently. Always make it cite the source for any figure, and validate against your own data before you quote it in the room.