Prepare a salary negotiation case
Build a negotiation case from market data, your achievements and a target number — including fallback asks and responses to likely pushback.
Last updated
Fill in the blanks
Assembled in your browser — nothing you type is stored or sent anywhere.
Your prompt
Model-agnostic — works in Claude, ChatGPT, Gemini or any AI chat.
Why this prompt works
- →Separating the case into opening, evidence, fallbacks and pushback responses mirrors how a real negotiation conversation actually unfolds, so the output is a script to use rather than a essay to summarize on the spot.
- →Restricting supporting points to the achievements actually listed prevents the model from inflating the case with generic claims like "consistently exceeds expectations" that fall apart under a follow-up question.
- →Asking the model to flag a weak case instead of overstating it is the escape hatch here — a negotiation prompt with no honesty check will always find a way to sound confident, even when the underlying evidence doesn't support the ask.
Part by part
- states the ask directly without over-apologizing
- Names the specific failure mode of a first draft, which is usually three sentences of preamble before the number appears.
- drawn only from my achievements below
- Confines the case to things you can defend. An invented supporting point collapses under one follow-up question and takes the credible points with it.
- each tied to a concrete result rather than a general claim of hard work
- Effort is not negotiable currency; outcomes are. This is the difference between a case and a request.
- two fallback asks
- Prepared alternatives, so a no on the number does not end the conversation. Improvising a fallback in the room usually produces one you did not want.
- tell me honestly where the case is weak
- The most valuable instruction here. Knowing the weak point before the meeting is worth more than a stronger-sounding draft.
Common mistakes
Write me a script to ask for a raise.Produces confident-sounding prose with no evidence behind it, because none was supplied. The first specific question exposes that, and the position is weaker than opening with less.
How much should I ask for?The model has no reliable, current data on your market, location or company, and will produce a number that sounds researched. Bring your own data and use the prompt to build the argument around it.
Make my achievements sound more impressive.Inflation is the one thing that reliably damages a negotiation, because a claim that does not survive scrutiny casts doubt on the ones that would have. Understated and verifiable beats impressive and vague.
Questions
Where should the market data come from?
Salary surveys for your market and level, recruiter conversations, and public ranges from comparable postings. The model cannot supply this — anything it offers as a figure is a plausible-sounding guess.
Why prepare fallbacks at all?
Because a straight no ends a conversation you were not finished having. A prepared alternative — a review date, a title change, additional leave — keeps it going and often produces more than the original ask.
What if it tells me my case is weak?
That is the output doing its job. Knowing which point will not survive a follow-up lets you strengthen it, drop it, or adjust the target — all of which are better than discovering it live.
Should I name a number first?
Usually yes, if you have researched it, since the first number anchors the range. The exception is when you genuinely do not know the market, where asking for their range first costs you less than a badly placed anchor.