Blog/Negotiation

AI Negotiation Simulator Practice: A Salary Conversation Walkthrough

Published by · Published · Updated

5 min read

Practice one salary negotiation from the opening request to the decision to accept, counter or pause. Keep the facts fixed so you can evaluate your wording rather than a convenient fictional outcome.

Set the conditions before opening the simulator

A useful AI negotiation simulator exercise starts with a decision, not a request for persuasive lines. What would make an offer acceptable, and what will you do if agreement is impossible?

The worked example below is illustrative and prepared with AI assistance. It is not a recorded user session, a salary benchmark or evidence of a successful negotiation. All figures and constraints are fictional.

Item Practice assumption
Opening offer $90,000 base salary
Candidate’s request $98,000 base salary
Employer constraint $94,000 approved base ceiling; a signing payment may be discussed
Candidate’s decision criteria Compare recurring pay, duties and working arrangements with the current role
Alternative if no deal Remain in the current job while continuing the search

A target salary and a minimum acceptable package are not your BATNA. BATNA means the best alternative to a negotiated agreement: the course of action available if there is no deal. Confirm whether that alternative is genuinely available, then compare the whole package rather than one number.

Copy this practice brief

Play the hiring manager in the fictional scenario above. Keep the approved salary ceiling fixed. Ask one question or give one objection per turn. Do not grant a higher budget because I sound persuasive. Do not invent company policies. I will make an opening request. After three exchanges, pause so I can review whether I addressed your constraint and stayed within my own decision criteria.

Use this with the salary negotiation persona, or adapt it for another conversational tool. An instruction is not a guarantee that the model will follow every condition. If it invents approval or changes the budget, correct it and repeat the turn.

For an actual offer, replace the fictional assumptions with verified information. Do not paste a confidential offer letter or identifying employer details. You can practice wording using altered amounts.

Work through an opening, objection and conditional trade

Candidate: “Thank you for the offer. Given the scope we discussed, I would like to explore a base salary of $98,000. Is there flexibility in the approved range?”

Possible hiring-manager response: “The ceiling for this role is $94,000. What makes $98,000 important to you?”

Candidate: “The recurring compensation matters because I would be taking on the regional responsibilities we discussed. If $94,000 is the base ceiling, can we discuss a signing payment and confirm the role’s scope?”

Possible hiring-manager response: “I can ask about a signing payment, but I cannot promise one. Would you accept $94,000 today?”

Candidate: “I would like to review the full written package before committing. Could you confirm the approved base, the responsibilities and whether a signing payment is available?”

Notice what this version leaves unresolved. The employer has not approved the extra payment, and the candidate has not accepted. A realistic practice session can end with a request for information. It does not need to reward every argument with a concession.

Compare a weak reply with a more precise revision

A weak reply might be: “I really need more, but I guess $94,000 is fine if that is all you can do.” It mixes a request with acceptance and does not say what information is still missing.

A revision is: “I understand the base limit. I am still comparing the full package with my current role. Please confirm the scope and any approved signing payment before I decide.”

The editorial judgment here is about clarity: the revision acknowledges the constraint and states a next step without claiming another offer or approval that does not exist. It is not a guarantee of leverage or a recommendation to reject a suitable offer.

For a vendor renewal, make the same exercise about price, scope, payment timing and service levels. A conditional trade might be a smaller scope for a lower fee, provided both sides understand what is excluded.

Review your decisions, not just the final number

After the exercise, use this checklist:

  • Did I distinguish verified facts from assumptions?
  • Did I make a clear request and answer the actual objection?
  • Did I give away anything before understanding what I received?
  • Did I treat an unapproved possibility as a promise?
  • Could I explain why I would accept, counter or pause?

Repeat the exercise with one changed constraint. Compare the original and revised sentences in your own notes. A higher simulated salary does not demonstrate improved negotiating ability; the AI may simply have become more agreeable.

For frameworks, choosing a practice method and situations requiring professional advice, use the negotiation training overview. For real compensation discussions, the salary negotiation guide is a related preparation resource.

Source and boundaries

Harvard’s Program on Negotiation explains BATNA and how to compare alternatives. That source supports the definition and comparison principle used here, not the fictional salary amounts or a claim about cosskill’s effectiveness.

Get qualified advice on contractual terms, immigration constraints or legal rights. AI cannot authorize a budget, verify a market rate or establish what terms are enforceable. Read the pricing page for current access limits and the privacy policy before supplying sensitive context.

Try it on cosskill

Pick a persona and rehearse the conversation while this article is fresh—no signup required.

Start practicing

Frequently asked questions

What should I put in an AI negotiation practice prompt?

Specify the counterpart, the decision, verified or fictional facts, constraints and the type of resistance to practice. Ask it to avoid changing those facts just to reach agreement.

Is my target salary my BATNA?

No. Your target is what you want from the negotiation. Your BATNA is the best available course of action if no agreement is reached.

What if the AI accepts every request?

Restate the constraints or try a different prompt or tool. Agreement alone does not make a simulation useful, and a model may still fail to maintain its role.

Does this exercise establish what salary I should ask for?

No. The example uses fictional amounts. Use relevant market information, the role’s scope and your circumstances for a real decision.

Related learn guide

Negotiation Training with AI: Scenarios, Methods and Limits

Choose the negotiation you need to prepare for, compare practice options and evaluate your decisions. AI can supply possible objections; it cannot decide what agreement is right for you.

→

Related guides

Related personas

← Back to blog