How to prepare a negotiation with AI: before, during, after

You have a negotiation that matters this week. A client contract, a supplier pushing up prices, a budget to defend at the next committee, a deal to close. And you will probably prepare for it the way you did last time: a few notes scribbled in your calendar, two arguments in your head. The rest, you will improvise on the day.

The problem is that negotiations are largely won before the first word. Whoever understands the other side's constraints and has already anticipated the objections walks in with a real edge. That preparation is exactly what almost nobody has the time to do properly. Which is precisely where AI earns its place.

Here is my method in three moves: before the negotiation for research and scenarios, during for listening and transcription, after for the debrief. With specific tasks to hand your AI assistant at each stage. And at the end, the tool that assists you live during the call and writes your debrief for you.

Before: put AI to work on the other side

The first move is to feed your AI assistant everything you have on the other party. Their website, recent press releases, the annual report if they are a listed company, your counterpart's LinkedIn profile, the history of your email exchanges. Paste all of it in and ask for a negotiation-focused summary: what their current priorities are, what constraints weigh on them, what carries value for them and costs little for you, what questions they are likely to throw at you.

A prompt that works: "I am negotiating the renewal of this contract with this company. Here is what I know about them and about my counterpart. Draw me a portrait of the person across the table: their probable goals, their room to manoeuvre, what they cannot accept, the arguments they will respond to." You get a prep sheet you would never have taken the time to build on your own.

Keep your judgment switched on. AI can get things wrong, or invent a detail that sounds true. Check the facts that matter before you build a strategy on them. Your experience is there for exactly this: to read and challenge what the machine hands you.

Build your scenarios and rehearse the objections

Second job: simulation. Ask your AI assistant to play the other side, using the information you just gathered. Tell it to become a difficult buyer, the kind who contests every line of your offer. Then run the negotiation in writing, exchange by exchange. You will quickly see where you buckle.

Then the objections. Ask for the ten most likely objections to your offer, each with a short answer and a follow-up question. Set your limits before the call too: your target, your walk-away point, the concessions you can give and what you ask in return. A concession with nothing in exchange is a gift. Write it all down, and AI lays it out in a table in a single request.

One last scenario to prepare: the one where it goes badly. What do you do if the person across the table threatens to walk? If they pull out a cheaper competitor? Having already played those moments through with AI completely changes how calm you stay on the day.

During: listen instead of taking notes

During the negotiation, your worst enemy is your own notepad. When you write, you stop listening. And the decisive information often slips out in half a sentence, buried in an answer: a budget to spend before a deadline, a boss pushing to sign fast. If you are busy noting the previous point, you miss it.

The fix is transcription. Have the call transcribed, with the other party's agreement, and free up all your attention for the conversation. Afterwards you get the exact words, the figures quoted, the commitments made, the hesitations. No more "I think he said something like": you have the verbatim.

Stay in active listening: reflect back what you hear, and let the silences do their work. It is counter-intuitive, but the less you talk, the more you learn about the other side's position.

After: the debrief almost nobody does

The call ends, you jump straight into the next one, and the negotiation is forgotten until the next round. A shame, because the debrief is where you actually get better. Give the transcript to your AI and ask for an honest read: which signals you missed, which concessions you made with nothing in return, the moments you talked when you should have stayed quiet, what the other side revealed without meaning to.

The debrief also sets up what comes next. Have it draft your summary email with the agreed points and the open ones, then list what needs to be locked down in writing. Negotiation after negotiation, you build your memory as a negotiator: the same tics come back, and so do the same mistakes. Seeing them in black and white is where fixing them starts.

An AI negotiation coach that runs this method live

Everything I just described, you can do with a general-purpose AI assistant and a bit of discipline. The weak spot is the during: you cannot ask ChatGPT for advice in the middle of a tense call. That is exactly the gap NegoBrain fills.

NegoBrain is the AI negotiation coach built by the AI x Leaders Lab. It transcribes your call live and surfaces tactical suggestions during the conversation, right when you need them. Once the call ends, it generates your post-negotiation debrief. The during and the after of the method, handled for you.

It covers sales negotiations, client meetings, supplier discussions and internal budget trade-offs. The tool is available now at negobrain.ai.

Read next

An executive AI coach who starts from your real files Everyone says AI is a priority. Almost nobody has a strategy. Shadow AI: your teams already use AI without you