Your negotiation changes after the first round

Dear Negotiation Explorer,

In week seventeen of building the NegoAI platform, we added ways to keep working on the preparation after the report arrives. You can question the specialists, bring in another one and review the changes they propose, then record a meeting and prepare for the next.

Brama, my partner in Claude Code, built these changes, with Codex reviewing the work.

Here is a fictional walkthrough, using the report and specialist answers from our PharmaCrest teaching case. Sunita buys consulting for PharmaCrest. She is about to renegotiate an $18.5 million annual agreement with McKinsey. The NegoAI app has given her the preparation report.

It says to buy a smaller service and make McKinsey's senior people more accountable for it. The proposed opening is $13 million. Another firm has quoted $11.5 million, so before she uses that opening she wants to know why she should pay more.

This is where self-organization comes in. Sunita asks the team a question, and the app chooses which specialist should answer. It tells her why, and she can choose someone else.

She asks, "Why should I open with $13 million rather than match the cheaper quotation?" The app chooses Deepak, the negotiation expert.

A price match could leave her buying the same work, with the same problems. What she wants is to stop paying for routine analysis her own team can do, and to get senior judgment when she needs it. The cheaper quotation still has to be checked against that.

When she asks about her purchasing strategy and dependence on McKinsey, the app chooses Kraljic, the buyer expert. PharmaCrest could move some of the work today, but it isn't ready to replace the whole relationship. A quotation doesn't tell you the other team is available, can take over in time or covers the same work.

Her next question is, "How should I compare alternative suppliers and their quotes?" The app chooses Porter, who helps find and assess suppliers. He sets out what she needs to check: what work is included and which senior people are available, then what it costs and what it takes to change provider.

All three answer inside the same case, so Sunita doesn't have to start again in a new conversation.

Sunita now needs the accountability proposal written into contract terms. After the first report the app can bring in another specialist, and here it recommends Macneil, the contract specialist. She adds his analysis to the preparation. He has the report and the case material, but no signed agreement, so he proposes terms to negotiate. He doesn't claim to know what rights Sunita already has.

He makes the accountability proposal concrete. Someone has to decide whether a milestone has been met. The contract has to say what happens when the work fails that test, and when payment is due. And naming a senior partner doesn't say how much of that person's time is reserved, or who takes over if they're unavailable.

Macneil's analysis comes back as proposed changes, each showing the old passage next to the new wording, with the reason. Sunita can accept or reject each one. The earlier report stays available.

Now a fact changes. In our fictional case, Deloitte clarifies that its $11.5 million quotation covers routine research and presentation work only. It excludes implementation support and high-stakes strategic advice, including material acquisition work.

The price is the same, but it covers different work.

Sunita enters the correction, and the app sends it to Kraljic, Deepak and Macneil, who each check which of their conclusions depend on it. Their proposed changes reach the summary and the short reference sheet too.

One answer Sunita had prepared for McKinsey changes. She can now admit that Deloitte quoted for different work, then take David back to the service she wants McKinsey to price. The report also drops the suggestion to test Deloitte's implementation support, because that quotation doesn't include it.

Sunita accepts the changes. Deloitte's price still helps her compare the routine work, but it doesn't tell her what McKinsey's smaller service should cost or give her permission to spend more. The opening stays at $13 million.

Von Neumann plays the other side. Here he played David, the McKinsey partner, and Codex wrote Sunita's replies. The whole exchange was a simulation.

David was willing to talk about a smaller service. Before promising senior coverage, he wanted a decision calendar and named PharmaCrest owners. For him, urgent review of work already in the scope was a different thing from a new assignment arriving at short notice.

Sunita offered a predictable calendar instead of a higher fee. In return she asked for a costed proposal showing the senior capacity McKinsey could actually reserve.

We stopped before any agreement and asked for the debrief. It was precise about what they had: a basis for a proposal, with price and staffing still open. A friendly conversation hadn't reserved any of the people she needed.

This is multi-round preparation. After a real meeting, Sunita can write down what happened or attach the transcript. The app uses that record to prepare the next round, and she can rehearse from there.

We gave it our rehearsal transcript and said plainly that it was a simulation. Deepak built the next round around the question David had left open, which was how much senior capacity McKinsey could reserve. Kahneman, the behavioral expert, left his profile of David unchanged, because an AI playing David says nothing new about the actual person.

The second rehearsal didn't start the first time. It asked whether we wanted to resume, though we had already given it the roles and the difficulty, so we had to tell it again. After that David carried on from the first meeting: urgent review only for decisions already in the scope, and no agreement yet on the fee or on which senior people.

On Thursday, 26 November, I'm teaching the fourth free Lightning Lesson in the series: Prepare your next negotiation with AI in 15 minutes, live.

I'll use the NegoAI app to work through the fictional PharmaCrest negotiation in this issue. We'll question the recommendation and see how the report changes when we correct a fact. Then I'll rehearse the conversation using the revised preparation, so you can see what still needs resolving before the real meeting.

The session lasts 45 minutes and starts at 18:30 Rome (CET), 12:30 New York (EST) and 23:00 Mumbai (IST). A recording will be available afterward.

Reserve your free place on Maven.

Put the report you prepared from next to what actually happened. Separate what the other side proposed from what they committed to. Correct the facts that changed, then follow each correction into the advice that depended on it.

You can do that with a colleague. Before you practice again, make sure you're both working from the revised preparation.

What changed in your last negotiation meeting that should have changed your preparation for the next one?

Just hit reply and tell me. I read and answer every one.

#Negotiation #NegotiationSkills #NegoAI #AI #AIAgents #AugmentedNegotiator