Building Your AI Negotiation System

Dear Negotiation Explorer,

Welcome to issue 31 of our NegoAI series.

Over the last six issues, we've covered a lot of ground. Metaprompting. Prompt frameworks. Output assessment. Context engineering. Memory templates. Each one a building block.

But building blocks aren't a system until someone assembles them.

This week, I want to show you what that looks like — not in theory, but through someone actually doing it. Meet Sarah.

Sarah Chen is a Chief Procurement Officer at a major pharmaceutical company. She's preparing for one of the biggest negotiations of her year: renegotiating an $18.5 million consulting contract with a top-tier management consulting firm.

The stakes are real. The contract has been running for a decade. Her internal review shows the firm is underdelivering — implementation rates well below industry benchmarks, junior consultants doing work that senior partners were supposed to lead. She has competitive bids from three rival firms, all significantly cheaper. And her own team has started using AI tools that match the consulting firm's output on routine strategic work.

Sarah knows negotiation. She's been doing this for years. But she's never used AI systematically to prepare. She's done what most professionals do — opened ChatGPT, typed a question, gotten a generic answer, closed the tab.

This time, she decides to build the system first.

Sarah starts where we started in issue 5 — with a folder structure. But instead of jumping to the full eight-folder system from issue 6, she keeps it simple. Six folders. Clean enough to build in five minutes, complete enough to support everything she needs.

Knowledge Base — where her negotiation expertise will live as a structured document. Not there yet. That's step 2.

Cases — where the deal materials go. Sarah drops in her internal performance review of the consulting firm, the competitive proposals from rival firms, the current contract terms, and her notes on what's worked and what hasn't over the past decade. Everything relevant to this negotiation, in one place.

Outputs — where she'll save what AI produces for her. Analysis, scenarios, strategy options. She'll iterate, and she'll want to compare versions.

Prompts — where she saves the prompts that work. The metaprompt from issue 1. The assessment prompts from issue 3. The KB builder prompt she's about to use. A reusable library.

Memory — where the learning lives. After this negotiation, she'll write a deal debrief. What worked, what didn't, what she'd do differently. Over time, this folder captures patterns across all her deals — the compounding intelligence we discussed in issue 6.

Capstone — her live negotiation workspace. The deal brief, the session state, the active prep. This is where the work happens.

Five minutes. Six folders. Sarah's filing cabinet is ready.

But a filing cabinet with no documents is just empty folders. The most important document — the one that turns generic AI into her AI — doesn't exist yet.

This is where most people stall. They think building a knowledge base means sitting down with a blank page and writing out everything they know about negotiation. That sounds like a weekend project. So they never start.

Sarah doesn't have a weekend. She has a negotiation next week.

Here's what she does instead: she gathers what she already has.

Every professional has more negotiation knowledge captured in existing documents than they realize. Sarah pulls together her preparation checklists from past deals. Email threads where she explained her approach to colleagues. Notes from negotiation training she attended. A slide deck she presented on vendor management strategy. A post-mortem from a deal that went sideways last year.

She doesn't curate. She doesn't organize. She just dumps everything into a pile.

Then she opens Claude, uploads the whole pile, and runs this prompt:

The AI reads everything Sarah uploaded. It finds her philosophy scattered across three different emails. It pulls preparation frameworks from her training slides. It extracts her anchoring approach from a deal post-mortem she'd almost forgotten about. It identifies her industry-specific patterns from the vendor management deck.

Then it tells her what's missing. For Sarah, areas 4 (Value Creation) and 7 (Difficult Situations) were thin — her documents covered what she did but not why. The AI asked five targeted questions. Sarah answered in a few sentences each.

Twenty minutes total. Not a weekend project — a focused session.

The output: a single structured document covering eight areas of her negotiation expertise. Her philosophy, her methods, her instincts, her industry knowledge — organized so that any AI tool can reference it.

Sarah saves it to her Knowledge Base folder. She uploads it to her Claude Project. From now on, every conversation inside that project starts with her expertise loaded. The AI doesn't guess how she negotiates. It knows.

Before the Knowledge Base, Sarah's AI conversations went like this: explain the deal, explain her approach, explain what she's tried, ask a question, get a generic answer.

After: she attaches the KB, attaches the case materials from her Cases folder, and asks. The AI already knows her philosophy, her frameworks, her style. It gives her analysis grounded in how she thinks — not how a generic negotiation textbook thinks.

The difference isn't subtle. It's the difference between a consultant who just met you and one who's worked with you for years.

And the system grows. After the McKinsey negotiation, Sarah will write a deal debrief and save it to Memory. The lessons feed back into the KB. The next deal starts smarter. The deal after that, smarter still.

That's the compounding effect we talked about in issue 6 — but now you've seen someone actually build the foundation for it.

Build yours.

Step 1 (5 minutes): Create the six folders. Google Drive, OneDrive, your desktop — wherever works.

Step 2 (5 minutes): Gather your existing materials. Don't curate — dump. Checklists, emails, training notes, slide decks, post-mortems. Whatever reflects how you negotiate.

Step 3 (15 minutes): Run the KB builder prompt above. Upload your materials, let the AI extract, answer the gap questions.

Save the output to your Knowledge Base folder. If you're using a Claude Project or ChatGPT Project, upload it there too.

You now have the two things that make everything else work: a system that organizes, and a knowledge base that personalizes. Every AI session from here forward starts from you, not from zero.

Next week: Sarah prepares for the meeting.

The Second Brain is built. The Knowledge Base is loaded. Now Sarah uses the system to actually prepare — analyzing the consulting firm's position, identifying her leverage points, developing her strategy, and walking into the room ready.

One scenario. Every step. The complete preparation workflow.

Questions? Reply directly — I read every response.

#Negotiation #NegotiationSkills #NegoAI #AI #KnowledgeBase #SecondBrain #ChatGPT #Claude #NegotiationPreparation