Practicing with AI: rehearse the conversation you dread, risk-free

You don't get good at difficult conversations by reading chapters: you get good by practicing. The problem is that these conversations are rare, high-stakes, and impossible to "rehearse" in real life. This is exactly where conversational AI changes the game: it offers a training partner that's available, patient, and consequence-free, with whom you can rehearse as many times as needed.

This section is the program's signature angle: crossing the psychology of communication with AI to turn an intimidating skill into a trainable one.

Three complementary uses of AI

flowchart TD
    A[AI for practice] --> B[Simulate the other person<br/>role-play]
    A --> C[Get feedback<br/>on your wording]
    A --> D[Prepare<br/>anticipate reactions]

Simulate the other person: the AI plays the difficult party (the defensive colleague, the unhappy client, the rushed boss) while you lead the conversation. Get feedback: the AI analyzes your sentences and spots judgments, accusations, or safety breaches. Prepare: the AI helps you anticipate objections and separate facts, story and emotion.

Prompt 1 — The realistic role-play

The secret to a good simulation is a precise brief: give the AI the role, its emotional state, and the instruction to stay in character.

Prompt: "You're playing Marc, a colleague who consistently delivers his work late. You're defensive and tend to justify it with your workload. I'm going to give you feedback. Stay in character, respond realistically (sometimes annoyed, sometimes shut down), and don't make it easy for me. Start by waiting for me to speak."

Lead the conversation. Let the AI resist. You'll discover in real time whether your opening creates safety or instantly puts your counterpart on the defensive.

Prompt 2 — Structured feedback

After the exchange (or by pasting a draft message), ask for a targeted analysis rather than a vague opinion.

Prompt: "Here's how I opened the conversation: '[your sentence]'. Analyze it on three criteria: (1) is it an observable fact or a judgment? (2) is my intention clear and perceived as well-meaning? (3) do I leave an opening for the other person? For each criterion, say what works, what risks triggering defensiveness, and propose a rewording."

This grid — fact/judgment, intention, opening — is exactly the principles of the earlier chapters. The AI becomes a mirror that applies them to your words.

Prompt 3 — Anticipating reactions

Prompt: "I have to tell my team the training budget is cut this year. List the five most likely emotional reactions, and for each, suggest a response that validates the emotion without promising what I can't deliver."

Preparing the likely reactions reduces the risk of an amygdala hijack on the day: what's anticipated surprises us less, and surprising us less means keeping the prefrontal cortex online.

The limits to keep in mind

AI is an excellent training ground, but it's not a human being. Three precautions:

AI training sets the stage; it doesn't replace real experience, nor the courage to hold the conversation for real.

Practical exercise

Take a real difficult conversation that awaits you. Run Prompt 1 to simulate it three times in a row, changing your opening each attempt. Then move to Prompt 2 to have your best version analyzed. Note the wording that creates the most safety — that's the one you'll carry into the real conversation.

Summary

AI offers a risk-free training partner for a skill you can't rehearse in real life. Three uses: simulate the difficult counterpart (a well-briefed role-play), get structured feedback on your wording (fact/judgment, intention, opening), and anticipate reactions to limit the amygdala hijack on the day. Mind the limits: agreeableness, no real emotion, occasionally inaccurate advice — AI prepares you, it doesn't replace the real conversation.