The Prompt I Used To Prep For The Most Contentious Executive Meeting I've Had This Year — Full Breakdown

I'm a PMO director. Last week, I had to walk into a room full of my counterparts of senior leadership and my boss and tell them we couldn't do everything they were asking of our engineering teams.

That meeting had "this is going to get ugly" written all over it before it started.

Here's what I did before I walked in.

The setup

I'd already done the hard work on the data side — took our Microsoft Project plans, ran exports, had Claude build a dashboard that showed every active effort, the business value tied to each one, and a toggle system where leadership could literally turn projects on and off and watch engineering capacity shift in real time. Monte Carlo simulations for schedule risk. Delay scenarios. The whole thing.

But data alone doesn't win rooms where people have already made up their minds. I needed a meeting strategy — not just a presentation but I also had to get my mind right so I wasn't combative in the meeting.

So I used my app, RACEprompt, to build out a structured prompt. RACEprompt is built around Role / Action / Context / Expectation — it forces you to not just ask AI a vague question but to actually think through what you're trying to get out of it. Even though its my own app I was please with the clarifying multiple choice questions it gave me beofre the prompt such as What type of responses do you typically get from your leadership team? with selections such as "We need more resources" and "We need to be able to execute it all with the resources we have" and this was easy for me to curate this prompt specifically for our leadership team.

The prompt I ran:

Why this structure worked

A few things I want to point out:

The Role isn't just "you're a PM." It's describing the specific flavor of the problem — composure under sustained challenge. That changes the output. If you just say "you're a PM," you get generic PM advice.

The Context is doing heavy lifting. I'm telling the model the pattern of the room — leadership who defaults to "do it all." That context shaped every objection response Claude gave me. Without it, you get generic negotiation tips.

The Expectation is explicit about format and tone. "Avoid apologizing for constraints — frame them as strategic levers" — that one line changed the entire register of the output. Without it, the default is often a defensive or hedging tone.

What I got back

A full meeting script. Five sections. Opening framing with language designed to redirect the room before the first pushback even lands. Objection-by-objection responses with specific data callout structures (even placeholder fields for my own numbers). Three different scripts for when you're on the 5th version of the same objection — one direct, one stakes-framing, one authority clarification depending on what's actually driving the resistance.

I'm not going to claim the output was perfect out of the box — I customized it with my actual project names, real velocity data, and my own read of the room. But having that scaffold meant I spent 90 minutes refining something instead of 3 days building something from scratch.

The outcome

Complete alignment. Toughest decisions got made. Projects got deprioritized with leadership owning that call, not fighting it. I walked out with a signed-off tier structure.

The prompt is the thing most people skip. They either ask too vaguely or dump everything in without thinking about what kind of output they actually need. RACE forces that discipline.

Happy to answer questions on the prompt structure or the dashboard build if anyone wants to dig in.