The $0 Fix That Makes Claude 10x Smarter For Your Business

Rob Collie spent thirty years inside the machine, and AI still scared him.

Funding stage: Bootstrapped. P3 Adaptive is a founder-led data, BI, and AI consultancy Rob co-founded in 2013, growing on client revenue with no announced outside capital.

He was there in 1996, a fresh Vanderbilt grad walking into Microsoft. He led the business intelligence work inside Excel. He was a founding engineer on what became Power BI. He wrote three books about it and, by his own account, sold more than 92,000 copies. If anybody had a right to shrug at the next wave, it was Rob Collie.

Then ChatGPT landed, and the guy whose entire career was telling computers what to do realized he ran the exact kind of company AI eats first: a data and BI consultancy built on code, scripts, and formulas. Easy mode for a language model. By late 2024 he calls it dread, not curiosity. He did not fight the fear, he followed it, and what he found on the other side was almost insultingly simple. AI success is not an AI problem. It is a data problem and a traditional software problem, the thing he already knew how to do.

That discovery became Fair Game, his fourth book, aimed squarely at the founder who is not getting briefed. If you run a Fortune 500, the hyperscalers are in your office with slide decks because they want to keep you. Everyone from mid-market down is left to guess while the labs sprint past them, too busy racing each other to slow down and teach anybody anything. Meanwhile the internet offers two flavors of fear: the hype side needs you scared so you buy the course, the doom side needs you scared so you click the next video. Rob's read is blunt. Both are engagement strategies, and neither one is running a business.

What he offers instead is a diagnosis. When you sit down with vanilla Claude or vanilla ChatGPT and try to do real, repeated business work, you hit a wall, and you blame yourself. You should not. That model knows a staggering amount about the Roman Empire and nothing about your company. It is a PhD in everything and a brand new hire at your business. So you spend an hour explaining your world to it, get 80% of the way there, grind on the last 20%, and then the chat ends and all that context evaporates. Rob has a name for the feeling: bot sitting.

The fix is not a better prompt. It is grounding. Rob's own hello-world moment was connecting Claude Desktop through MCP to a Notion knowledge base holding nothing more exotic than his company's brand guidelines. Suddenly the model was not just more accurate, it felt more confident, more willing to take a creative swing. Same model, different context. Ryan pushes on the entry point, and Rob's recommendation for anyone still on the sidelines is refreshingly unglamorous: open Claude Cowork, make a folder, and let a project accumulate its own memory. Motivation, history, guidelines, goals, all injected on hello. No engineering degree required.

There is also a moment in here that has nothing to do with AI and everything to do with why this show works. Ryan asks Rob what he has not forgiven himself for, and Rob does not dodge. He goes to high school, to filling holes in his soul with achievement, to making his own problems other people's burden, a habit he calls externalization and names as the root of most misbehavior in the world. He also gets vulnerable about cigars, rum, and a mouth cancer diagnosis during COVID that turned out to be a straightforward removal but still handed him a long look at his own mortality. His words: it is all yoga and run clubs now.

The through line, and the reason the book is called Fair Game, is that AI capability currently feels like an ivory tower priesthood. Rob's position is that this is a temporary information advantage cosplaying as expertise, and tearing down that tower is, in his words, one of his favorite things.