The FP&A Guy Network

The episode contains several valuable insights about AI's impact on data roles and the strategic importance of semantic models, but also includes substantial amounts of tangential discussion (lightning strike, Mac purchasing decision, family anecdotes) that dilute the substance-to-filler ratio. The core insights about job displacement, the shift to semantic models as AI infrastructure, and the reframing of data as fundamentally an AI problem are solid but not densely packed throughout.

“whatever our existing job is today, like, it's kind of already gone”

“AI really is a data problem. It's about getting the right information to the LLM, um, at the right time”

While the core argument that AI requires better data infrastructure and semantic models is sensible, it is not particularly fresh or contrarian in 2024. The framing of data professionals needing to 'create more value' is familiar industry advice. The analogy about the sun's light taking 8 minutes to reach us is creative but does not yield novel strategic insight. The discussion of Claude/Cowork is more product review than original thinking.

“whatever our existing job is today, like, it's kind of already gone. In the same way that like, if something happened at the sun right at this moment, we wouldn't know for eight minutes”

“the new jobs for people who are wired like us, who are in this data and analytics space, the new jobs are probably better than our old jobs”

Rob Collie is a genuinely credible guest with 13 years at Microsoft as a founding engineer on Power BI and deep technical expertise in data modeling. He has founded and runs a relevant consulting firm (P3 Adaptive), has published multiple influential books, and has direct ongoing experience building AI systems into client work. He is neither a pure theorist nor a generic 'thought leader.' His background and current role make him well-qualified to speak on this topic with authority.

“Rob Cawley is the founder and CEO of P3 Adaptive. You can see it on his shirt there if you're watching online, a national Microsoft analytics and AI consulting firm. During his 13 years at Microsoft, Rob led the BI focused capabilities in Excel and was subsequently one of the founding engineers on Power BI”

“I just really started a little over a year ago, like really diving really super deep into AI because not. I needed to know what our company was going to look like. Like it was, it, it was my responsibility as the leader of this company to plot us a course”

The episode lacks concrete data, client examples, metrics, or measurable outcomes from AI implementations. Rob discusses his experience using Claude/Cowork for editing and his daughter's job search, but provides no specifics about how P3 Adaptive has deployed AI for clients, no revenue impacts, no quantified efficiency gains, and no case studies. The discussion of semantic models is conceptual rather than backed by hard numbers or named implementations.

“There's this like, like, uh, you know, these silly things. It's almost like my whole career had been building to that moment”

“I could send Eddie three paragraphs and ask him what he thinks. I'd be like, this is, you see what I'm trying to do here, Eddie? But it's not kind of coming out right”

The host (Paul Barnhurst) asks reasonable open-ended questions but rarely pushes back, challenges claims, or goes deep on specific points. Much of the episode drifts into personal anecdotes about lightning strikes, Mac purchases, and family stories without the host steering back to substantive ground. There is little evidence of sharp follow-up questions or productive disagreement. The conversation feels more like a friendly chat than a probing interview designed to extract actionable insights.

“So you were one of the founding engineers of Power bi. Obviously, you worked on Power Pivot. Would love to learn a little bit about that experience, how it came about”

“how are you using AI in your company and how do you think it's going to change your business?”