AI doesn’t win games. Decisions do.

In my latest conversation with Mike Band (NFL Next Gen Stats), we explored how one of the most data-driven environments in the world actually uses AI — not as hype, but as a real decision-making system.

What stood out the most is something many companies still get wrong: It’s not about building better models. It’s about building better decisions.

A few key takeaways: • 80% of the work is data preprocessing — not the model itself • In real-time systems, latency vs accuracy is a constant trade-off • “Good process, bad outcome” is often the right decision • AI should support your judgment, not replace it • Context is everything — without it, even the best models fail

In the NFL, every decision is high-stakes and time-sensitive. And that’s exactly how leaders should start th ...