Across many weekends in any given year, in cities from São Paulo to Monaco, a small group of people sit in front of dozens of screens and make decisions that determine whether our drivers win or finish last.

Energy strategy. Tyre temperature. When to deploy maximum power in Attack Mode. The teams have minutes. Sometimes seconds. They are making split-second decisions based on multiple data stacks.

A few years ago, the obvious answer to that problem was to hire more analysts. Today it’s to deploy AI agents alongside the humans who are already there. We’ve done both, and I can tell you the second is harder, more interesting and considerably more useful.

I want to make a case that runs against the grain of how many large companies are approaching AI. The race to embed generative AI into business has produced a strange pattern: enormous investment in tools that are technically impressive but operationally underwhelming. Chatbots nobody uses. Copilots switched off after a quarter. Innovation labs producing demos that never reach production.

This is a deployment problem.

In motorsport, you can’t escape it because the feedback loop is brutal. If a system doesn’t work, the........

© Sports Business Journal