Uber spent its whole annual AI budget in the first four months of 2026. And when asked what they got for it, the company’s own COO admitted they can’t yet draw a clear line between what they’re spending on AI tools and the useful features actually shipping.
Sit with that. Uber has hundreds of engineers, a massive budget, and access to tools most businesses won’t see for years. They spent the whole AI line item by April, and the people running it can’t point to a clean return.
If that’s the situation at that scale, what does it mean for a 50-person company in Ventura County?
It means AI spending and AI results are two different things, and most businesses confuse the two.
Plenty of businesses are paying for AI tools right now, ChatGPT seats, Copilot licenses, and AI-powered CRMs. The tools are running. The invoices are going out. But ask the owner what’s actually different, how much time it’s saving, what it’s producing, whether it’s paid for itself, and the answer gets vague fast.
That’s not a knock on AI. Uber’s CEO is a believer; he’s said the right tools make his engineers far more productive. The problem isn’t the technology. It’s spending without a target.
For a small business, the risk is actually lower if you’re intentional. You’re not deploying AI across a global network. You’re looking at one or two tools and one or two workflows. The feedback loop is faster, and you can kill what isn’t working in 30 days without a boardroom fight.
But you have to go in with a measurable goal. Not “we want to use AI more.” Something like: “Cut the time we spend on client onboarding documents by 50% before the end of Q2.”
That’s a test. Something you can pass or fail. That’s how you avoid Uber’s problem at your scale, for a fraction of the cost.
AI works. “We’re spending on AI” is not a strategy. A goal is.
Thinking about adding an AI tool, but not sure it’ll pay off? Book a 20-minute consultation, and we’ll pressure-test it before you spend a dime.
