Speed Isn’t Strategy: Why AI Doesn’t Equal Expertise

I want to say something that might sound strange coming from someone running a software company built around AI: I don’t think you can implicitly trust it.

Everyone reading this has seen an AI hallucination. You’ve watched a model insist it’s right when you know for a fact it isn’t, and you’ve had to push back two or three times before it finally acknowledges the mistake. That experience should tell you something important. AI does not equal expertise. Using it does not automatically mean you’re doing something better. It’s a technology that gives you access to a whole range of new capability, and what you do with that capability is entirely up to you.

Speed vs. Velocity

There’s an analogy from physics I keep coming back to, because I think it captures the moment we’re in better than anything else I’ve heard: the difference between speed and velocity.

AI has given all of us enormous speed. You can analyze extreme amounts of data far faster than you could a few years ago. You can compress cycle times on innovation, especially in areas like coding where the gains have been dramatic. But speed, on its own, has no direction. Velocity has a direction. And right now, a lot of firms have all this new speed without necessarily knowing whether they’re pointed the right way. You could be racing at full speed in the wrong direction and not even realize it until you’ve already burned the time and the resources.

That’s where the human still comes in, and I don’t think that changes anytime soon. Applying direction to speed requires judgment. It requires skilled people who understand your business well enough to know whether an output is actually right, useful, and appropriate for the situation, not just fast and confident-sounding. Without that layer, you’re not deriving a benefit from AI. You’re just compounding your risk faster than you used to.

Building Expertise

This connects directly to something I get asked a lot, which is what actually builds real human expertise in an AI-saturated environment. My answer is that it hasn’t changed: expertise is built by solving problems. People have been solving problems for thousands of years, and that fundamental mechanism hasn’t gone anywhere. What’s different today isn’t how expertise gets built, it’s what happens to that expertise after it’s built.

For decades, the expertise a firm generated every single day mostly sat still. It lived in someone’s head, or in a document nobody opened again, or in a data silo that never talked to the rest of the business. A brilliant piece of problem solving from three years ago might as well not have existed for a colleague facing a similar problem today, because there was no way to access it.

What’s changed is that we now have the ability to index that expertise, share it, and apply it consistently across a business. That’s the real shift. It’s not that AI is creating expertise for you. It’s that AI is finally giving expertise a way to compound instead of sitting isolated.

Shifting Your Approach

I think that distinction matters enormously for how firms should be thinking about AI adoption. If your AI strategy is oriented around getting faster answers, you’re optimizing for speed. If your AI strategy is oriented around capturing, connecting, and reapplying the expertise your people already have, you’re building velocity. Those are not the same project, and they don’t require the same investment.

My advice to any firm leader evaluating AI right now is to stop asking whether a tool makes your team faster, because almost everything on the market will, at least in the short term. Ask instead whether it’s helping direct that speed toward outcomes that are actually better for your clients, and whether it’s helping your organization retain and compound the expertise your people are building every single day. If it isn’t doing both, you have speed without velocity, and that’s a much riskier place to be than most firms realize.