We're making the same mistake with AI that we made with machines.

Walk into almost any organization right now and you’ll see the same thing.

Someone on the team has found an AI tool they like. They’ve figured out how to get it to produce something that looks like a finished work product, a report, a summary, an analysis, a client-facing document. It takes a fraction of the time it used to. So they use it. Regularly. Without much review.

The output looks polished. It reads confidently. And somewhere in the third paragraph there’s an error, a number that’s slightly off, a conclusion that doesn’t follow, a recommendation that doesn’t fit this client’s actual situation, that nobody caught because the document looked done.

The AI did exactly what it was asked. The human did exactly what required the least effort. And the result was something that looked like work but wasn’t.

I’ve watched this happen across industries. Not because people are careless. Because nobody designed how it should work.

Thirty years ago, we built machines and forgot about the people using them

I spent the early part of my career walking manufacturing floors, watching workers contort themselves around equipment that had been designed for throughput, not humans. The machine determined the posture. The process determined the reach. The workflow determined the cognitive load. And somewhere in the design process, the actual human doing the work had become an afterthought.

The injuries that followed weren’t accidents. They were the entirely predictable result of a system that was never designed with the human in mind.

The fix wasn’t complicated. Move the work to the person. Design the environment around human capability. Stop asking people to adapt to systems and start designing systems that fit people.

We called it ergonomics. The principle was simple: fit the work to the worker, not the worker to the work.

Most organizations never fully learned it. And now we’re watching the same mistake happen again, faster, at larger scale, with higher stakes.

AI has arrived in your organization. You probably didn’t design that.

Ask yourself: does your organization have a real framework for how staff use AI? Not a policy. Not a reminder to “use it responsibly.” A genuine, designed approach to how AI tools fit into the work, which tasks they’re suited for, which they aren’t, how outputs get reviewed, and what standard of judgment is expected before something gets used?

Most organizations don’t. AI showed up, staff started using it, and leaders quietly hoped for the best.

The result is the same pattern I watched play out on manufacturing floors for three decades. The tool is determining the workflow. The AI is setting the standard. And the human, the most important variable in the entire system, has become the afterthought again.

The problem isn’t the AI. It’s the absence of design.

I’m not anti-AI. I use it daily. I think it’s transformative when it’s used well.

But “used well” requires something most organizations haven’t built: the deliberate design of how human judgment and AI capability work together. Which decisions belong to the human. Where AI accelerates without replacing. How you build a team that treats AI as a thinking partner rather than an answer machine.

That’s not a technology problem. It’s a human systems problem. And it’s precisely the kind of problem that requires someone who understands how people actually think, decide, and work, not just how the tools function.

I’ve spent 30 years studying human-system interaction. AI didn’t change the question I’ve been asking. It just made it more urgent.

The two-inch fix exists here too

In manufacturing, the fix was rarely dramatic. Move the bin. Raise the surface. Redesign the reach. Small, deliberate, structural changes that repositioned the human for success.

The same principle applies to AI adoption. You don’t need a six-month transformation program. You need a clear, deliberate framework for how your people engage with AI outputs, one that builds judgment rather than bypasses it.

What gets reviewed before it leaves the building? What questions do your people ask before they trust an output? How do you build a team that uses AI to think harder, not less?

Small. Deliberate. Structural.

This is what I mean by Deliberate Intelligence: the practice of intentionally designing how human thinking and AI capability work together inside your organization. Not leaving it to chance. Not hoping staff figure it out. Designing it the way you’d design any other critical system.

The organizations that get this right won’t just avoid the errors. They’ll outperform the ones that didn’t bother, the same way a well-designed workplace outperforms one where workers are fighting the equipment all day.

Everyone has AI. Not everyone thinks with it.

The question isn’t whether your people are using AI.

The question is whether anyone has designed how.