What a Digital Camera Taught Me About AI

Every time we walk into a business to talk about AI, there's a moment you can feel. No one says anything, but you can see it on people's faces. They're thinking, "Is this the thing that's going to replace us?" And honestly… that's a fair question.

Because a lot of what gets said about AI doesn't exactly help. Automation. Efficiency. Reducing labour. If you're running a line or managing a shift, that doesn't exactly sound like good news.

I remember working at Toyota in the early 90s. To create standardised work, we had to draw everything by hand, with a pencil. Parts, cars, sequences — the lot. For every 60 seconds of work, you could end up with 10 Work Element Sheets, each one needing a drawing. It was a massive job.

This was before production had properly started. We were building the process from scratch, and everyone was involved in defining their own role, supported by our Japanese trainers. But it wasn't just about the tools. We were being taught the fundamentals — what good flow looks like, why standardisation matters, how to see waste properly, and the behaviours that sit underneath it all. Respect for people. Ownership of your work. Improving the process, not blaming the person.

By the time the line started moving, people didn't just know the job. They understood it.

A couple of years later, the first digital camera arrived. It felt like a big step forward. No more drawing, no more trying to sketch something that vaguely looked like a car. The pencils could be made redundant. And to be fair… that wasn't a bad thing. Some of our drawings were closer to Edvard Munch than engineering diagrams.

At first, one camera was shared between about 30 people. Then, like most things, it evolved. One person took the pictures. One person created the standardised work. More efficient. More consistent. It all sounded like progress.

I remember talking about this with my Japanese trainer. I said, "At least now we can actually see what the pictures are meant to show." He agreed. Then he asked, "Have you noticed the other outcome?"

I hadn't.

He pointed out that people weren't as involved anymore. There was less ownership, less connection to the work. Even the team spirit had dropped a bit. Then he left me to think about it.

It didn't take long to decide what to do. We changed it. We trained everyone to use the camera and went back to building standardised work as a team.

Almost straight away, engagement went up. Ownership came back. And the number of improvement ideas increased. Because the people doing the job — the ones who really understood it — were back involved in improving it.

That experience has stuck with me, because what we're seeing with AI today isn't that different. The tools are better, obviously. But the risk is the same.

If you introduce technology in a way that takes people out of the process, you don't just lose involvement — you lose improvement.

There's something else happening now as well. A lot of organisations are rushing straight into AI. Tools first. Platforms first. Capability later.

But if the fundamentals aren't there, it doesn't work.

If people don't understand how the process should flow, what "good" looks like, how to identify waste, or how to improve safely and consistently, then AI doesn't solve the problem. It just analyses it faster. Or worse, it gives you a very detailed view of a process that nobody really understands.

AI can't replace Lean thinking, good behaviours, or a culture of continuous improvement. It depends on them.

We've seen this over and over again. The moment people think that improving something might cost them their job, Kaizen stops. Straight away.

Why would anyone point out waste, fix inefficiency, or improve a process if the end result is they're no longer needed?

You don't get ideas. You don't get engagement. You get silence. Or you get, "Yeah, everything's fine." And we all know what that means.

If AI is introduced in the wrong way, it just makes that worse. It stops being a tool and becomes a threat.

From what we've seen, the people closest to the process already know where the problems are. They don't need AI to tell them that. What they don't always have is the time, the data, or the ability to move quickly.

That's where AI helps. Not instead of them. Alongside them.

The shift happens when people see it in their own environment. Not in a presentation. On their line. In their process.

Watching video and spotting waste they hadn't seen before. Turning audits into something real, not just paperwork done at the last minute. Solving problems in minutes that used to take days.

That's when it changes. From "This is being done to us" to "This actually helps."

This is why we don't start with AI. We start with the work. Understanding the process. Seeing the waste. Improving how things run. Building the fundamentals.

Because if that's not in place, AI doesn't fix anything. It just makes the mess faster. And no one needs a faster version of a bad process.

Roles are changing — that's true. But not in the way people think.

We're not seeing people disappear. We're seeing better decisions, faster problem solving, and more confident teams. The job doesn't go away. It gets better. And usually a bit less frustrating.

That lesson from Toyota still applies.

Technology can make things quicker. But if it takes away ownership, involvement, and pride in the work, you lose something important.

And if you skip the fundamentals and rush straight to the tools, you lose even more.

The future isn't people or AI.

It's people, grounded in Lean thinking, using AI to do their jobs better than they could before.

— Richard Pounder, Founder, CLS