Why the Issues of the Day Still Trumph AI and Data Strategy

Meeting a deadline feels more urgent than checking a data field. Finalizing a contract feels more important than leaving a dataset in good order. But decisions will be made based on that data, argues Pim van Meer in this edition of Digitale Doolhof. And if the data isn’t right…

“How did you know this, Pim?” That was the question that led to this column. It’s as if I have a gift for prediction. As if I know in advance where a project will hit a snag, which contracts need attention, or which parts will spark debate down the line.

The answer is actually pretty boring. I think supply chain integration and consistent data are sexy. Call me crazy.

For more than fifteen years, I’ve been advocating for digital building models where information is stored in one place, follows the same conventions, and relies on a single source of truth. Not because data is interesting in and of itself, but because good data enables better decisions—both within a single project and across projects.

And yet, I see the same thing happen time and time again. We stray from our agreements, and that leads to mistakes.

The issues of the day prevail.

I get that, too. Meeting a deadline feels more urgent than checking a data field. Finalizing a contract feels more important than leaving a dataset in good order.

No one wakes up in the morning thinking, “Today I’m going to sabotage the data strategy.” But that’s ultimately what ends up happening.

What’s remarkable is that many organizations believe AI will eventually solve this. That messy data will sort itself out. That the technology is smarter than the process.

I’ve put quite a few hours into this by now. And yes, AI can do a lot, but not this.

Take a simple HSB wall, for example. AI can verify whether the coding is correct, whether the object is a wall, whether the material used is appropriate, and whether other properties support this conclusion. The more clues there are, the greater the likelihood that it is indeed an HSB wall.

But certainty? That only arises when someone verifies the information and corrects it where necessary. AI doesn’t turn that into truth. At best, AI turns it into a probability.

And that’s exactly where it goes wrong. Because if that final check disappears because the issues of the day take precedence, we won’t be working smarter. Instead, we’ll be automating our mistakes. Every incorrect assumption finds its way into dashboards, reports, benchmarks, and ultimately into decisions. That’s not digitization—that’s pollution on a massive scale.

Sometimes it feels like we’re attaching a jet engine to a shopping cart. Impressive technology, enormous computing power, amazing AI models… but it’s still just a shopping cart if the foundation isn’t right.

The strange thing is that this problem isn’t technical at all. It’s organizational. People don’t enter data carelessly because they don’t take their work seriously. They often simply don’t see why their piece of information matters. That one field in a model seems unimportant—until you explain that the same information will later be used for quality control, contract drafting, permits, operations, maintenance, AI analyses, and strategic decisions. Then, suddenly, a parameter becomes much more than just a parameter.

Then it comes down to a decision.

Perhaps that’s why we should talk less about AI strategies and much more about data strategies—about explaining individual information needs, about why that information is important, and about how all those little pieces ultimately come together to form a whole.

Because when every link in the chain understands which subsequent link depends on its information, quality improves naturally. Then, good data ceases to be an administrative burden and becomes a mark of craftsmanship.

And yes, even then, the issues of the day will win out from time to time. But not every day anymore.

Filing moment

Maybe I’m way off base. Maybe in five years, AI will be so good that it’ll effortlessly clean up our mess. That would be fantastic. But to be honest, I don’t believe that will happen.

I actually think the opposite is true. AI isn’t the solution to a poor data strategy. AI amplifies its consequences.

So before we invest heavily in increasingly powerful AI, I’d rather ask an uncomfortable question: Is our data strategy actually as mature as our AI strategy? Because otherwise, we’ll just keep doing what we’ve always done—only this time with a jet engine attached to a shopping cart.

Go ahead and share your counterargument. I’m genuinely curious to know where you think I’m wrong.