My Big Miss: AI Wasn't Supposed to Happen This Fast

We knew AI would change construction. What I didn't understand was how quickly it would become part of everything else.

by Will Johnson

A couple of weeks ago, I went back and looked at one of the first articles we published on Buildernauts. In February 2024, we wrote a piece called Revolutionizing Construction: Top 5 Emerging Technologies Enhancing Project Delivery. This one caught my eye in particular because it's actually the most viewed article ever on our site.

Naturally, I wanted to see how well it had aged. There's nothing quite like going back and checking your old predictions to find out how smart you used to think you were.

The premise was pretty simple. What technologies were going to have the biggest impact on construction over the next several years?

We picked five: Building Information Modeling, drones, augmented and virtual reality, the Internet of Things and modular construction.

Two and a half years later, I don't think we did terribly.

BIM has continued becoming more embedded in how projects get designed and built. Drones have gone from being a cool way to take pictures of a jobsite to becoming a serious data-collection tool. IoT largely happened, although I'm not sure anybody really calls it IoT anymore. Equipment, tools, cameras and all sorts of devices are just connected now. Modular and industrialized construction continue to move forward, although probably not as quickly as we imagined.

AR/VR? I'm still waiting.

So, overall, I was feeling pretty good about us.mAnd then I noticed the fairly enormous thing that wasn't on the list... Artificial intelligence. Whoops.

We didn't miss AI. We missed what AI was going to become.

Here's where it gets a little worse for me. It wasn't like we didn't know AI existed. Eleven days after publishing that article, Buildernauts published an entire piece about the developing role of AI in construction technology.

So I can't even use ignorance as my defense. We knew AI mattered. What I don't think I understood—and I'm not sure most of us really did—was how quickly and how deeply it was going to work its way into everything else.

Forget construction for a minute and just think about how differently many of us work today compared with February 2024. I use AI to research things, summarize documents, analyze data, prepare for meetings, create images and, yes, help me research and write articles like this one. Two and a half years ago, I wasn't doing most of that.

There is something mildly amusing about using AI to assist in writing an article about how badly I underestimated AI, but here we are.

Increasingly, I'm not even necessarily making a conscious decision that I'm going to “use AI.” It's just becoming part of the software, the search and the workflow.

That's the part I missed. I was thinking about AI the same way I was thinking about drones or BIM or modular construction—as another category of technology.

I'm not sure that's what it is anymore.

Construction is obviously a little different

Now, before we get carried away, construction has certainly not become some sort of AI-native industry overnight. Not even close.

A 2025 RICS survey of more than 2,200 construction professionals found that 45% said their organizations hadn't implemented AI at all, while another 34% were still in the early pilot stage. Just under 12% reported regularly using it within specific processes, and fewer than 1% said AI was fully embedded across their organizations.

That sounds a lot more like the construction industry I know.

But the same survey also found that people see significant potential for AI in things like progress monitoring, scheduling, resource optimization, contract and document review, risk management and cost management. Roughly two-thirds of project managers and quantity surveying/construction professionals believe AI will eventually help them deliver greater value.

So there is a pretty big gap between where we are and where people think we're going. That's not particularly unusual in construction technology. What is unusual is how quickly the gap appears to be closing.

And there is another interesting signal: a lot of money is starting to chase it.

Follow the money

Trunk Tools has now raised $70 million, including a $40 million Series B last year, around an AI platform designed to make sense of the enormous amount of information generated by construction projects—drawings, specs, RFIs, schedules, submittals and everything else.

Document Crunch went after a bit of a different problem: using AI to understand construction contracts, documents, compliance and project risk. After raising significant venture funding of its own, Trimble acquired the company earlier this year for roughly $250 million.

Then there's Digs, which is much closer to the residential homebuilding world I spend most of my time in. Digs just raised a $25.3 million Series A led by Builders FirstSource, along with a five-year strategic partnership aimed at expanding its AI-powered workflows across the homebuilding and homeowner experience.

I'm not bringing those companies up because I've done some exhaustive evaluation of their products. I haven't. What interests me is how different the problems are that they're attacking: documents, contracts, risk, project information, preconstruction, homebuilder workflows and homeowner handoff.

That's when this starts to feel bigger than another ConTech category.

For the last 15 or 20 years, a huge portion of construction technology has really been about digitization. We took things we used to do on paper, in binders, in somebody's head or on a whiteboard and started putting them into software.

We digitized the plans, contracts, schedules, photos, RFIs, equipment and project records.

In the process, we created an enormous amount of data. The problem is that having a bunch of data and actually being able to do something useful with it aren't necessarily the same thing.

I recently interviewed Steve Dell'Orto, founder of ConCntric, for Buildernauts. Steve spent decades on the builder side before becoming a technology founder, and we ended up talking quite a bit about this exact issue.

At one point he said:

“Pre AI, they were still coming to that. Now they're like, holy crap, I really need to do something with my data big time.”

That line stuck with me because I think he's right.

For years, digitizing construction information was valuable because it made that information easier to store, share and access. Now we're entering a world where the quality and structure of that information may determine what AI can actually do with it.

It can find something buried in thousands of pages of project documents. It can connect information from different places. It can summarize it, interpret it, identify risk, answer questions and increasingly take some action based on it.

That's a pretty meaningful jump from simply moving another construction process from a clipboard to an iPad. And it's a much bigger idea than the neat little “AI” box I apparently wanted to put it in back in 2024.

The people are still the hard part

There is another reason I keep coming back to the conversations I've had for Buildernauts. No matter what technology we start out talking about, eventually the conversation seems to get back to people.

I've written about this before. I don't think construction has a technology problem nearly as much as it has an adoption problem. There is plenty of good technology available today. The harder part is getting people to trust it, understand why it matters and actually change the way they work because of it.

AI isn't going to magically make that problem disappear.

When I talked with Contech Recruiter , Romey Oulton, she brought up the fear side of AI, particularly around jobs.

“People hear AI and immediately think jobs are disappearing. I don't see it that way. With the labor shortages our industry faces, AI has the potential to be a game changer. People should embrace it rather than fear it.”

Steve had a slightly less diplomatic version of the same argument.

I asked him what he would say to a superintendent or project manager who was worried that AI might eventually replace them. His response was pretty direct:

“Your bigger threat to losing your job is not being proficient with AI... because other people are.”

I tend to agree with both of them.

That doesn't mean AI isn't going to eliminate jobs. I'm sure it will eliminate some, and I think pretending otherwise is probably just as silly as assuming it's going to eliminate everybody.

But construction already has a labor problem. We need more people. We need more skilled people. And we're asking the people we do have to manage increasingly complicated projects with an almost ridiculous amount of information coming at them.

What happens if AI makes one really good estimator dramatically more productive? Or helps a superintendent spend less time digging through documents and more time actually managing the project? Or allows somebody in preconstruction to evaluate alternatives that previously would have taken three people a week to work through?

That's where this starts getting really interesting to me.

Maybe the important dividing line over the next few years isn't between jobs that

AI can replace and jobs that it can't. Maybe it's between people who figure out how to work with it and people who don't.

Maybe AI never belonged on our list

Which brings me back to that original article.

Maybe our mistake wasn't simply forgetting to put AI in the top five. Maybe the bigger mistake was thinking about emerging construction technology as five separate categories in the first place.

Look at the list now. BIM can be paired with AI. Drone and reality-capture data can be analyzed by AI. Connected equipment generates data that AI can interpret. Robotics and industrialized construction can be powered by it. Scheduling, estimating, project controls and document management are all being affected by it.

AI isn't really sitting next to those technologies. It's starting to work its way through them.

And I think that's why this feels different from some of the other technology cycles we've seen.

Eventually, we may stop talking about “AI construction software” altogether.

Nobody talks about “cloud-based construction technology” with the same excitement they did 10 or 15 years ago. Of course it's cloud-based. We just expect it to be.

I wonder if AI ends up the same way.

Which, if it does, makes leaving it off a list of emerging technologies in 2024 both more understandable and considerably funnier.

At least that's the explanation I'm going with.

Two and a half years isn't very long

That's probably the part of this that I find most fascinating.

February 2024 doesn't feel like ancient history. Yet think about how you researched something then, how you prepared for a difficult meeting, how you summarized a 50-page document, analyzed a dataset, created an image or got yourself unstuck when you were staring at a blank page trying to write something.

For a lot of us, at least some of those answers are already completely different.

Steve made this point during our conversation too. AI has technically been around for a long time, and the technology people will rightly point that out. But for most normal people going about their jobs every day? This version of AI has really only been part of our lives for a couple of years.

And suddenly everybody is trying to adjust. Companies are trying to figure out what it means. Universities are trying to figure out what they should teach.

Employees are wondering what it means for their jobs. Software companies are rebuilding products around it, sometimes while their customers are still figuring out what questions they should even be asking.

That's a lot of change in a very short period of time.

Construction will probably continue moving more slowly than some other industries. We have jobsites, physical assets, fragmented data, thin margins, safety considerations and decades of deeply ingrained workflows to contend with. None of those things disappear because somebody added AI to a piece of software.

But AI doesn't necessarily require the industry to reinvent itself overnight either. It can work its way into the tools and processes people already use and start making pieces of those processes better. And that's already happening.

Two and a half years ago, Buildernauts tried to predict which technologies were going to change construction. Looking back, I think we did okay.

But we missed the biggest one.

Not because we didn't know AI was coming. We just didn't understand how quickly it would become part of everything else.

And after going back through that old article, I think that's what I find most interesting. The technologies that ultimately change an industry the most may not be the ones that create some shiny new category. Sometimes they just work their way into everything else until eventually we stop thinking about them as technology at all.

Maybe that's where AI is headed.

Or maybe, two and a half years from now, I'll go back and read this article, laugh at myself again and write another one about everything I got wrong.

I'm starting to think that's half the fun.