Your contractor just told you they use AI. You nodded. It sounded good. But when you asked what that actually means for your project (for your schedule, your budget, your risk exposure) they changed the subject.

That’s the dirty secret of AI adoption in construction right now. Firms are buying tools. They’re not building capability.

And as the owner, you’re the one paying for the gap.

The tool isn’t the problem

Software vendors will tell you their platform will transform your projects. Maybe it will. But transformation doesn’t come from a license. It comes from people who know how to use the tool, know when not to, and know how to protect your data in the process.

Here’s a question worth asking every firm you work with: What does your AI training program look like?

If the answer is a shrug, or “we let people figure it out,” you have a real problem on your hands.

What “figuring it out” actually costs you

When professionals use AI without training, the mistakes aren’t obvious. They look like normal work. A specification section that’s subtly wrong. A cost estimate that used the wrong context. A project update that sounds confident and is completely misleading.

This isn’t hypothetical. Teams without guardrails are running client data through public AI tools with no understanding of confidentiality implications. They’re producing outputs they don’t fully understand and presenting them as professional judgment.

The liability lands on you. Your project. Your building.

The training budget question

Most construction and design firms spend thousands sending project managers to conferences, bootcamps, and leadership programs. Ask them what they’ve spent on AI training in the last twelve months.

For most firms, the answer is zero.

That’s not because AI is new. It’s because training feels optional when the tools are free or cheap and the pressure to appear innovative is high. Slap a ChatGPT subscription on the company card, declare yourself an AI-forward firm, and move on.

Appearing innovative and actually being capable are different things.

What you should demand from your team

If your contractor or designer is using AI on your project, that’s a reasonable place to start a real conversation. Not a technology audit. Just a few direct questions:

What AI tools are you using, and for what tasks? You deserve to know whether AI is being used for scheduling analysis, cost estimating, specification writing, or something else entirely. Not because you need to approve every tool, but because it affects your risk profile.

What’s your data policy? Is your project information going into a public model? Is it being retained? Shared? Most professionals don’t know the answers. A trained team does.

Who’s reviewing the AI outputs? The right answer isn’t “anyone who has time” and it isn’t “everyone.” It’s a specific person with specific accountability. If no one owns the output, no one owns the risk until something goes wrong, and then it lands on you.

The transparency test

There’s a version of AI-enabled construction that works in your favor. Real-time job site monitoring. Automated progress tracking. AI-assisted schedule analysis that flags risks before they become overruns. These tools exist. They work. They give you the visibility you’ve always wanted and rarely gotten.

But they only work when the people running them know what they’re doing. When they understand the tool’s limits, know how to interpret the outputs, and have standing procedures for when the AI gets it wrong.

An untrained team with powerful tools creates expensive surprises. A trained team with the right tools gives you something closer to honest, real-time control over your project.

What to do right now

Before your next project kicks off, or on your current one, ask each major firm you’re working with to show you their AI use policy and their training approach. Not a slide deck. The actual policy. Who’s trained. On what. How recently.

If they can’t show you, that’s useful information. It tells you where the risk is hiding.

The firms that will earn your trust in the next few years aren’t the ones with the most sophisticated software. They’re the ones who know how to use it responsibly, transparently, and with your interests in mind.

The tool is cheap. The judgment to use it well is not. That judgment comes from training. And right now, most of your vendors aren’t doing it.

You should know that before you sign the next contract.