When Construction Companies Build Their Own AI Teams

A student who recently completed a civil engineering program asked me why construction appears to lag behind banking and tech in AI. I encouraged him to take a closer look.

Last month, Suffolk, one of the largest construction companies in the U.S., announced it would place AI engineers directly on job sites. Rather than working in a corporate lab, these engineers join project teams on-site, using real-time data from ongoing builds to improve schedules and reduce expensive rework.

The CEO described this as an "inflexion point." Over the last 10 years, the company has invested more than $100 million to create what they call a "clean data trough." They have collected 293 terabytes of structured construction data, with another 50 million pages added daily from active sites. Now, they are ready to put this data to use at construction sites.

The industry has moved. Quietly.

A new benchmarking report from Unanet shows that Suffolk is part of a larger trend. Today, about three-quarters of architecture, engineering, and construction firms use AI in some form — a roughly 20% increase from last year. AI is now used in project management, proposals, forecasting, and increasingly in field operations.

One detail in the data stood out to me: only 13% of these firms expect AI to reduce jobs. The other 87% have different priorities. Around 71% want their teams to work more efficiently, and 64% are looking for better decisions, not just quicker ones. That focus matters.

The problem underneath the adoption numbers

There is an important detail here. Even with this rapid adoption, only 29% of AEC firms say they fully trust the data going into their AI tools.

Companies are investing in AI, but many quietly admit they don't fully trust the data these systems use.

The main challenge is not the technology itself, but how well data is managed. In this industry, that is actually an opportunity. The firms that move ahead may not have the most advanced AI tools. Instead, they will have people who understand how project data moves, where it can fail, and how to keep it accurate from the beginning. This skill combines project management and data analytics — but most engineering programs do not teach them together.

A role worth preparing for

The Suffolk announcement marks a real shift. Construction companies are now creating roles for people who work at the intersection of AI and fieldwork. Not every company will assign dedicated AI engineers on-site like Suffolk does, but every project team will need someone who can question AI results, spot bad data, and make decisions when the model misses something about site conditions.

This skill is just as valuable for a project manager as it is for a data scientist.

What are you seeing on your projects? Is AI starting to play a role in your work, and does your team feel ready to work with it?