Artificial Intelligence

Artificial intelligence isn’t just for back office tasks anymore. It increasingly can help make jobsites safer, experts say.

Construction Safety Week, which runs the first week of May every year, provides contractors with an opportunity to pause work and double down on safety. In 2026, those conversations featured more references to AI.

Indeed, builders including Skanska, Turner Construction and Balfour Beatty are already using AI to keep workers safe. 

Skanska’s safety planning helper

Building onto the generative AI boom, Swedish builder Skanska has leveraged its own homegrown solutions, known as its Sidekicks, on its jobsites. 

The contractor revealed its initial Sidekick in February 2024. On release, it was capable of generating answers to help workers with project questions based on a well of Skanska’s proprietary data.

In 2025, the contractor introduced its Safety Sidekick, a generative AI tool that plumbs Skanska’s EHS Manual, OSHA construction standards and supplemental safety documentation.

Currently, Skanska mostly leans on the Safety Sidekick for planning, but it still has reactive use cases, said Brian Karas, national environmental health and safety director for Skanska USA Building. 

“It doesn't leap out your phone and tell you to get off the ladder, but it definitely gives us a leg up in terms of learning from this wealth of information,” Karas said.

To that point, the Safety Sidekick uses data from the company’s jobsites to improve planning and learn from past events. This is on top of a vetting process, where members of the builder’s EHS team review the information and make sure it’s accurate, valuable and timely, Karas said. 

That review process also helps protect against so-called “hallucinations,” where AI tools generate answers that are not based on any real fact or data. 

Mitigating those false outputs meant training the tool to respond “I don’t know” when it does not have the answer, said Will Senner, Skanska USA’s data solutions team leader.

“You know, I won't sit here and definitively claim that we have it 100% perfect, but I think we did a really good job at minimizing potential erroneous answers,” Senner said. “Making sure that the tool will respond to say, ‘I don't have information about that,’ if it's a question that is not in its knowledge base, as opposed to trying to make up an answer that sounds plausible.”

Turner Construction’s SafeT Coach

New York City-based Turner Construction, on the heels of its partnership with OpenAI, has made a tool available that the rest of the construction industry can now also use on jobsites across the country, the contractor shared in May 2026.

Known as SafeT Coach, the generative AI tool is trained on Turner’s environmental, health and safety database and gives builders plain-language answers to questions. The current version was built in OpenAI’s ChatGPT environment, while the next will be in Google’s Gemini, which will give builders the ability to select whichever tool they prefer. There will also be two versions, one for internal Turner use and one for external use.

One superintendent, according to the news release, used SafeT Coach to ask whether a vertical shaft qualified as a permit-required confined space. SafeT Coach, in response, generated a decision flow chart, a start-of-day permit checklist and policy citations, which helped the superintendent have a safety conversation with a trade partner’s safety manager. 

Since the initial pilot, SafeT Coach has logged more than 25,000 interactions with Turner staff, trade partners and field teams.

Balfour Beatty’s smart alarm

London-based infrastructure builder Balfour Beatty knows how dangerous roadside jobsites can be — for 2025’s Safety Week, the firm recognized traffic as construction’s fifth fatal risk.

On those jobsites, heavy equipment operators have a lot to worry about, said Jason Sikora, construction manager for Balfour Beatty's infrastructure work in the Southeast.

“So, the operator has to worry about the GPS telling him what to do. He's got to worry about the traffic next to him. He's got to worry about the people on the ground, the other equipment moving around, like a roller,” Sikora said.

That’s where Balfour Beatty makes use of equipment manufacturer Caterpillar’s Cat Detect system. Amid all that movement, the system uses intelligent cameras to track people who go near heavy machinery. If they get too close, the system sounds the alarm. 

The sensor, Sikora said, is based on speed and movement, with scaling alarms that heighten as a person gets closer to the machine. It’s like a regular backup camera, Sikora explained, that can differentiate between a person and an inanimate object. When a person breaks into preset ranges, an alarm, visual indicator or both will activate.

For the people on the ground, there may also be an alarm affixed to the equipment to get their attention, a big help on a roadside jobsite where the noise of passing traffic, heavy equipment and other distractions may drown out one source of noise.

With that in mind, however, Sikora was still adamant that the person was ultimately the focal point for making safety decisions. He echoed a sentiment he’d heard that a human brain was the best computer in the world.

“So we want to give the person in that seat all the information we can, and they're going to make the right choice,” Sikora said.

The most interesting thing happening in artificial intelligence right now isn't in Silicon Valley. It's in a mechanical room in Dallas where a second-year tech just diagnosed a chiller issue in minutes using context that used to live exclusively in a 20-year veteran's head.

While the tech world ties itself in knots debating whether AI will eliminate jobs, the skilled trades are showing the rest of us what AI is actually good at today. Not as a replacement, but as a capability multiplier. That distinction makes all the difference in the world.

The AI blind spot

I'm deep into conversations around AI right now. And I'm seeing two camps clearly emerge.

One is all about anxiety, the fear that AI is going to hollow out entire functions, and maybe whole companies along with them.

The other is all about doing the same work, faster and cheaper.

Both miss the bigger story. There's a third reality playing out right now, and it's moving fast enough that by the time most leaders recognize it, the gap will already be hard to close. It's not that AI shrinks what exists. It expands what's possible. The real question isn't how to do the same work with fewer people. It's what becomes possible when every person on your team has capabilities that didn't exist a year ago.

If you're not asking that question yet, you're building your strategy around yesterday's version of what a team can do.

How trades use AI today

I co-founded BuildOps, a platform for commercial contractors — the people who physically go to a building and make it work. No amount of AI is going to install a chiller or pull wire through conduit. The work is irreducibly human.

And that's exactly why these workers have figured out the best use of AI in construction faster than the rest of the industry.

Because the core of their work has always been physical, the conversation was never "Will AI replace us?" Instead, long before AI made its widespread debut, they asked, "How do we do this?" Answering that question is something AI is good at today, something that gives them a clear advantage in figuring out the best way to do the physical work.

That clarity, forced on them by the nature of the job, is the same clarity every industry — not just construction — needs right now. Once you stop worrying about what AI takes away and start paying attention to what it adds, you make better decisions, in any industry.

Construction’s can-do narrative

The replacement narrative falls apart the moment you look at what's actually happening on the ground.

When we surveyed hundreds of commercial contractors last fall, 78% said AI can improve how they work; 80% said it will be essential to stay competitive within three years; and 81% said they feel confident in their ability to adopt it. These aren't people bracing for a layoff wave. They're gearing up to do more.

This pattern is showing up everywhere: small, focused teams attempting and pulling off things that were out of reach 18 months ago. Not because they got bigger. Because each person on the team got more capable.

AI is changing what a small, focused team can credibly attempt. That’s happening, to a degree, in every industry. But it’s really apparent in construction with teams on the jobsite. 

Beyond cost cutting

The easy way to use AI is as a cost-cutting lever. Ask it for summaries. Use it to shave a few minutes off a workflow. Celebrate the incremental gain.

The harder and more important move is to treat it as a capability shift.

If you walk into this moment asking "How many people can I cut?" you'll get a smaller company. You might hit your margin target this quarter and then spend the next few years wondering why you're losing market share to competitors you've never heard of.

If you walk in asking "What can my people do now that they couldn't do before?" that's a different trajectory entirely. That's how a 40-person team starts to operate with the force of 100.

But that only happens if you're willing to redesign the work itself. Job descriptions, team structures, even who you hire first, they all change when you assume every person on your team has access to a constantly improving set of capabilities.

Stop measuring AI by what it removes. Start measuring it by what it unlocks.

The trades got to this conclusion before most workers because they had no choice. When someone has to be on a jobsite tacking fasteners in place, you can't entertain the replacement fantasy for long. You're forced into the more interesting question: how can this person do this job more effectively when I put these additional AI tools in their hands?

You should be asking the same thing right now. Because if you’re only counting what AI might take away, you’ll miss what matters most: what your people are suddenly capable of doing

McCarthy Building Cos. is looking to get a leg up on its artificial intelligence adoption, and it’s signed a deal with a prominent software development firm to get the job done.

St. Louis-based McCarthy has signed a multiyear, multimillion-dollar agreement with Palantir to power up its internal AI capabilities, according to a June 2026 announcement from the developer.

The contractor will use Palantir’s Artificial Intelligence Platform to develop a connected AI operating system, per Palantir, which will help McCarthy employees on construction steps that range from design to building in the field.

The partnership is centered around Pulse, McCarthy’s AI-native system that offers field teams real-time insight, scenario planning, risk analysis and decision orchestration, according to the announcement. 

Miami-based Palantir provides AI-based software solutions for a wide range of customers in sectors such as military, retail, food manufacturing and investment management firms. The firm has gained recognition as a defense contractor via its “Gotham” software platform. 

Pulse is modeled in Palantir’s Ontology segment of its AI business, which the developer describes as a “tool factory” that lets people develop applications, set parameters and ready them for use, according to its website. On McCarthy’s side, Pulse helps superintendents, project managers and field operators make decisions and evaluate options on the jobsite.

McCarthy will use AIP to connect aspects of the business, such as contracts, estimating and execution, through that same Ontology.

The partnership also marks a turning point for the builder — McCarthy will model its internal tech team around embedded Palantir engineers and build enterprise-grade software natively with its own internal applications team.

“Beyond the ontology and technology itself, Palantir brings exceptional engineering talent and strategic thought leadership that has accelerated our ability to transform complex operational concepts into scalable solutions faster than we thought possible,” said Justin McFarland, McCarthy’s chief digital officer, in the announcement.

That build versus buy calculus is not a new consideration in the construction industry. As construction firms become more tech-literate, they’ve found they often understand their own needs more acutely than third-party vendors.

To that end, builders have begun to partner with later-stage startups, particularly those which can tailor solutions to meet their own specific needs.

Indeed, Turner Construction, the largest contractor in the U.S. by revenue, has taken a position further down that line of inquiry. The builder has questioned the economic incentive to pay a startup that provides a solution, rather than building its own. 

Turner took a step in the latter direction last fall, when it announced its companywide partnership with OpenAI.

In jobsite trailers across the country, Suffolk Construction builders may have a new voice at the table: an artificial intelligence engineer, per a new directive from the company.

The initiative, dubbed “Jobsite of the Future,” will embed AI engineers at Suffolk construction sites around the country to help solve pain points in the building process, the contractor announced in June 2026.

As part of the initiative, AI engineers will sit in on project meetings and iterate on solutions in three core areas — design, schedule and process — per the announcement. John Fish, Suffolk’s CEO, cited cost increases and growing labor woes as part of the drive to create this program.

Jobsite of the Future builds on Suffolk’s ongoing investment in data and general technology, according to the firm. Over a decade ago, the company invested more than $100 million in data, technology infrastructure and innovation capabilities. 

“We believe Jobsite of the Future and our use of artificial intelligence and data will fundamentally change that trajectory and redefine how America builds for generations to come,” Fish said in the news release.

According to Suffolk, the program has so far yielded AI-assisted design review tools that spot drawing conflicts and coordination gaps before construction starts, as well as AI-powered procurement and delivery tracking systems that help teams see into supply chain risks and lead times, among other improvements.

Suffolk also cited a multibillion-dollar project in the Midwest, where the Jobsite of the Future initiative helped the contractor build and pilot an AI-powered requisition process for monthly payment applications. The system aims to help reduce delays caused by missing documentation, approval bottlenecks and repetitive review comments.

Overall, the company’s investment in data has helped the builder create a groundswell of opportunities in its business. Its clean data lake, or collection of data, contains approximately 293 terabytes of structured construction data; the equivalent of roughly 75 billion pages of PDFs.

“Jobsite of the Future connects data, technology and operational expertise so our project teams can make smarter decisions, reduce risk and deliver projects with greater predictability,” said Jit Kee Chin, chief technology officer at Suffolk, in the announcement.

Indeed, the firm has established a reputation for being innovation- and tech-focused. The company’s industry-famous tech accelerator, BOOST, closed out its sixth cohort of member startups in December 2025. The company has also aimed to put new tech solutions, both from outside firms such as Arrowsight and in-house developers for steel tracking, out into the field. 

The push also coincides with other construction firms bringing these development initiatives in-house. New York City-based Turner Construction, for example, has questioned the value of hiring companies to build solutions for the firm, when the contractor could build the solutions on its own.

Global consulting firm McKinsey has partnered with ALICE Technologies to offer artificial intelligence-driven generative scheduling software to contractors across the infrastructure and construction spectrum, according to an April 2026 announcement. The terms of the partnership were not disclosed.

The two companies have already introduced the optimization technology to more than 35 clients in infrastructure, data centers, energy and manufacturing, per the announcement, achieving schedule acceleration of up to 20%.

“Generative scheduling creates a step change in capital project performance,” said Erikhans Kok, senior partner and leader of McKinsey’s Capital Excellence practice, in the news release. “When embedded within the right operating model and supported by strong project controls, it can help organizations make faster, more informed decisions.”

Kok told Construction Dive that a combination of often-cited issues — including labor scarcity, stagnant productivity and supply chain hurdles — were driving the demand for streamlined services on large capital projects.

“If you combine all those factors together, like my simple thesis is that I think we're going to need just better, smarter ways to do capital projects,” Kok said.

The offering can analyze BIM models and Oracle P6 schedules to generate millions of possible execution paths, while identifying the most efficient options, according to the announcement. By treating labor, equipment, materials, space and sequence as adjustable variables, builders can then stress-test alternatives, understand the ripple effects of a given decision and make choices between cost, speed and risk, the firms said.

Mark Pitcher, a partner in McKinsey’s Capital Excellence practice, told Construction Dive that a McKinsey team will build models in ALICE’s software to perform schedule analyses.

“We use those models to perform deep schedule analytics to identify opportunities to improve the project plan,” Pitcher said. “In parallel, we put in place the operating model and build the capabilities of our client teams that will sustain this novel approach over the long term.”

René Morkos, ALICE Technologies’ CEO and founder, said the tech is quickly shifting the scheduling approach on large builds.

“Generative scheduling is reshaping how complex capital projects are planned and delivered,” said René Morkos, ALICE Technologies’ CEO and founder. “Together with McKinsey, we are helping organizations integrate this capability into broader transformation efforts and drive sustained impact at enterprise scale.” 

Clients have used the product to get results and deliver streamlined products, the firms say. San Antonio-based Zachry Construction, a client of Menlo Park, California-based ALICE Technologies, used the tech to analyze schedules and reduced a highway project’s timeline by 28 days.

A McKinsey data center client, meanwhile, reduced the construction schedule on one project by 40%, per the announcement.

“We really use it across sectors at this point, and also internationally,” Kok said.

Pairing up on the tech front has been trending in construction. Contractors searching for ways to speed up their building processes and cut back on waste have entered partnerships with contech firms in search of solutions. 

Beyond partnerships, companies have been combining as well, particularly in the contech space. Industry stalwart Trimble signed an agreement to acquire Document Crunch in April 2026, while Autodesk finalized its acquisition of Rhumbix in March 2026.