As Adrian Quinto sat in the crowd for an AI talk at the 2025 MEP Conference, he wondered just how much the novel technology was being used among construction contractors industry-wide.
He got his answer right away.
“Show of hands, how many of you guys have used ChatGPT or something like it in the last month?” the session moderator asked the crowd of 150-plus.
Quinto raised his hand, as did nearly everyone else.
“How many of you have used it in the last week? In the last 24 hours?”
Quinto’s hand stayed up—and as the project engineer at ICOM Mechanical Inc. in San Jose, California, panned the room, he noticed most hands were still raised.
“I was taken aback,” Quinto said. “I didn't think many people were actually using (AI) in the construction industry. Because I have this opinion that construction, in general, is slow to adopt technology.”
How to use AI to streamline operations and grow your business was a big topic at MEP this year, when the conference brought hundreds of commercial and construction contractors from across the country to Los Angeles for three days in late January.
But within that topic came an all-important question:
How do you grow comfortable with this fast-evolving technology?
Developing trust
Like many, Quinto is skeptical about AI.
“The way I see it is, AI is kind of a black box,” Quinto said. “We don't know what the inner workings are. I don't know how it comes up with its answers. I don't know if my data is being used in a particular way.”
That’s why, after the session moderator explained that his electrical shop uses AI to rewrite technician summaries, Quinto raised his hand and asked a question.
“Just from a general manager quality control standpoint, how long did it take you to actually trust that the (AI) summarizes accurately compared to what your techs were doing?” Quinto said.
The moderator explained that he doesn’t implicitly trust it—he still reviews every AI summary to make sure things are worded correctly. And that’s OK. Because it’s still a more efficient process.
“It saves time, really, and it makes us look more professional,” the moderator said.
‘Boring AI’
Lindsay Williams realizes that calling AI like ChatGPT “boring” is a bit ironic.
But when looking at the grander scope of what AI can do for the trades—such as using an all-in-one software with advanced AI capabilities to power your shop—ChatGPT feels more like a sandbox to play and experiment in.
That’s why, in her MEP roundtable session about predictive maintenance, IoT-enabled monitoring systems, and AI-driven analytics, she suggested easing into these technologies by trialing what she calls “boring AI.”
“I think it provides a comfort level for people and an understanding of what tools can do,” said Williams, the service manager at Murphy Company. “Even though it's something super basic, culturally, it changes your team, it changes their mindset, so that now when they want to go solve a problem, they're automatically thinking, ‘I wonder if ChatGPT could do this.’”
For example, a technician at her company used ChatGPT to troubleshoot how to solve a problem in the field. ChatGPT replied with six steps—and a journeyman technician at Murphy validated that those steps were the exact ones he would’ve taken, too.
“Our technicians are using (boring AI) to cover knowledge gaps,” Williams said. “And in an industry like ours, where a lot of our workforce is retiring and there's a lot of young people coming in, there's a big knowledge gap in the middle where college was emphasized and the trades were not super popular.
“We're seeing (that AI) can support people with a little bit less experience and develop their thought processes.”
Perfect data
Experimenting with AI is important—but it’s only useful if you’re inputting good data.
That’s what Jonathan Marsh, the CEO of SteelToe Consulting, stressed during his roundtable discussion on the topic.
“Up until a few years ago, having data be perfect was not a big deal. But right now, it's everything,” Marsh said. “If your data is not formatted in a way that it can be ingested by other programs, if it's not something that can be compared to others, you are at a disadvantage.”
Simply put, bad data leads to the “garbage in, garbage out” result, Marsh explained. “Most technology is looking at you and saying, ‘We're never going to solve your data problems, but we'll give you the mechanisms—if you’re a good contractor—to use your data to make predictions.’”
Quinto attended this session, too. And after a few days at MEP, he said he knows things are moving fast in this arena—and they aren’t slowing down
“It just seems people are coming to a realization that (AI is) not going anywhere. It's not necessarily like a fad,” he said. “It'll just become a matter of time for AI to touch every aspect of what we do—aside from the physical labor.”
