Construction AI is no longer a pilot in Florida. Builders are using it to finish paperwork in minutes.
Gulf Shore Business reports that Suffolk Construction, Stevens Construction and DeAngelis Diamond are investing heavily in AI. Suffolk also has AI work under way in South Florida.
What Is Suffolk’s Jobsite of the Future?
Suffolk runs 275 active jobsites. Eight carry its Jobsite of the Future designation, including the $1.2 billion Phase 2 terminal expansion at Southwest Florida International Airport.
That site has a software engineer on staff and a conference room dedicated to AI. It will help shape Suffolk’s AI best practices.
Cary Greenwood leads the effort. He comes from software engineering, not construction, with past roles at NV5 and Raytheon.
Greenwood is tuning data from Requests for Information. These are the questions contractors and engineers exchange about plans and specifications.
The change is large. Before Greenwood joined in February 2026, preparing an RFI could take six to eight hours.
Six months later it takes as little as 10 to 15 minutes. “There was a pretty noticeable difference when Cary got here,” said assistant project manager Dwight Moe.
Greenwood described the work. “It’s a large amount of data, definitely gigabytes of data. We essentially mine it.”
Suffolk’s South Florida Work
Suffolk is also using AI in Miami’s market. NBC 6 reported in August 2025 that the contractor had AI tools and robotics on nine South Florida projects.
The tools included blueprint robots and a robot dog with a 360-degree camera. Other software creates submittals in hours instead of days.
Suffolk said its AI tools save an average of 840 hours a year per project.
Daniela Ariza of Suffolk framed the goal as workforce readiness. “AI is here to be an assistant more than anything else,” she said.
She noted that 41% of the construction workforce is expected to retire by 2031.
Stevens Construction Trains Everyone
In Fort Myers, Stevens Construction uses AI in every department. That includes estimates, plan coordination, requests for information and accounting forecasts.
Founder Mark Stevens said use has accelerated during the past two years. “Pretty much everyone in the office has been trained in AI,” he said.
He credits younger staff for the push. “The young folks in our industry have forced me to embrace it.”
Stevens founded the company in 2003. He admits one habit has not changed.
“I’m still the only one in the office who likes to touch and feel a set of blueprints.”
Stevens added a caution. “We’re not using it irresponsibly or depending too much on the information, because it’s so new.”
DeAngelis Diamond Built a Spinoff
Naples-based DeAngelis Diamond went further. It launched Future Bylt in April 2025 to build AI tools for construction.
Future Bylt has grown from six to 18 employees. The team used Anthropic’s Claude Code to speed up software development.
Its tools include Squads, which recommends team members for roles, and Scout, which searches the web and databases for recruiters.
“Talent in construction is very scarce,” said chief innovation officer Brett Diamond. “We’re trying to figure out how to overcome that.”
The parent has 326 employees, about $700 million in revenue this year and roughly 70 active job sites.
The offshoot also gives the company its own brand, people and culture. “It’s our beta model,” Diamond said.
“Once we have these platforms in place, we’re looking to offer these services to the outside world.”
Diamond plans to offer the tools to other contractors. Robotics is next. “We do think there’s a combination of human experience and expertise along with robotics.”
What It Means for Miami Builders
Miami’s construction market faces the same labor squeeze. Faster RFIs and smarter staffing matter on tight schedules.
The University of Florida won a $750,000 National Science Foundation grant to build AI tools for construction education, the Rinker School reports. That grant targets the same talent gap.
The lesson from Southwest Florida is simple. Start with paperwork, measure the hours saved and train everyone.
Sources: