Blueprints to Building: AI Becomes Smart Tool for the Construction Industry

But labor is just one challenge among many. Another issue is construction data, which is plentiful but often messy, according to Miles Smith, senior industry growth and strategy manager at software company Graphisoft, developer of Archicad, a building information modeling (BIM) platform used by architects and designers in more than 100 countries.

“Imagine the illustration of taking all of the materials required for a house and putting them into one giant dump truck, and then just dumping them on the site,” Smith told The Forecast in an interview. 

“That’s a data challenge that we face right now. There’s so much data, and it’s so hard to wrangle.”

Although the digitization of architecture, engineering, construction and operations (AECO) is underway, it’s not a simple endeavor, said Smith, who pointed out that construction projects include multiple trades, each of which has its own interests and doesn’t always share data with the others. For example, a mechanical engineer focuses on heating, cooling and ventilating a building, while a structural engineer is more concerned with mechanical ducts not impacting major load-bearing beams and columns. Meanwhile, a supplier needs the correct specifications to properly do their job. All of them are dealing with unique requirements within their particular focus area.

“Building practices, standards and codes vary enormously from global region to global region,” Smith said. 

“Localization is a challenge within AI, and building codes are full of contradictions and dependencies and new issues that come out every couple years.”

Using AI in Construction for Design and Collaboration

To address challenges in areas like labor and data sharing, contractors and construction companies are turning to AI.

Procore Technologies, for example, is a technology partner for all stages of construction. Its unified technology platform uses AI and data-driven insights and decision-making to improve efficiency, reduce rework and minimize risk. Its "The Future State of Construction" report discovered that 28% of a project’s time is spent on rework, and 18% is spent searching for data. The company therefore built Procore AI to turn all those fragmented pieces into a single system for intelligence and collaboration.

“Anything that is soul-sucking should be automated,” Kris Lengieza, field chief innovation officer at Procore, told The Forecast. 

“Submittals, RFIs, schedule updates, progress tracking. People spend a day, a week, a month putting these things together. And it should be a matter of minutes.”

These benefits extend to all elements of construction, according to Scott Flores, CEO of Empire PLS, who has worked with concrete, grading and surface issues for more than 25 years. His team’s projects vary from simple pothole repairs to complex seven-figure, multi-phase parking lot restorations.

Flores believes AI in construction opportunities include estimate preparation, scope verification and proactive risk identification. He offered an example: A parking lot rehabilitation project of 85,000 square feet may require up to 170 tons of material to use asphalt at varying depths. An AI program can compare past paving projects; utilize site imagery, measurements and material specs; and identify outliers prior to submitting the estimate.

“In practice, this allows estimators to catch missed labor hours or material loads that could impact the budget by over $10,000,” Flores said.

He added that AI analysis paired with field imagery can help contractors proactively manage potential challenges. In the previous example, for instance, the construction team could discover inconsistencies sooner and adjust schedules if AI confirms onsite that the initial material quantities or site conditions were incorrect.

“AI provides secondary oversight during the planning phase, which reduces the chances of human error during mundane tasks,” Flores continued. 

“Essentially, your schedule and budget have a better chance of staying on track.”

Building a Foundation with Voice and AI Agents

Even when they’re not facing a backlog of projects, construction workers must juggle many different tasks at once. For that reason, making it easier to collect and consume project information is critical, Lengieza suggested.

“Voice and visual are going to be king,” he said. “If I can just walk and talk with a pair of Meta glasses on, it’s seeing where I’m seeing, it’s hearing what I’m saying, and that all is context.”

Lengieza said contractors could use that input to create daily logs. An AI agent could then offer even more support by flagging items humans forget to mention — for instance, recognizing a delivery ticket from a photo album — and adding those to the log, too.

“It’s helping me remember things I didn’t even say out loud,” Lengieza said.

AI agents are also vital to building the data centers that make them and other AI innovations possible, noted Szoeke-Schuller, who pointed out how complex data center construction is thanks to components like individual server racks, HVAC units and routing cooling systems, to say nothing of design and operating costs. AI agents can optimize all these moving pieces, many of which may change throughout a project.

“You want the whole process, from the smallest components to the largest, to be as efficient as possible,” Szoeke-Schuller said. “These agents allow you to keep all the different levels in check. Otherwise, it’s a very tedious, cumbersome and error-prone process.”

Although data has traditionally been the main bottleneck with AI tools and agents, Szoeke-Schuller argues that the main challenge now is seamlessly integrating data with infrastructure and organizational strategy. 

“People know a lot about the specific design, intent, constraints and performance metrics, but they’re not always able to put them into writing,” he said. 

“Once the agent has the right context and tools, it’s really independent in evaluating many design options and running the simulation.”

While AI agents can support contractors in planning and on jobsites, Smith stressed that the technology should complement human work rather than replace it.

“We have to remind those folks using an agent in their practice, ‘Hey, you wouldn’t send out the work of a first-year intern without quality assurance,’” he said. “You wouldn’t put your stamp on that without verifying what’s going on. Let’s take AI-based information the same way.”