AI becomes a construction tool

During a recent webinar by the Maryland Center for Construction Education and Innovation (MCCEI), one panelist described the importance of construction companies adopting Artificial Intelligence by quoting American futurist Stuart Brand:

“Once a new technology rolls over you, if you’re not part of the steamroller, you’re going to be part of the road.”

AI is rolling through the business world and many construction industry leaders say companies that fail to productively harness its power will find themselves struggling with a competitive disadvantage.

“If you don’t adopt an AI model to improve your back-end operational efficiency, if you don’t adopt an AI model to improve your front-end sales efficiency, know that your competitors will,” said Kevin O’Brien, CIO at Advance Business Systems. “Then, rather than having a leg up on the competition, you will be fighting to maintain your business.”

That market prospect doesn’t only apply to large construction companies. Increasingly, AI bots, personalized GPTs, data analytics, forecasting functions, and large language models are becoming business tools for mid-sized and smaller firms. And while AI adoption is a major technological advance, it is also becoming easier to implement in small to medium-sized companies.

Yael Meretyk Hanan, Co-Founder and CEO of Pelles.ai and a speaker at the MCCEI webinar, advises trades contractors to approach AI adoption with a business focus, not a technological one.

“Let’s stop and understand where your challenges are,” she said. “What are the things you are looking to improve?… Where are the inefficiencies?”

Meretyk Hanan urges companies to pick one pain point in their operations and begin their AI adoption by addressing that. Even small contractors can afford to develop custom GPTs, which can cost as little as $20 a month, to automate inefficient aspects of their pre-con or back-end operations, she said.

Many companies can leverage software systems they already license to optimize business processes through AI, O’Brien said.

“Everybody seems to have an AI copilot right now. Microsoft has one, Procore has one, Sage has one,” he said. “You don’t have to invest in expensive, new applications. You can, for example, leverage the copilot in your Microsoft 365 Suite to source data and begin optimizing your processes.”

Software companies, IT consultants and other organizations offer training on how to effectively use various copilots, ChatGPT and other AI bots.

Successful adoption hinges on several core activities, including making data accessible often through a data warehouse or data lake, securing data against hackers or inappropriate use, and directing the AI bot to focus on productive questions and use cases.

“The biggest thing to think about when you are using a GPT is your prompts. You want to be sophisticated in your prompts,” said Ricky Venters, a contractor with Hartman Executive Advisors and founder of Success Journey Strategy. “Instead of saying, ‘Hey ChatGPT, give me the answer to this problem,’ ask GPT to help you think through the right questions you need to ask in order to solve that problem.”

That approach of using an AI bot as a brainstorming partner delivers multiple benefits.

It facilitates deeper problem analysis, brings more relevant data into the process, helps prevent AI hallucinations or results based on incorrect or insufficient information, and “it does not infringe on your authority, your thought leadership and your understanding of your business,” Venters said.

“Over the next decade, firms that want to win in the construction industry will need to adopt AI to solve the problems our industry has experienced for decades – too many delays, too little coordination and too much wasted time,” he said.

DPR Construction is already using off-the-shelf products, including ChatGPT and Microsoft CoPilot, as well as some custom AI tools developed inhouse to boost its operational efficiencies.

The company’s efforts have focused on AI use cases that can deliver the greatest benefits, said Hrishi Maha, a Data and AI practice leader with DPR. To date, those have included using large language models to rapidly access information about a workflow or project, and accelerating access to information embedded in construction specifications, RFIs, and other project documents.

DPR’s AI adoption has also tackled a perennial construction challenge, namely staffing levels.

“Working with our preconstruction team and our business developers, we developed an algorithm which helps people predict what staffing will be required for a project,” Maha said. “Based on work we have completed previously, it will estimate the FTEs required and the individual traits needed. It might say you will require two project managers, two schedulers, a safety engineer and two project investigators to execute a job, and it will give a range of hours for those people.”

That prediction, Maha cautioned, isn’t foolproof. “There are so many nuances in construction which AI is not going to know about. But it gives teams a good starting point and, if the AI estimate is different from their own estimate, it will prompt people to assess the situation more deeply.”

DPR, which has been gradually integrating AI-enabled processes throughout its project and business workflows, expects the technology to fundamentally improve performance within the construction industry.

“AI is the next technology frontier… It can change how we do construction,” Maha said. “The result will be time savings for our people, accelerated jobs, better schedules, more predictable outcomes.”

Faster, safer, smarter
How could AI improve your operational efficiencies today?

Smart schedules:

Incite Automation has built AI-supported project management systems that assess tasking for construction projects. The systems know the roles and certifications of each company’s employees. Based on that knowledge and project specifications, it will propose assignments to a particular job.

“It is able to make determinations based on the location of the jobs and the employees, their certifications and past performance,” said Dustin Sitton, Principal at Incite Automation and a speaker at the MCCEI AI webinar in May.
The system, he added, can also generate proposals for the equipment that should be assigned in conjunction with those individuals and propose adding “fluff time” to a schedule, based on the project’s circumstances.

The tasking plans aren’t always 100 percent correct, but they still save construction companies a great deal of time, Sitton said. “I know organizations that use a big, Excel spreadsheet and spend a day, a week or more on resource management. Now they are getting suggestions [from the AI system], making minor tweaks and they are done in an hour or two.”

Safety:

The combination of AI technology and security cameras, drone videos and other surveillance can enhance jobsite safety while also reducing the time needed for safety compliance.

One Incite Automation client employed a “hard hat guy” to walk jobsites and ensure everyone was wearing their hats. Incite linked the jobsite’s network of security cameras to the contractor’s Power Automate program in the Microsoft Ecosystem and trained an AI bot to recognize a person without a hardhat. The system then scanned the site every 30 seconds and alerted the company to any hardhat issues.

Code compliance:

Higher models of ChatGPT are able to answer questions about building code compliance, including what is required in local, international and other codes, and the changes between different iterations of each code.

Preconstruction and back office operations:

In many companies, AI-supported systems are streamlining payroll processing, invoicing, supply chain monitoring, updating of project documents, employee onboarding, and more.

AI is even helping preserve the deep expertise of veteran employees who are ready to retire.

One option is to “record interviews with those industry pros about how they do things, their thinking processes, problem solving, constructability issues,” said Kassy Slaughter, Senior Manager, Industry Transformation at Procore Technologies and a speaker at the MCCEI webinar. “AI can then convert that into a podcast or other training materials.”