Bechtel The artificial intelligence hype cycle might be firmly entrenched in industry, but the architecture, engineering, and construction sector has been somewhat slow to adopt the technology.
A global survey of 1,000 AEC professionals from Bluebeam, a developer of solutions and services for AEC professionals, found only 27% of respondents use AI in their operations.
Further reading:
- How artificial intelligence can benefit construction projects
- Digital twins show great promise in civil engineering. But what’s next?
- Texas DOT works to leverage artificial intelligence’s power
Such cautious behavior is to be expected, said Jeff Sample, senior industry development manager at Bluebeam. After all, for a technology that runs on data, it does not help that 52% of survey respondents still use paper during the design phase and 49% during planning. Forty-three percent report relying on physical signatures and approvals.
Companies looking to use AI need to lead with the problem first.
“AI is not going to do everything for everybody, but the 27% in our report who are using AI knew what their core problems are and how AI could solve them,” Sample said.
AI-based assistance could boost construction sector
Bechtel, a global engineering, procurement, construction, and project management company, knew which problems it could solve with AI. (Because survey respondents were anonymous, it’s unclear whether Bechtel was one of the Bluebeam survey participants. The anonymity aspect also means it’s not possible to break the responses down by job categories. The survey, though, focused on roles actively driving digital decisions and workflows. Respondents spanned design, construction, and operations, including project managers, architects, and engineers, among others.)
One of the challenges for Bechtel was the volumes of documentation. When each jobsite has dozens – if not hundreds – of pieces of equipment, each with thousands of pages of documentation, staying on top of operational and maintenance protocols can be intimidating, said David Wilson, manager of functions at Bechtel.
“We have got a historic library of information, and that has to be better tailored for the end user who needs it in real time and with information in the right context,” Wilson said.
To address the problem, Bechtel corralled proprietary information into a tailored large language model. Instead of poring through operation and maintenance manuals, users can query the model, much like how we use ChatGPT. The result is a handy punch list that extracts relevant information and prioritizes tasks.
“It’s a great example of where AI is being used as an assistant, and it’s changing days of activity into minutes,” Wilson said.
Somewhat similarly, global construction company Skanska developed Safety Sidekick, an AI-powered assistant that delivers safety guidance to its teams.
Bechtel The assistant consolidates Skanska’s internal environmental, health, and safety manual; the Occupational Safety and Health Administration’s construction standards; and supplemental safety documentation into one resource that workers can query through mobile devices or desktops.
The AI assistant is modeled on OpenAI’s GPT-4o model. In addition to providing real-time access to best practices for safety, it delivers practical scenario-based policy applications for daily operations.
The Safety Sidekick is part of a suite of so-called expert sidekicks that Skanska has developed. The My Skanska sidekick searches documents on the company’s internal intranet site. And the Operational Risk sidekick feeds on thousands of accumulated case studies and expert knowledge to provide reasoned risk mitigation strategies for current jobs.
It’s a way to democratize access to on-site job experience, which will likely prove a valuable tool as experienced AEC professionals retire. Having the wisdom of learned experience behind their backs can be a tremendous benefit for young professionals.
Such assistance-based implementations are early use cases for AI in construction.
Saving time, avoiding surprises
AI can also help bridge the gap between design and construction that can stall major projects. Given that design and construction often occur in parallel in engineering, procurement, and construction projects, AI comes in useful to keep projects on track.
When changes occur as a natural byproduct of concurrent work, projects have to be reevaluated in terms of potential new tools needed or time or cost overruns. Bechtel developed a “change” AI agent that evaluates changes in scope, how far along the project is in the execution process, materials purchased, and more, helping project managers quickly understand the implications of the changes. Having this ability makes execution more agile and without surprises for clients.
One of the more advanced use cases for AI involves modular construction, which applies principles of assembly line manufacturing to the sector. It involves breaking down a project into small modules that can be manufactured off-site and later assembled on-site.
Much rides on optimizing the size and components of single modules so they can be reproducible in factories while still capturing the requirements of megaprojects.
The work packages are optimized based on a dizzying number of parameters. Just a few of the many include the time spent on projects, the net area a module can span, and who needs to work on which parts of the module.
“With so many rules in place, in the past it would take a substantial amount of time to take a design model and break it into installation work packages,” Wilson said.
Now, led by a human, Bechtel’s AI agent optimizes breakdown of advanced work packages faster than before – in minutes or hours instead of days or weeks. Advanced work packaging teams tweak these results further. Especially important, as jobsite requirements change, the optimized product can also be updated quickly to reflect new parameters.
Somewhat related to project management, Skanska is using Hakimo, a company that uses AI for physical security, to monitor jobsites when needed. AI-based computer vision can detect intruders and intelligently alert managers when real problems surface.
The AEC industry will also benefit from AI as a tool for workforce development and retention. Fifty-six percent of the Bluebeam survey respondents believe that AI will compensate for the ongoing shortage in construction skills, perhaps by making existing workers more productive. A little less than half (44%) cite advanced digital tools as key to attracting and retaining top talent, alongside culture and compensation.
The AI approach that companies like Bechtel and Skanska are adopting reflects the belief in the technology’s ability to increase productivity and safety of workers. In the U.S. alone, Bechtel has 18,000 craft professionals, Wilson points out, with growth expected to hit 30,000 over the next few years. To achieve that number, “it’s really important for us to be the employer of choice for the craft professional, and the way to do that is to send them home safely and enable them to be as productive and engaged as possible,” Wilson says.
To keep workers safe, Bechtel is using a third-party AI solution from Detect Technologies that uses visual cues from jobsites to detect nonuse of essential personal protective equipment and get ahead of safety incidents by triggering alerts.
Implications specific to civil engineering
Beyond construction, AI plays a role in civil engineering disciplines, such as structural engineering, and damage assessment, including that from hurricanes and other natural disasters. Machine learning algorithms using computer vision can estimate and identify structural problems to ensure ongoing integrity of public infrastructure.
Structural health of bridges and related civil engineering structures are prime candidates for AI. Digital twins, which are virtual replicas of the built environment, have many uses in civil engineering, especially for ongoing building management of large structures like airports and hospitals. These data-intensive operations are also significantly improved by AI.
While AI’s promise is heady, adopting the technology is not easy. The Bluebeam survey respondents report data-sharing security (42%) and cost and complexity (33%) as the top challenges. And 69% say uncertainty around potential AI regulations has affected plans to implement the technology.
The skills gap is also real. Nearly a fourth of respondents (23%) mention it’s hard to keep up with rapidly changing technologies.
A view of the larger picture is critical to AI’s success, said Usman Shuja, CEO of Bluebeam in the report. “The biggest barriers to AEC technology adoption in 2026 aren’t cost – they’re complexity, culture, and connection. Success requires not just tools but training and an integrated approach that connects the dots across teams, project phases, and workflows.”
Sample agreed. “You have to prepare yourself for cultural change and your ability to adopt new technology as it comes,” he said. “The rest will usually follow.”
Advice for nonadopters
The tsunami of digital data in EPC has made information difficult to consume, and AI has the potential to tame the beast. “The data volumes made us think of better ways to consume and incorporate it in our systems, and AI has been a natural step in that direction,” Wilson said.
Bechtel encourages employees with problems to try AI tools, and successful experiments can be scaled through the organization. Bluebeam also advises companies to start small to achieve early successes with AI. Data interoperability – datasets working between siloed departments in a company – is key. “AI can’t tell you anything without data that it cannot find,” Sample said, “so start with data.”
The Bluebeam report found that 94% of the 27% who have adopted AI plan to increase usage next year. There’s a good reason for that.
“AI translates into means and mechanisms to improve the ability of our craft professionals to be successful because it enables quicker resolution of constraints, better predictability for work, and an ability to optimize their workflow,” Wilson said.
Get ready for ASCE2027
Maybe you have big ideas about AI in the AEC world. Maybe you are looking for the right venue to share those big ideas. Maybe you want to get your big ideas in front of leading big thinkers from across the infrastructure space.
Maybe you should share your big ideas at ASCE2027: The Infrastructure and Engineering Experience – a first-of-its-kind event bringing together big thinkers from all across the infrastructure space, March 1-5, 2027, in Philadelphia.
The call for content is open now through March 4. Don’t wait. Get started today!
Bechtel
