The AEC industry has a reputation for adopting new technologies at what might diplomatically be described as a “measured pace.” Artificial intelligence (AI) is no exception. While firms have moved quickly to use AI for automating routine tasks, generating marketing content and supporting back-office functions, its broader integration into project delivery and core operations has been slower to take hold.
This is beginning to change. As the technology matures and becomes more embedded in daily workflows, firms are starting to see real gains. What remains less clear is where exactly AI is delivering meaningful return on investment (ROI) and where it is falling short.
That’s the subject of this edition of The Friedman File, featuring insight from two participants in an AI-focused panel at the recent AE Advisors CEO Forum—Leslie Moulton-Post, president and CEO of Environmental Science Associates (ESA), a 750-person firm based in San Francisco; and Rod Sommer, CEO of 350-person, Dayton, Ohio-based engineering firm LJB Inc.
Assessing AI’s Real Value
As with most AEC firms, the AI journey for ESA and LJB has had its share of stops and starts, bullseyes and misfires. Rather than proving to be an immediate and sweeping transformation of how business is done, it has been a pattern of incremental gains.
This is not for lack of AI tools or the capabilities that they offer; there are plenty. What is in shorter supply is clarity on what actually works. While nailing down the specifics of AI’s ROI in your firm may remain difficult, there are some ways to clarify where AI is creating real value and where it is not.
1. Focus on Outcomes
AI’s true value is judged on what it actually accomplishes, not only on whether it reduces the cost or time required to do it.
Like many firms, ESA recognized early that AI had the potential to accelerate work they were already doing, particularly in database management and in seeking more sophisticated ways to analyze and assess information. This position was strengthened by the firm’s acquisition of Sitka Technology Group, which brought additional technical depth and a mindset aligned with integrating technology into workflows.
But Moulton-Post cautions against viewing AI primarily through a financial lens. “When AI first came along, it seemed like everything was a race to cost-cutting. But we quickly realized that’s not the real power of using AI. The real value is in how it helps us deliver more capabilities to our clients,” she says.
This distinction matters. Firms that focus only on efficiency may see short-term gains. Those that use AI to enhance problem-solving and expand services are more likely to realize long-term value and transformative results.
Both firms have examples of how this can work in practice.
At ESA, the firm is coordinating with the Pacific Northwest National Laboratory to leverage their work and understand how AI can be used to support automation of environmental review and permitting-related processes. The challenge is not simply to complete tasks more quickly, but to do so in a way that is legally defensible and maintains ESA’s rigorous quality standards.
At LJB, Sommer described a project involving fall protection safety. The firm used an AI tool to review a PDF of a design and identify potential compliance issues with OSHA and ANSI regulations. In a matter of minutes, the system flagged multiple areas for review and cited the relevant standards.
What would traditionally require a senior-level analysis over the course of days was completed almost instantly, allowing less experienced staff to engage with the process while still benefiting from expert-level guidance.
The implication is significant. AI is amplifying expertise, not replacing it. This allows a firm’s technical staff to evaluate more options, identify issues more quickly and—most importantly—deliver better outcomes.
2. Look for ROI in Small Gains
Building on that point, the most reliable returns are coming from incremental improvements applied across many workflows.
LJB has approached AI the same way it more broadly considers operational efficiency: by identifying small, repeatable improvements.
“If you can take a function that takes eight hours and get it down to one, it’s not going to immediately save you tens of thousands of dollars. But find 25 or 30 of those functions and it can amount to a huge savings,” says Sommer.
3. Be Deliberate in How You Invest in AI Tools
The right tool depends on the task, so it’s critical to understand your needs before making long-term investments.
Unlike some AEC firms that have standardized or developed a firmwide AI tool, ESA seeks the best tool for the specific need. Moulton-Post says, “We want to ensure we are using AI responsibly and we have a policy and specific guidance for staff on how to choose the right tool. We don’t want to use a chainsaw when a pair of scissors will do the job.”
Because the firm has data scientists and software engineers in-house, ESA has also developed a habit of experimenting with AI solutions before committing to an off-the-shelf product. “The obvious question is why not just buy it off the shelf. But this way we can better understand how we work with our own data and make a more informed decision when the time comes to buy,” she says.
That thinking aligns with what Sommer is seeing at LJB, where many of the strongest returns have come from applying existing tools rather than building new ones. In his experience, out-of-the-box solutions have delivered more consistent value than custom development efforts that can quickly become outdated.
4. Build the Foundation Before You Scale
As firms look to expand their use of AI, many are discovering that the limiting factor isn’t the technology but the underlying infrastructure.
At ESA, this has meant investing in systems that connect data across the organization, making it more accessible and usable. The goal is to enable AI tools to draw from a broader and more integrated set of information, rather than isolated project files or disconnected databases.
This also ties directly to knowledge management. In many AEC firms, institutional knowledge still resides largely in individuals, passed down through an apprenticeship model. Or it is scattered among folders and files, with no easy way to organize or retrieve what you need when you need it.
AI has the potential to make all this knowledge more accessible, but only if it is captured and structured effectively. Without this foundation, even the most advanced tools will struggle to deliver meaningful value.
5. Guide Your Firm Through Change
Successful AI adoption depends largely on training, communication and clear expectations. When there is no strategy in place and the process flows without direction, ROI suffers.
“We’ll have a session on AI, and there will be very few or no questions,” says Sommer. “But that may be because most people don’t know enough about it to even know what to ask.”
Without proper context and direction, even powerful tools can sit idle, or worse, be used inefficiently or harmfully. At LJB, the solution has been to provide simple, practical entry points, showing employees a handful of relevant use cases and building from there. Peer sharing, including regular internal discussions of successful applications, has also helped drive adoption.
Another challenge is that some employees, particularly those newer to the field or in less technically demanding roles, feel threatened by AI. The concern is understandable.
This is where it is critical for leadership to convey that AI is not replacing the role of the professional, though it may change how those roles function or develop. Embracing AI as a tool to optimize one’s skills, and thus guard against becoming obsolete, is the best line of attack.
Employees also need to be cautioned to avoid letting AI take full control of the creative process. “We have to use AI as a tool to assist critical thinking, not replace it,” says Sommer. “If people are going to use it as a replacement to critical thinking, how can they become a truly value-added engineer. What is even the definition of engineer?”
Training has to stress that AI, used properly, can accelerate learning and expose staff to more complex problem-solving earlier in their careers. But they must still be expected to question, interpret and apply judgment to the results.
6. Rethink Value and How You Measure It
AI is forcing firms to reconsider how they define value and how they measure return.
For many AEC firms, efficiency gains raise immediate questions about business models. If a task that once took hours can now be completed in minutes, how should it be priced? And how should its value be communicated to clients?
These are not new questions; from the earliest days of CAD and 3D, the industry has struggled to adequately account for the efficiency gains delivered by machines. AI is the latest frontier in this long and difficult battle.
At the same time, measuring ROI remains a challenge. Similar to business development, much of AI’s impact is felt indirectly. Tying an investment to revenue growth can be difficult.
This hasn’t stopped firms from holding themselves accountable. “Everybody admits that AI is a work in process, but that’s not stopping the board from holding our feet to the fire,” says Moulton-Post. “We’ve made notable investments in a few key areas and are reporting on them regularly. We don’t have all the answers yet, but we’re serious about both the level of effort and being accountable to the investments we make.”
What are you doing to measure AI’s ROI in your firm? Any other thoughts on this issue? Write me at rich@friedmanpartners.com or call 508-397-9213.
