AI is accelerating fast, but in AEC the real opportunity lies in using it to expand capacity, not replace people.
According to a 2025 global survey conducted by Bluebeam of more than 1,000 AEC professionals, only 27% of AEC firms stated they were currently using AI for automation. However, 94% expressed that they planned to expand their use of AI in 2026.
The prevailing skepticism in the AEC industry may have been warranted initially. As Matt Shumer, CEO of HyperWriteAI, notes, “In 2022, AI couldn't do basic arithmetic reliably. It would confidently tell you that 7 × 8 = 54. By 2024, it could write working software and explain graduate-level science.”
The rapid advancement of these platforms shows no signs of slowing down, either. In fact, METR, an organization that tracks the time it takes AI models to complete real-world tasks without human help, showed in late 2025 that Claude Opus 4.5 was able to complete tasks within 10 minutes that would normally take a human expert nearly five hours to complete.
If that scares you, it shouldn’t. The new age of AI in the AEC industry is about capacity, not reduction – freeing people from repetitive, low‑value tasks so they can focus on the work that really matters. At Dunaway, we have done exactly that. Here is how Dunaway has managed to give employees nearly 10,000 hours of their time back, and how your firm can do it, too.
Fitting AI into the AEC world
Dunaway found that AI implementation works best when it is intentional and owned by a dedicated group – that means assigning ownership to a dedicated innovation team, whose job is to explore, test, and apply emerging tools to workflows.
Another critical factor is domain knowledge. Dunaway’s innovation team is led by Vice President of Technology and Principal, Brian Bowden, PE, who formerly served in Dunaway’s structural engineering department. As Brian explains, “because I came up through an engineering role, I’ve 'lived' the processes we are now automating.”
Innovation teams that include people with engineering or technical backgrounds have a significant advantage because they understand the workflows they are automating, where bottlenecks exist, which steps are critical versus redundant, and where automation can safely intervene. If AI initiatives are driven solely by IT or external consultants without that operational context, implementation tends to slow down, and adoption suffers.
Automate to innovate
One of the most important lessons Dunaway has learned is that automation works best when a process is already clearly defined – if a team cannot articulate how work moves from start to finish, no amount of AI will fix that.
For this reason, Dunaway prioritizes use cases where standard operating procedures already exist or can be documented cleanly. A written process becomes the foundation for automation. Once that clarity exists, the innovation team can implement AI tools to automate processes quickly and effectively. This discipline prevents wasted effort and ensures that automation enhances, rather than disrupts, existing operations.
To decide where to focus, Dunaway's innovation team categorizes potential AI initiatives into three primary areas: sales, internal support processes, and operations or revenue generation. From there, ideas are evaluated using a criteria matrix that emphasizes speed, scope, and impact. Initiatives that involve fewer stakeholders but impact many people, shorter timelines, and clear business value rise to the top.
This approach reflects an important mindset shift for AEC firms. AI does not need to solve the biggest problem first. In fact, starting small often accelerates the adoption of AI. When teams see a process improve quickly – and reliably – skepticism fades. Automation becomes tangible, not theoretical.
A regulatory challenge at scale
As Dunaway approached its 70th year as a firm, leadership recognized that regulatory complexity had become a growing operational burden. Engineers were spending hours each weekday manually searching, interpreting, and cross-referencing regulatory documents from more than 50 cities across Texas. This work was necessary, but it was not the best use of licensed professionals’ time.
Rather than outsourcing the problem or purchasing another license-based software tool, Dunaway chose to build an internal AI agent tailored to this specific challenge. Using Microsoft Copilot Studio, paired with Microsoft Power Automate, the firm created a system that continuously prepares, updates, and structures regulatory information.
The result is not just faster access to information, but greater confidence in its accuracy. Engineers can spend less time searching and more time applying judgment. The automation reduces manual effort, minimizing the risk of outdated or missed requirements, and significantly accelerates productivity across teams.
AI agents versus SaaS tools
One of the most strategic decisions Dunaway made was choosing to build internal AI agents rather than relying solely on third-party SaaS products. Traditional SaaS pricing models are typically seat-based, meaning costs scale directly with headcount, and for growing firms, this can quickly become unsustainable.
By contrast, building an internal AI agent is largely a one-time investment. Once the solution is developed, it can be deployed broadly without incremental per-user fees. Just as important, the solution can be designed around the firm’s existing workflows rather than forcing teams to adapt to someone else’s product.
Today’s AI platforms have lowered the barrier to entry. AEC firms no longer need to wait for vendors to build niche tools for AEC-specific problems. With the right expertise, they can create custom agents that reflect how their business actually operates. This shift represents a fundamental change in how technology decisions are made – moving from buying capability to building it.
Accountability, accuracy, and trust
A common concern about AI in the AEC industry is accuracy. Dunaway addresses this head-on by tightly controlling what its AI agents can access and how they respond. Agents are restricted to approved knowledge sources and instructed explicitly not to infer, assume, or summarize beyond what is documented. They also make citations and source links front and center so users can verify the information themselves.
This approach reinforces an important distinction: AI accelerates access to information, but it does not replace professional judgment. Humans remain responsible for decisions. Training employees to use AI responsibly and to understand its limitations is just as important as the technology itself.
Capacity, not reduction
Perhaps the most important takeaway from Dunaway’s experience is that AI is not about replacing people – it is about giving them time back. By automating work, such as data entry, document retrieval, or tasks with repetitive administrative steps, the firm can scale revenue without increasing headcount.
This reframing is essential for the AEC industry. The goal is not fewer engineers, planners, or surveyors. The goal is to let those professionals focus on the expertise they were hired for. When AI handles the surrounding noise, the human contribution becomes more valuable, not less.
As AI capabilities continue to advance, firms that invest early, intentionally, and thoughtfully will be better positioned to adapt to the technology. Dunaway’s experience shows that meaningful results do not come from chasing trends, but from aligning technology with real business needs. For AEC firms willing to do that work, “Let’s automate that” is not a threat; it is an opportunity.
 |
Sara Grayum is a senior digital content specialist and associate at Dunaway. Connect with her on LinkedIn. |