Brett Poulos doesn’t just look for office experts for his preconstruction teams — he wants jobsite expertise as well.
Poulos, national director of preconstruction and estimating for Kansas City, Missouri-based Burns & McDonnell, wrote in the company’s Q1 2026 update that owners want faster deliveries and better visibility, which has led to preconstruction taking on another level of importance in the construction life cycle.
To that end, Poulos has maintained that adding tradespeople to a builder’s preconstruction team creates a level of understanding unobtainable without people who have built a job themselves.
Hands-on experience also helps as builders continue to leverage artificial intelligence; Poulos advocates for tradesworkers in the room to push back and vet the tech and the answers it gives.
Here, Poulos talks with Construction Dive about how preconstruction took on this new level of importance, how Burns & McDonnell uses AI in preconstruction and the value tradesworkers add to the process.
Editor’s note: This interview has been edited for brevity and clarity.
CONSTRUCTION DIVE: You wrote in your post that preconstruction is the “new center of gravity.” Can you elaborate on that?
BRETT POULOS: Clients are consistently asking construction and design-build firms to deliver faster. They want earlier cost certainty. They want to better manage the volatility of the markets when it comes to tariffs and pricing volatility, commodity volatility — obviously, we’re seeing that right now — and fuel pricing.
So, the earlier a team can provide that input in the design process, the more empowered those clients are.
From the center of gravity perspective, your schedule, your costs, how you’re going to build it, what means and methods you’re potentially going to use, all of that is determined in the preconstruction phase of the project. The more confidence and more data you have in that phase of the project, the stronger delivery you have for that client or that owner.
How does Burns & McDonnell use AI when it comes to preconstruction?
We have a kind of a progressive roadmap. We use large language models, like everyone else does currently. But we’ve also built agentic capabilities, or agents that have discipline-specific tasks.
Brett Poulos
Permission granted by Burns & McDonnell
We still have human oversight to be able to vet it and review it, but with the documents and data we have as a design builder and EPC firm, having all of that in house and having access to it directly under one umbrella is really powerful.
We’ve deployed enterprise-wide AI tech, and everyone who’s within the company has gone through that training. We have a roadmap to be able to have future capabilities at work as well that are not necessarily deployed today, but AI is absolutely at the forefront. All of our decisions, all of our technology spend, are all about how things will fit into the AI roadmap over the next five years.
How does Burns & McDonnell’s use of AI in preconstruction set it apart from other builders?
I think it’s our ability to inject our design data and preconstruction and estimating metadata — so our historical costs from past projects, understanding our productivity rates on past projects.
Being able to add that data into the design model is what we’re working on for a future capability, and how do we bring up our ability to develop cost models in real time, as opposed to waiting for an engineering deliverable output at specific milestones?
That’s the value that we see in AI as part of our roadmap. It’s how we input our cost history, our schedules, our procurement lead times, and pair it with our design data and design knowledge much earlier in the process.
You’ve said preconstruction and estimating teams need more tradespeople. How so?
Within our preconstruction team of approximately 100 people, we have 17 people without a college degree that came directly from the trades. I don’t want to say it’s necessarily an anomaly, but it’s not common practice.
What we found is having teams injected into the early project lifecycle who have been there — who have actually built it, who know the crew sizes it would take to install this type of work, the exact type of equipment, how they would go about doing it — having that direct field experience within your team is invaluable.
Having someone who understands the cost, or maybe understands technology extremely well, is important, but really getting down to the accuracy and insights that someone can provide who’s actually physically built that in the past is huge. The value is extreme.
And as we and other firms continue to deploy artificial intelligence, having troops on the ground who have the ability to challenge that with confidence and know that “that is not how I would do that,” only gives you even more capability. That’s how you correctly challenge AI outputs, by having someone who’s physically done it in the field in the past.
