Whether you're a chief operating officer at a 25-person firm, like Justin Hoff at Emergent Architecture in Cedar Falls, Iowa, or a chief information officer at a 1,300-person firm, like Brooke Grammier at CannonDesign in Houston, artificial intelligence sits heavily top of mind.

"I spend a significant amount of my time trying to help people understand how data and AI aren't really here to replace jobs," said Grammier, who spoke in two sessions in PSMJ's AEC INNOVATE conference in 2025. "But if they don't learn how to use it, they'll be replaced. It's more about that than the actual technology. AI is a new tool, and if people don't learn how to use it and infuse it in their work, there will be other people who know how to use it who can fill those roles. People should think about how AI can help them do their work better."

"We have spent a lot of time educating our people, our marketing staff, our finance staff, and designers on their own data," Grammier continued. "And how they have control over the outcomes of AI and machine learning based on how they manage their data and do and don't collect it. It's about breaking it down to something that's not scary and helping them see how they can use it to better do what they do."

A couple of months ago—after 17 years at a previous firm—Hoff was brought on by a growing Emergent Architecture to "help them catch up, technologically and organizationally, with processes," he said.

In the pre-AI days, Hoff said, the silos for a COO could be draining, doing HR, payroll, accounting, business development, and so on. "But with the advent of ChatGPT, Gemini, and other models, I lean heavily on using those processes and tools to wear those hats more efficiently. So before, if I needed to write a standard operating procedure or make policy changes, I could stumble my way through it for 2-3 hours.

"Now I can spend time with a chatbot and be done in half an hour. If I have a proposal to respond to, I can do the research using those chatbots of the organization and then feed information like, 'Help me write a really good section about our process.' Then, point to a couple of things in the RFP, and we would tailor the message to go out."

In that way, Hoff could act like a four-person marketing department. "I'm bringing over a lot of that line of thinking," he said. "I'm also helping Emergent's marketing to build some of these processes and policies that they do not have in place."

Grammier also referred to helping her firm's marketing department. "You must ask the right questions," she said. "'Where are the pain points?' 'What do you spend most of your time on?' We learn that people are rebuilding marketing proposals from scratch every single time. Why can't we build a database [where] our marketing team members can ask, 'create a proposal for this, this and this," and then AI just builds a template off past proposals that empower them with a baseline."

"I don't have to be an expert in what [my departments] do. I just need to get people to [see the potential and] move down their own individual path. It's helping them understand the future and seeing what's in it for them. That's what's so great about AI. It sells itself."

Grammier added that there's been a "surge" of people asking about data, mostly how to clean it and how to get more out of it. "So it's become more of a push than a pull, which is a nice thing, after 20-some years, to have people willingly want to build this future rooted in data."

Huff has had success on two special fronts:

ChatGPT’s paid version. "With this version, you get to train multiple different bots to do multiple different things,” he said. “I have bots that I’ve trained to be specifically my HR, marketing or process bots, to name a few. Because we're a multi-state employer, I'm asking it to consider all the states that we work in, and then it already knows my prompts from before, saving me time and reducing the cost of our final legal review.”

ChatGPT’s deep research mode. “I'm starting to use that to do analyses of our competition,” Hoff said. “So that when we're responding to RFPs and RFQs, we know what we're up against. Now we can start tailoring the project’s proposal. Marketing materials and proposals can be public record, so we can learn what other competitors have said about themselves. That gives us a better way to say why we're better or more suited for a project.” Grammier said that CannonDesign has built an internal AI tool called Billie that its staff can ask questions to, and it only feeds them information based off company data. So, they can quickly get information that used to take hours to sift through documents or the Internet. “Programs like Chat GPT and Midjourney (a Gen AI program that generates images) “has opened their eyes and brains to what's possible, and they want more,” she said.

This is content from the PSMJ Newsletter, exclusive to PSMJ PRO Members. PSMJ PRO is the fastest-growing network of AEC firm leaders. Not a PRO Member? Try a 50-day trial (no credit card required). You can request a trial here: https://bit.ly/50dayLI