AI Engineers and Testing: How Suffolk Integrates Technology into Workflows

For years, experts in the construction industry have emphasized that for artificial intelligence,data quality is crucial. Doug Harrison, Vice President of Operations at Suffolk Construction, is one of the proponents of this view.

Harrison has been with Suffolk for nearly a decade, joining the company in 2018 as a project engineer. Today, he is deeply involved in how the company integrates AI into its workflows, a process aimed at consolidating disparate data and presenting it in a way that builders and executives can understand.

"The construction industry is a highly fragmented, disconnected environment, and the AI world is really bringing all of that together," Harrison told Construction Dive.

To that end, Suffolk hasestablished the "Future of Construction" program, which embeds AI engineers within project teams to help solve critical problems and build the new technological tools needed for successful construction.

Here, Harrison discusses with Construction Dive how Suffolk integrates AI into workflows, how it standardizes technology across the business, and what the company considers when weighing purchasing technology solutions versus building them in-house.

Editor's note: This interview has been edited for brevity and clarity.

CONSTRUCTION DIVE: What initiatives has Suffolk launched to embed AI into construction processes?

DOUG HARRISON:It starts with our "Future of Construction" program, an initiative aimed at embedding AI into our project teams.

It pairs AI engineers with project teams in specific sectors, because most of our sectors differ in how they deliver. Many of the fundamentals of construction are the same, but sectors like mission critical, healthcare, or higher education each have their own nuances.

Pairing AI engineers with project teams really has two aspects.

On one hand, it provides project teams with licenses for more typical tools, such as Claude, ChatGPT, and others that are widely used in the industry today. Our question is, how do we achieve day-to-day efficiency and train the people on our projects to actually use these tools? We want them to use these tools to read documents, perform change order reviews, RFI creation, and similar comparisons. That's part of their responsibilities.

On the other hand, what we believe will truly drive transformative change in the industry are the tools and applications we are developing in close coordination with project teams.

How do you approach deploying technology to teams?

The most important thing we've realized is that you first need to deploy consistent technology to capture data.

Before declaring "this is a required tool," it's crucial to establish the backend infrastructure, clarifying: what data will we extract from this tool? Will it become the source of truth for how we manage projects at the regional level?

Because all of this data not only flows to project executives, but also into our dashboards, where COOs, general managers, and even presidents use it to manage their project portfolios.

You need to view consistent data collection and technology deployment in the same way.

That's exactly why we've been able to successfully standardize tools across our project portfolio.

What does standardizing technology tools involve?

Tools must prove their long-term value and must improve our current execution processes.

Currently, we are looking for the right partners and platforms to support procurement and delivery, for example.

Yes, all of these applications are looking for ways to embed AI into their solutions, but some applications' workflows are better suited to how our company operates. Finding that match helps us manage procurement and delivery.

So, functionality matters, right? You need to have field users actually test the technology, ensure you're getting the expected results, and make sure project teams endorse its ease of use. That's where you want early adoption.

How you actually standardize and deploy technology is really a training strategy issue, and how you ensure appropriate, robust training is provided for these individuals and specific platforms.

Contractors today often think: is it better to build or buy technology solutions? What is Suffolk's stance on the build-versus-buy tradeoff?

Before we prototype any solution, this is an ongoing debate.

I think the partnership withSuffolk Technologies' BOOST programallows us to have a comprehensive view of the industry landscape and where it's heading.

What we value most is that if a technology is at the forefront of the platforms we currently use, or if there's something new on the market to test, we won't build that application ourselves unless it's truly a custom need based on how we work. We understand that for standard applications, we can't build and surpass the models already available in the market.

We're not a software company; at the end of the day, we're a construction company. We want to provide first-class value to our clients and trade partners, and ultimately provide the most productive solutions for our team members so they can grow and scale within the organization.