Most contractors buying AI right now are buying software. Suffolk is buying people.
The Boston-based builder’s Jobsite of the Future program embeds an AI engineer directly with project teams, and it does so by sector: mission critical, healthcare, higher ed. The differences between those markets are real enough that a tool built for a data center shell won’t survive contact with a hospital fit-out.
Why Suffolk starts with data, not models
Doug Harrison, Suffolk’s vice president of corporate operations, has been with the firm since 2018, when he started as a project engineer. His argument is the one every serious practitioner eventually arrives at: the model isn’t the hard part.
“This is such a fragmented and disjointed business and environment that we work in here in construction,” Harrison told Construction Dive, “and the world of AI is truly bringing that together.”
The practical consequence is that Suffolk won’t mandate a tool until the back end is ready to receive what it produces. Before anything becomes required, the question is what data comes out of it, and whether that data can serve as a source of truth at the regional level. Standardize the capture, then automate on top of it. Get that order backwards and you end up with 40 project teams generating 40 incompatible datasets, which is where a lot of the industry currently sits.
Two layers: licenses and purpose-built tools
The near-term layer is unglamorous. Project teams get licenses for general-purpose models and use them to ingest documentation, run change order comparisons, and draft RFIs. That’s a productivity play, and it’s available to any contractor with a credit card.
The layer Suffolk thinks actually moves the needle is the custom one: applications built hand-in-hand with the project teams that will use them. That only works if an engineer is sitting in the trailer, which is the whole design of the program.
The part nobody wants to hear
Embedding engineers is expensive, and it doesn’t produce a demo you can put in a pitch deck next quarter. It produces a data foundation. Contractors that skip this step and go straight to buying an AI-branded platform will get a dashboard that reflects whatever their field teams happened to enter that week.
Suffolk is one of the larger private builders in the country and can afford to run this experiment. The interesting question is what the mid-market does. Peer firms tracked by Construction Dive are largely still in the pilot phase, which is a polite way of saying the data problem hasn’t been solved yet.