Fish spoke with ENR about Suffolk's "Job Site of the Future" strategy to embed artificial intelligence in its processes, an initiative that formally launched June 9 after years of internal investment and testing. He says the AI-enabled operating model is the first-of-its-kind and intended to transform project delivery In innovation workspaces at jobsites, with professionals using AI tools having access to real-time project data. These "AI engineers" focus on delivering improvements in three core areas: design, schedule and process. This interview has been lightly edited for clarity.

ENR: Suffolk's AI engineers are now embedded directly on jobsites and participating in critical project meetings. How did you did you get project stakeholders on board?

John F. Fish: Let's step back, because I want to unpack a little bit for you. We all know construction costs today are greater than the value of the build. You and everyone at ENR also know about the labor shortage. You also know about five in two, five out, two in, and you know we have no immigration policy and about [the state of] interest rates. I often tell the story about the Empire State Building being built in 13 months, and now it takes five years to build a building like that. So at the end of the day, something's got to give.

We are aware of all of that, and about low margins contractors have to deal with.

All of those factors are really compromising the ability to have the American dream come true for people. what I really mean by that is that the only way people build wealth in America—I'm generalizing—is through home ownership, and right now that home ownership issue is not within reach of people between 27 and 35 years old. So if we don't solve this problem in theory, the founding principle of the American dream is at risk. A lot of people don't look at it that way, but that's the way I look at it. People say we [are] 4 million housing units short. We [are] about 8 million units short in America today, So when I think about that, I think that is the overall picture of our country: the commercial, residential, retail market right now is basically at a standstill. 

So as I'm thinking about this, we just did [a study of AI in construction] with the Massachusetts Institute of Technology Media Lab. As a result of that [we asked ourselves] can we build a building at 20% less cost and with 25% less schedule?

And that brought you to the Job Site of the Future strategy?

It created the three-step process to the Job Site of the Future. It's creating a seamless platform in the business, bringing back the master builder at its core, like we did in the 1930s and 1940s, but introducing technology and artificial intelligence and machine learning into that—so you see we're embedding that in the master builder process. The idea came out to be creating this seamless platform that's outlined in the paper, and the seamless platform is really at a high level. It's procure, design and build. So it's turning the traditional process on its head a little bit.

Not engineer-procure-construct, but procure-design-build?

Exactly. So if you take a look at your regular shop drawings on a typical residential project, [the structure] has doors, frames, kitchen cabinets, flooring, and you know exactly what you need. You don't need shop drawings later on, in theory, following this process four or five years down the road. The designer then incorporates those elements into a design. You know your cost, you know your delivery, which is important, and then you know how to build. So the theory is having Suffolk Design—which has 125 architects, structural engineers and mechanical engineers now, and probably 250 a year from now—do that design. We have Freedom Source, which is a procurement group, and we have Liberty, our construction group with about 1,500 carpenters and laborers right now across the country. So the idea of the job site of the future is really to solve the issue of the seamless platform.

So, how do you integrate the procurement, design and build into that seamless platform that can achieve 20% less cost and 25% schedule savings? 

The idea was to start with three particular environments at Suffolk that are investment environments that I've made as the owner of the company. We do three things:  AI as relates to design, AI as it relates to schedule and, AI as it relates to process. I'm going to give you one example of process. We're doing a data center job, about a $2-billion job, [with] $200 million worth of billings on a monthly basis. It takes three to three-and a-half weeks to do the requisition. Can we leverage AI to get down to a week? We're down to two weeks right now. The idea is ultimately get down to get down to one week at the end of the day.

What about design and build?

Let's go to schedule. When we resequence a job or something doesn't arrive at the right time, the critical path gets disrupted because material delivery date or something happens on a job. How long does it take us to resequence a critical path on a large-scale $2-billion job? Sometimes it takes a week to 10 days to do that. Can we do that in two-and-a-half days with AI? The answer is yes. Okay, let's go back to design. When the client gives us all the material cut sheets, because they supply a lot of equipment on data centers, how long does it take us to incorporate, not coordinate, but incorporate material equipment that the client has already purchased into the overall design at the end of the day? 

So the Jobsite of the Future is to take three particular use cases: design, schedule, and process, and figure out problems with people on that jobsite. We are leveraging AI in a way that's not going to take your job. We prove that out to them. We're looking to take your paper, not your job. On a jobsite, you've got people doing requisitions. They don't want to be doing requisitions for three-and-a-half weeks. They would rather be doing something else, so you're taking menial tasks that otherwise frustrate people, and we put that in the hands of AI.

How does this physically play out on said job site?

On the Job Site of the Future we have a full-time AI engineer there, with a trailer that's 2,000 sq ft that's bolted onto the existing construction trailer with a glass panel, so people can look in from a curiosity point of view. The whole job site is all digitized. You're creating an environment that speaks to what you're trying to accomplish. It's non-threatening, and you have somebody there that is has a construction background, but is AI fluent. and it can make people feel comfortable in asking any question at the end of the day, even what they might consider a stupid question.

What's the relation to 100MAG, your modern construction process group based at Suffok's expanded 100 Magazine Street office and warehouse in Boston?

The idea there is we take the ideation in the field, digitally transfer that to 100MAG in the core of the company, and we will have probably 10 to 12 AI engineers and people working together to take that ideated ideas and transfer them into something that's practical from a solutions point of view on the job site, and then put a training program around it. Job Site of the Future, 100MAG, ideate, develop, train and then we bring it into a collab. 

The fourth environment we have is in Bangalore, India. We just opened up our office [there] six months ago, and the idea is I'm not using that as a labor arbitrage. I'm doing that as an intellectual arbitrage, okay? Because there's extremely bright, intelligent PhDs over there that we can leverage to turbocharge our AI strategy, the Job Site of the Future, our 100 Mag, and our collaboration at the end of the day.

Can you give me an example of how AI has helped on a project using Job Site of the Future strategy? 

when you hire Suffolk, I want to be able to say to a client, that we're going to give you that procurement, and then we're going to give you options of three very specific offsite manufacturing capabilities for sections of the building. Only three, not 17, three—the three that can create the most value and create the most predictability. 

Take a mechanical room. We're doing Dana Farber Cancer Institute's world headquarters right now. It's a $1.2-billion job. The mechanical room on the building that's coming out of the ground right now was about 32,000 sq ft. We're doing offsite manufacturing of the mechanical room and we bring that in in pieces. Mechanical room is 24,000 sq ft. We're saving 9,000 sq ft of roof of the building because we're off-site manufacturing coordinating mechanical room. Can we do that with an electrical room, and can we do that with a third component of the building that makes sense? Maybe bathroom pods or something else like that. Standardization is the key at the end of the day.

I know Suffolk's clean data lake is key to this whole process, but are there any AIs you can mention that are you know being used in this strategy?

Light Table. Right now—because we have our own architectural, mechanical, structural group—we call them TVs. When we get involved in pre-construction. We have two stages of TVs—technical validation reports. Okay, so we will take the design at a mature design documents. We'll leverage Light Table with our own engineers and architects, and we'll give a comprehensive report to the architect of record and the client, not point fingers when somebody calls level of detail. They're 400 and they're nor right. This is where the drawings are from a baseline comparison point of view, and then they come out with the the construction documents. We'll do another technical validation report.

And that's how you're getting to 20/25?

If we can anchor this in on the preconstruction process, leveraging AI, getting things geometrically, dimensionally correct and coordinated, and now also specifying material up front, ultimately, we don't want to be doing shop drawings four years from now, three years from now. The level of predictability [is there], and I can assure you, the 20% cost savings and  25% schedule savings.