In a matter of months, 400 Central Avenue, the site of St. Petersburg’s future tallest building, has gone from a ground-level foundation to a multi-story cement structure rising rapidly from the ground.
Parker Mundt says artificial intelligence has accelerated the pace of construction for the Residences at 400 Central, a 46-story condominium tower that will become the tallest structure on Florida’s Gulf Coast.
Mundt is the vice president of platform for Suffolk Technologies, the venture capital partner to Suffolk Construction. His company chooses tech start-ups to fund and implements some of that tech into downtown job sites.
Suffolk Construction, meanwhile, is the lead construction contractor for 400 Central, a project funded by developer John Catsimatidis Sr. of Red Apple Group. Suffolk has 19 ongoing projects in Florida.
At 400 Central, two devices help catch mistakes before they happen and avoid costly leaks and water damage, Mundt said. Those tools are called WINT and OpenSpace.
Artificial intelligence in construction has boomed in the last year, Mundt said. Here’s how he screens AI systems that could speed up his crews’ work. This conversation has been edited for length and clarity.
How are AI advancements speeding development up at 400 Central?
They’re all designed to solve different problems. WINT is a water intelligence system. It’s designed to detect and mitigate leaks on job sites. What’s unique about it is as the machine learning learns the use of water, then it has the ability to detect anomalies in your water usage. If from the hours of 1 a.m. to 4 a.m. you rarely use any water, but then WINT notices 220 gallons rushing out of one pipe, it’ll be set so it will ping superintendents and say “we’ve detected a water anomaly.” If you don’t respond in 2 minutes, it automatically shuts off the flow of water. It protects from very costly and schedule-ruining events on a job site.
What about OpenSpace?
OpenSpace uses a computer and an off-the-shelf GoPro camera. As construction crews are walking, the camera is constantly taking pictures. What you end up with is the Google street view of your job site that’s time stamped. It will tell you that you walked the entire fourth floor today and you are 60% complete with metal studs and 40% complete with the drywall work. It uses a little bit of AI and machine learning. It’s inherently boosting construction crews’ productivity and actually giving them more time to focus on bigger issues with the job site.
Follow trends affecting the local economy
Subscribe to our free Business by the Bay newsletter
You’re all signed up!
Want more of our free, weekly newsletters in your inbox? Let’s get started.Explore all your options
How does Suffolk vet new AI technologies?
We have a department dedicated to experimenting with tech solutions on our job sites. A few of those become standardized across the projects. In particular, we focus on adding productivity, increasing safety or increasing our level of sustainability on a job site.
Tell me more about your venture capital activities.
We were one of OpenSpace’s’s first customers in 2018. We were able to invest in the company using Suffolk Technologies, which is a venture capital fund that sits alongside Suffolk.
We invest in six to 10 companies per year. We raised a $110 million fund that will deploy over the next five years or so. We also run a program called Boost. Think of it as Shark Tank but for the construction technology space. Applications are open in the spring for companies in pre-seed or Series A, meaning they’re typically generating $1 million or less in revenue. We got 200-plus applications from over 30 countries worldwide, and we select less than 10 that are admitted to our fourth Boost program that starts this fall.
How many of these companies are advertising AI in their business model?
This year, 45% of the companies that applied to our Boost program are claiming to have some AI in their business model. That increased twofold from last year.
This is different than Chat GPT. Large language models cannot read drawings. AI can do that. What’s really interesting is we’re seeing companies coming in finding innovative ways to read and digest and understand 2D or 3D models.
How do you sell your construction crews on using new tech?
I think we have been very fortunate at Suffolk that we have a very forward-thinking workforce. But we’re not going to introduce 10 projects to something that’s not been piloted. We run in isolated instances where we can make adjustments. We have an opt-in innovation champions network. People have opted in to talk about advances in the construction industry and talk about how we can leverage and use them at Suffolk.
What’s an indicator that a piece of tech might not work out?
There are a lot of metrics we look at. The two that matter most to us are productivity and cost. Another red flag is when our project teams don’t have a clear understanding about the inputs they put into an AI engine compared to outputs, like, are you likely to have a safety incident this week? Are you going to finish on time? If AI can’t explain why it’s coming up with the answers, it’s very hard to go back and tweak some of the inputs.
There’s also a little bit of a battle of how much of the AI can we trust at this point in time? ML and AI continue to prove themselves. Sometimes if we don’t have any data points that are proving if the tech is working, then we’re very cautious about new AI that cannot be explained to our construction team.
What improvements have you noticed since implementing this tech?
It’s increased our productivity. We’ve proved that we can be safer, proved we can mitigate more of the risk, and I think we will prove that we’ll be able to take on more projects, ideally trying to get more projects into the ground faster.