At the recent AGC National Convention in San Diego, two leaders in construction technology – Autodesk Senior Vice President and General Manager for Construction Solutions Jim Lynch and DPR Leadership Team member Atul Khanzode – hosted a fireside chat to discuss approaches to innovation, emerging technology like AI, and risk.

Excerpts of their conversation below have been edited lightly for readability:

Lynch: Atul, you’ve said DPR wants to find new ways to innovate to be more predictable, more sustainable and safer, but at the same time, you have to deliver projects on time, on budget. How does DPR strike that balance?

Khanzode: We won't be a great business if we’re not fundamentally delivering on that promise of predictable construction. We have a lot of bright people that want to try new things on their projects. So, we created a mechanism for funding outside the projects themselves so anyone can apply.

That created a huge pipeline of ideas and we realized we needed a more disciplined approach to innovation. For example, we’d see an idea out of one office and the same idea was being discussed in three other places. We needed a way for ideas to be shared.

Additionally, many times, what an owner or a customer considers innovative is tied to what they are trying to do with their project. So, we encourage teams to pair their ideas with customers whose interests align and may want to help out.

Lynch: I have a saying to my Autodesk product management team when we're trying to solve a challenge: the answer is not in our building; the answer is out on the jobsite.

In the last five years, the drive to apply technology to the biggest challenges has moved quickly. Things are moving at a much faster pace now in the construction sector than they were back in the early 2000s in the design sector, and things were moving fast then. So, for us to innovate, we have to work closely with contracting partners.

We also know many of our customers are developing technology themselves, but many of them don't want to be in the software business. So, in the past, we've acquired technology from our customers. Then the last thing that I find extremely powerful is investment in the construction technology sector over the last couple of years has just exploded. Literally billions of venture capital dollars going into construction tech startups. It's challenging to keep up with it.

How does DPR go about working with your partners on projects that drive broader innovation?

Khanzode: A good example where we have seen big success was with multi-frame racks for one of our largest data center customers. This was something that we've seen quite a bit in Europe, but in the U.S. it was novel idea. We had to bring trade partners together to make it happen. So, we created an incentive mechanism with those trade partners and it aligned with our customer’s desire to accelerate schedule.

Lynch: You hit on something: pressure from clients to move faster. Projects are more complex and delivery schedules are much shorter. Do you see your clients driving the use of technology?

Khanzode: Yes and no. We see that a lot of institutional clients, with portfolios of buildings that we build for in multiple locations, are very interested in standardizing their delivery approach for all contractors. So, they may bring ideas to their group of contractors to scale.

We also see others that leave it up to us, that expect us to bring technology and new ideas to them. In some of those cases, it's probably their first or second project, and they may not have the support or desire to try new things. In those cases, they're expecting us to help them understand the value.

Lynch: At Autodesk, we define industrialized construction really as the convergence between manufacturing and construction. How do you deliver more projects using manufacturing-type approaches?

Khanzode: Let’s start with the need to do more with fewer people. The superintendents I worked with 20 years ago are no longer in the industry and fewer younger folks are coming in. So, the approach we take is, to solve all those challenges, we have to get dialed in on designs that can be manufactured and that provide the opportunity to do component-based design.

We have invested quite a bit in manufacturing units. At many of our regional offices, we are seeing success with what we call prefabrication assembly facilities where our self-perform craft can manufacture elements and we can maintain a steady workforce.

Lynch: Let's switch gears to another topic that's near and dear to my heart and yours as well: data and artificial intelligence. I know it's a buzzword, but it's an important buzzword and I think it's really starting to have applications.

Khanzode: There's not a single day that goes by that somebody doesn't say something about construction AI. I think it's hard to sort out the hype versus reality, so we have an internal task force with people who are the experts in the technology working with others who really know our business and how we take care of our people.

I'll give you a practical example: Every customer we deal with is interested in what the general conditions of the contract are going to be. The better we can predict certain elements like labor will benefit us as well as the customer.

So we use machine learning to look at, say, the last 20 years' worth of data. When people work on projects, they charge specific project codes, so it's good data.

If we have a healthcare project in California, we can predict what type of staffing we need based on the last 30 hospitals we have built in California. So that's a very pertinent example of something that we can do.

Lynch: Absolutely. When you consider the data that is being amassed on all the projects we're all working on, there's a tremendous amount of data. So how do you turn that data into useful information?

At Autodesk, we approach AI through three lenses: the analyze lens, the augment lens, and the automate lens. Through that analyze lens, you gave a good example of how we help our customers use AI and machine learning to extract key insights and drive better outcomes. Another is we have a customer in Europe that uses our AI technology to look at previous project information to identify potential water penetration risks.

To us it's making sure that the knowledge that is captured in that data can be used to drive better outcomes and predict issues before they take place.

I'm also curious to hear more about risks. One of the biggest risks I hear as I speak with our customers is the labor challenges. But, construction projects themselves are fraught with risk. What are the biggest risks that you and your team work to overcome, and how do you go about that?

Khanzode: Execution of quality work is essentially the biggest risk that we have. We're trying to mitigate that through all the means that we have talked about, utilizing this virtual design infrastructure as much as we can.

On the labor side, a few years ago, we did not have the benefits for the craft we have today. We’ve done more with craft benefits so that our thousands of craft employees feel like they have a much better future with us than other places.

The other thing we see, and get challenged with all the time, is we know what happens when you involve all parties early on in the project and utilize their expertise to deliver benefits. But we still see a lot of resistance to alternative delivery like design-build. There's a whole educational component our industry needs to do with owners. Fortunately, I feel like we've been lucky because DPR has done a fantastic job using technology to help minimize those risks really across the board, which is impressive.