Recently, Ron Fritz, CEO of Tech Soft 3D, hosted a roundtable discussion with six other technology executives to gather their thoughts on what’s going on in the AEC space, what issues they’re surprised still exist in the year 2023, and what might be on tap for construction in the next 5-10 years.
Topics discussed include the challenges inherent to a project-based industry like construction and friction points created by a lack of industrialization. Additionally, the execs wonder if we should be paying more attention to the business model of construction itself, rather than any technical limitations, if we want to push the industry to new levels of innovation—and what role AI might have to play.
The participants for this discussion were:
Anand Mecheri, Cofounder & CEO at Invicara, a provider of digital twins for the built environment.
Yves Frinault, Cofounder & CEO at Fieldwire, a provider of field management solutions for construction teams.
Jesse Devitte, Cofounder & General Partner at Building Ventures, a venture capital firm focused on the built environment.
Thiago da Costa, Cofounder & CEO of Toric, a data analytics workspace for construction.
Viraj Voditel, Founder & CEO of Techture, a provider of cloud-based construction management and project management solutions.
Tyler Barnes, President at Tech Soft 3D, a 3D engineering SDK company.
Ron Fritz: Let’s take a look at the AEC space. Where have things changed over the years, and where do things still need to change? What, if anything, do you think is holding the industry back?
Yves Frinault: I think the business model of construction needs to evolve. There is very little incentive from a business model perspective for all the different people on a construction project to collaborate. Sure, everybody could show up 5 months ahead of the project and start working together and start solving problems, but that doesn’t happen. So, I would say the problem that is still holding construction back a little bit is the business model and the incentive between companies. That aspect has changed very, very little over the years.
Thiago da Costa: I think our industry has also been slow to educate the customer about the importance of a SaaS [software as a service] subscription model. When people first learned that they couldn’t own software anymore and that they’d have to pay a monthly fee, there were a lot of complaints. But as a vendor, how do you continually improve the software and make it better if someone pays once and then holds on to that license for the next 10 years? I think that the old business model prevented innovation. We have to find the right balance between doing business profitably and developing the tools that the industry needs because there’s a dire need for technology in the AEC space to continually evolve and improve.
Tyler Barnes: I think that the business model conversation is really interesting. Another thing that surprises me about the AEC industry—probably because I come from a mechanical background—is how much inefficiency there is, particularly in the design-to-build phase. Buildings are conceptually designed manually, then they’re redesigned in some modeler. Data is entered manually; it’s printed manually. It’s very different than the manufacturing space, where you use one application to do conceptual design, and then bring that design into a modeler and build off that design so that there’s no rework. When you think about efficiency frameworks like Six Sigma or lean manufacturing, it feels like that hasn’t quite reached AEC yet.
Viraj Voditel: I agree with what Tyler was just mentioning. We’ve seen that over and over again in this industry, especially when it comes to connected workflows—there are clear disconnects in this process, even in pretty standard use cases. So, it’s surprising that things still aren’t really happening as they should be in the industry.
Jesse Devitte: AEC is a project-based business, not a product-based business—and that’s historically been what has held it back from scaling. It fortunately is becoming more of a product business as construction becomes more industrialized. The entire way the industry works—the duplication and the reentry, from architects and engineers to the builders to the owner operators—is the reason why we mysteriously fail on things that should be so straightforward and logical, like handovers. But there is good news on all these fronts. It’s all evolving.
Ron Fritz: I was expecting a lot of conversation about technical limitations—which are certainly still there—but it seems like market dynamics and business models are very much on peoples’ minds. Anyone care to expand on any of the above—either to agree or disagree?
Anand Mecheri: I was particularly intrigued by Jesse’s comment about handover. It’s hard for anybody to really understand why it’s failing because every party is actually interested in solving that problem. The client wants a good handover. The contractor wants a good handover because it reduces risk. Even though we talk about the whole lifecycle of an asset, there are so many silos at any given stage of that lifecycle and so many different stakeholders—from the project manager, the contractors and subcontractors, the client, the operator—that it quickly becomes mind-bogglingly complex. Every stage along this lifecycle has its own set of problems that we need to find a way to solve because right now the industry is struggling.
Jesse Devitte: For years and years, it seemed like the only entities in the AEC space that cared about having visibility from the beginning of a project all the way through were Disney and the casinos. But now you’re starting to get companies like, say, Meta, who are having a data center built that they plan to use for a long time, and they’re more hands-on. The more you have these enlightened owners demanding—even requiring—visibility and a smooth, connected process, the better. And hopefully, that spills over into owners in other sectors.
Thiago da Costa: Just to add to that, I think that while owners are pushing for things to be better, with more predictability and more transparency, I would have expected by now that owners would be demanding it. I think that’s one reason construction companies have been slow to adopt technology: because no one has been requiring them to. I think if owners were to say that they have to provide these certain things or operate in a certain way, you’d see better adoption, which—by the way—is good for the construction people. Because as you adopt technology, you reduce risk, you increase productivity—you start reaping the benefits of all the things that technology can do. I think that shift from asking for things to be done in a better way to demanding it would just change the game.
Yves Frinault: I agree with all of the above. I think we’ve seen very different owner types, and that results in fundamentally different outcomes technology-wise. For example, companies like healthcare firms really care about getting a specific type of hospital. Companies like Meta know what they want in a data center. Those are the ones that push collaboration to the max and the technology tools and everything that comes with it. But at the end of the day, we’re a late-adopting industry when it comes to technology. I mean, we were like 20 years behind the U.S. Army on adoption of mobile technology.
Viraj Voditel: Adding to what Yves just mentioned about the construction industry being one of the later adopters of most of the tech out there, especially if you look at the 3D side of things: If we look at, say, the gaming industry, they’ve been able to access and interact and view 3D models in a fairly seamless manner for a long time now. The construction industry is still a long way away from being able to interact with 3D models in the way the gaming industry does. That’s because there are still a lot of clunky construction tools that require minutes and minutes of loading time for models. We still need 64 gigs of RAM to be able to open a standard-size model, not even an obscenely large model. There are a lot of struggles around some of these large dataset workflows that have been solved in other industries that regularly deal with large datasets, so it’s surprising that we’re still dealing with this in 2023. It definitely seems like a problem that should have been solved by now, but somehow it isn't.
Ron Fritz: So, where do you think we’ll be 5 years from now? What sorts of things do you think we’ll see in the construction space? And will AI play a role in any of that?
Yves Frinault: In the near horizon, I think we’ll have figured out the loop of data capture and progress capture on job sites so that there is more robotic automation. I think that’s the next 5 years for construction.
Viraj Voditel: I think we will see some fundamental challenges in the industry being solved by tools like AI. For example, this industry has some of the more difficult challenges around interoperability between different file formats. Being able to convert a particular file format or being able to extract information in automated and intelligent ways would be very beneficial. I think some of those practical applications of AI that solve everyday problems like that could quite possibly happen in the next couple of years.
Anand Mecheri: We’ve reached a point where customers realize that their data is valuable and they can do useful things with it, but they don’t know how to scale up in an efficient manner to take advantage of it. My personal prediction is that AI will play a role in helping these companies to build solutions on a platform where all they have to do is define their requirements, and the platform will provide automation assistance that helps create the applications that will start solving their problems. That’s something that we should be able to actually bet on happening.
Thiago da Costa: I agree. I think the most likely thing to happen is that we will see an attempt at reducing the amount of time that it takes to do things so that companies can become more profitable. If you can paint a wall with a robot, why wouldn’t you paint a wall with a robot? If you can easily do valuable things with your data, why wouldn’t you do that? And so on. I think that attempts at reducing the amount of time it takes to do things will come from multiple angles, but I think what will really change the game for all of these companies is how they work together in collaboration with the owners and the transparency they have with the owners. That will change how they can adopt technology, which goes back to the business model thoughts we had earlier.
Jesse Devitte: I think the industrialization of construction will pick up the pace. This is an industry that evolves—it doesn’t change overnight necessarily, but the pace is picking up. There’s no question that the productization of the process is happening, and I think AI will only help that. I think augmenting the design and engineering process is a really natural way for it to start out.
My hope is that we do reach an inflection point in this industrialization of construction so that more broadly, a fuller range of owners can be involved earlier in the process—and leveraging technology—so that at the end of the day, the outcomes are better. A better built world is what it’s really all about, and I think we’re on the path in that direction. There’s a lot of opportunity ahead.