Brick & Mortar Ventures’ Darren Bechtel On AI, Robots and the Next Big Thing(s)

Brick & Mortar Ventures founding partner Darren Bechtel views the construction world through the lens of both a builder — he’s a scion of a major construction company — and an early backer of proptech companies like Procore, EquipmentShare and Fero.

As the broader industrial market turns increasingly to technology sectors outside traditional software as a service (SaaS) companies and relentlessly toward artificial intelligence, a catch-up with Bechtel seemed in order.

On Sept. 19, Bechtel spoke with PropTech Insider about the rapid adoption of AI in construction and the necessity of AI-enabled solutions in startups; physical AI and robotics’ potential to address labor shortages and improve efficiency; whether AI will enhance or replace SaaS companies; and what he sees as investment opportunities for Brick & Mortar Ventures.

This interview has been edited for length and clarity.

PropTech Insider: From your perspective as a major contech venture capital investor, what do you wake up thinking about most each day?

Darren Bechtel: That it’s never been easier to build technology based on some of the modern-day coding tools. The need for someone to be a subject matter expert in how to build software is really starting to disappear. Those that understand the pain point or the problem now find it easier to build a solution or provide some sort of prompt to be able to realize some benefit of modern-day technology, without having to go to an incumbent technology solution provider and pay them a significant amount for a solution. People have had a taste in the last couple of years of using something like a ChatGPT to get an answer that they can validate and realize is good enough. 

So there’s been this sort of awakening, and not just from the corporate level or the IT and innovation officer level, but from the boots on the ground, out in the field. Those people are starting to dabble — and oftentimes go rogue — and use tech that may or may not be approved to realize material benefits in their day-to-day workflows. 

In the early days of construction innovation, you saw a top-down kind of mandate. Now, in the last two years, you’ve seen this bottom-up pull, and it’s created a little bit of a scramble. It’s been a bit of an accelerant, if you will, and really exciting. But you know what? I sleep pretty well.

That sounds somewhat chaotic. Like it could cause more problems than it solves.

Yes, indeed, because there’s great potential, and specifically through our lens of looking at the [architecture, engineering and construction] industry, the rate at which the average Jane or Joe has become familiar with and been open to adopting solutions like ChatGPT or you name it — that level of becoming a verb and a noun — and is synonymous with GenAI now. 

But, as IT departments across the construction value chain used to have to work hard to figure out how to go through change management in order to encourage people out in the field to use new tech, all of a sudden there’s been this about-face saying, “Oh, no, now we have to actually limit it or make sure that people aren’t moving too fast.”

We are figuring out what are the workflows we want to have deployed across an organization to make sure they’re really benefiting from the value-add of those, but also to make sure that there’s confidence in the accuracy of the information or the end result — the work product. And there is increasingly great concern around the potential privacy and data security risks, which are very real.

Given all that, are you looking to invest in proptech startups that use generative AI more than you have been, or are you now somehow leery? And how do you advise or deal with your limited partners, or your portfolio, to help them avoid such pitfalls?

Yes, on the investment front, definitely. But almost every startup right now that’s fundraising has added AI to their elevator pitch. I do think being AI-enabled is table stakes now, much like for a while saying that you were a cloud-based solution if you were software, or having over-the-air updates if you were software-enabled hardware. That used to be seen as novel and a value proposition. Then it just became assumed that was happening. 

The game for startups, depending on where in the venture cycles you are, usually is efficient growth and being capital efficient. Investors are not interested in growth at any cost, but in general being capital efficient and having capital efficient growth as critical. We work with our portfolio companies to try to help provide a little bit of a playbook if that is new territory to them. But, nowadays, if there is a founder that’s looking to raise money and is not already pretty well versed in some of those tools, you start to get concerned about their capability as a leader in the modern-day startup landscape.

That said, we’re not jumping on opportunities because it’s AI and we have to place a bet. We are still pretty disciplined when we’re looking at underwriting investment opportunities. Our approach is very much needs-based, or opportunity-based, instead of just pure tech for the sake of tech. So we’ve been a bit more skeptical, and have not done as many pure-play AI investments as perhaps some of our peers or other VCs, because we think most are not really up to snuff to meet the needs of the customers in construction.

Have you placed recent bets on proptech startups based on what you consider to be their real AI for construction?

Yes, I think where you’re going with that is more of that pure-play AI startup, instead of reaching for AI or adding it into their pitch. On the residential construction side, there’s Klutch AI, started by one of the first product managers at Stripe. So a very gifted, product-minded founder, with an incredibly strong founding tech team behind her. They’re trying to build a project management platform for residential construction developers. We did that investment in December 2024. That was an unconventionally early-stage bet for us, but we really like their disciplined approach of using AI to build a purpose-built solution for a kind of known pain point that this founder had experienced herself as she started getting into real estate development.

Another one, out of France, is called Freeda. They are focused on risk management, utilizing AI for checking to make sure that design documents and other plan documents adhere to required codes and guidelines for submissions. So it’s making sure that for any of the paperwork or designs that you’re submitting you’re complying. It’s sort of a debugging code that is the equivalent of spell check for plans.

What have you seen in AI for construction hardware?

A category of AI that gets me very excited — and I’m admittedly biased as a former product designer — is physical AI or robotics. It’s massive. If you talk to a builder for more than five minutes, depending on what they’re complaining about — material pricing, labor shortages, which are not getting better, only worse — it comes up either first or second. It’s long been the case that we’ve had shortages and fewer people coming in, more people aging out than coming into the sector. But now layer on top of that some of the immigration reform and deportation, a significant portion of that labor force was just removed or went into hiding. 

So far, there’s not great solutions or proposals around how we’re going to field all of the work that continues to stack up. And there are some of the most ambitious projects in our history all starting to stack up and looking to start in the next couple of years. We have a real challenge figuring out how we’re going to be able to provide everyone from laborers to electrical engineers to mechanical engineers. We need to sound the alarm a little bit, because in this race toward general intelligence that’s now deemed a national security priority, there’s a lot of electrical work, a lot of mechanical work, and you have to figure out how you’re going to power these data centers. 

We don’t have the people to do the workload that’s coming. The technology is not providing a replacement — not for manual work. The thing is, you can only do so much.

What can technology do to improve the construction sector?

There’s a lot of potential with prefab and modular, and we really should be embracing that more and be doing more work in a controlled environment that’s further away from where the actual site is. Nonetheless, at some point in time, you need to build parts of components or sub-assemblies of a structure or infrastructure to the final destination where that’s going to be permanently attached. 

That’s going to require people, equipment and tools. Construction has long operated with humans using tools, and increasingly better tools and equipment to do that work. But physical AI and the ability for robotics to start augmenting and automating certain tasks is needed because there’s just not people available to us, or it’s a task or workflow that you should not have people doing, either because of health and safety risks or the fact that robotics can do a better job. If there’s something that requires a really steady hand, robotics can have a steadier hand. If there’s something where fatigue is involved, the robot doesn’t get fatigued.

There’s still a lot of things that I think for a very long time humans are just better equipped to do. A perfect example I heard is that if you need to reach around a person and stick your hand into their pockets and pull out a set of keys, while that’s socially uncomfortable for humans, it’s very easy for everyone to do. A robot cannot do that very easily.

You don’t need to spend the time and money and brainpower of leading roboticists to solve for that pain point. But one of the companies in our own portfolio that we’re really bullish about, and I think is kind of right place right time, is Canvas, which is focused on robotic drywall finishing. Also, some of the solutions where you’re translating digital content into the physical world, like the layout process, is very exciting. 

And, frankly, it’s surprising that it’s taken so long to realize the equivalent of the last mile, where the digital world leaves off and hands the plans in some shape to humans. Our portfolio company Rugged Robotics has done a really good job with that.

Getting back to software, is AI replacing software as a service companies in the AEC sector?

There’s a good argument that those in the best position to succeed with AI are the incumbents that already hold all the data and can just layer on additional functionality.

We’ve asked some customers and our corporate investors, “Where do you come down on this spectrum of fear versus excitement? And how do you comb through the growing sea of startups all claiming to be the AI solution to all the problems?” If they are a Microsoft house, or Google Workspace house, the software that they’re already paying for is physically AI-enabled. So even if you did nothing, you will benefit eventually. I think there’s the incumbents and those that have been delivering solutions to their customers and have developed a reputation of trust as a solutions provider who already have the customer base and are afforded a bit more time to try to develop viable solutions. 

But it’s well known that radical innovation usually comes from outside existing platforms.

Looking ahead, of all this AI and tech in general, what is going to  propel your investment business?

It’s all of it. As long as humans continue to need to occupy physical space, construction and the construction value chain will be around. As technology evolves and there is the next great thing that’s going to help humanity, either from a clean energy standpoint, or transportation point, or any number of infrastructure solutions, that continues to create more work for an already strained construction value chain. 

While Brick & Mortar has a sector focus, as investors we have critical diversification through investing across all verticals within the built environment. So there’s a great opportunity.

Philip Russo can be reached at prusso@commercialobserver.com.