Oh, the Humanity! AI Forum Goes Long On Blending Tech With Talent

In a recent virtual forum about artificial intelligence innovation in commercial real estate, Steven Song, founder and CEO of Diald AI, sought to discuss whether AI had surpassed simply offering automation and efficiency, and instead had evolved to influencing “judgment itself” when it came to transactional and strategic decision-making.

“Where is AI creating value now?” Song asked during the Feb. 11 forum hosted by Commercial Observer Partner Insights, and presented by Diald AI and AI-driven construction platform LightTable. “Where do expectations still outpace reality, and what needs to happen for AI to meaningfully shape investment decision-making in commercial real estate?”

Ronald Kamdem, managing director and head of U.S. real estate investment trusts and commercial real estate research at Morgan Stanley, replied that there are three categories dominating AI implementation by public companies.

“No. 1 is front-facing applications,” said Kamdem. “So, for apartment companies today, if you want to rent an apartment, you don’t have to deal with a person anymore. An AI tool can handle that.”

Kamdem mentioned that it’s the same for public storage rentals, noting that “the largest self-storage owner in the U.S. talks about how 80-plus percent of their interactions today are happening through AI and digital tools.”

He then noted that back-office functions, like drafting or reviewing leases, and certain infrastructure and property operations functions for apartments, like smart locks and HVAC systems, have become greatly automated, often with AI.

Kamdem believes much of this will come to be seen as “table stakes.”

“For first movers that use these tools, the feedback has been pretty impressive,” said Kamdem. “Like lease abstraction tools that are 98 to 99 percent accurate, much more accurate than any sort of human work would be.”

Song then asked about a recent CNBC appearance where Kamdem said that AI will disrupt the investment space. 

“The real estate industry is about supply and demand, and investing in markets that have the best fundamentals,” said Kamdem. “Historically, we’ve had very good data on supply. Where we’ve struggled is on the demand side. There are a lot of data points that are really hard to aggregate, and we’ve never really been able to tap into unstructured data. This is the holy grail of what AI will bring to the space. For the first time, there’s a lot of unstructured data and local markets that can be used in the investment process.”

Kamdem believes that this sort of information will ultimately enable AI tools to suggest investment opportunities and screen deals.

“Deal screening seems like the most low-hanging fruit for AI,” said Kamdem, “because an investment committee can be pretty precise about the 15 to 25 factors they’re looking at and want AI to screen for.”

This discussion was followed by the panel “AI in Action: Conversations With the Leading Firms Using AI Today.” Participants included moderator Nikki Greenberg, CEO for Real Estate of the Future, along with Jeff Schacher, co-founder and chief AI officer for MetaProp Labs; Will Pearce, co-founder and CEO of Orbital; and Lindsay Ellis, vice president of asset management at Elizabeth Property Group.

Discussing whether companies should buy or build their AI capabilities, Schacher noted that while the best option will be different for different companies, AI has progressed to the point where building is generally the better option. 

“The building reality has changed,” said Schacher. “AI’s made it possible to build custom AI at a fraction of what it cost in the past.”

That said, Schacher noted that partnering with firms or individuals more experienced in AI is an option that most should consider.

“Building is not just code. You need expertise, infrastructure and pattern recognitions from doing this across dozens of companies,” said Schacher, who recommends a people-first approach over a tech-first approach when determining AI strategy. “My thesis is generally that you can’t necessarily buy the dream. You somehow have to build the dream, but you don’t have to build it alone, and you certainly shouldn’t build it by yourself.”

The panel then emphasized how companies are seeking ways they can to use AI to save money and resources.

Ellis, whose company is an affordable housing owner and operator, mentioned she recently spoke with a firm that used AI to cut down on security guard costs, a considerable expense for Elizabeth Property Group.  

“I recently discovered that Chicago’s O’Hare Airport is utilizing an AI tool that identifies if someone is smoking in the airport just by their hand movement. Or, if someone leaves a bag alone in the airport, the AI will page someone without having a human to watch it,” said Ellis. “Bringing that into the multifamily space could really reduce our guard costs. We can have an AI monitoring for break-ins.”

Ellis notes that AI adoption will include ongoing searches from all departments for ways to identify new AI tools that can help reduce costs. She also agreed with an earlier Schacher comment about every company needing AI champions — people who are dedicated to learning and staying up to date on the technology and its potential uses, and expanded it to say that every department should have one. 

Ellis mentioned that however deeply AI is brought into a process, concurrent human involvement is essential as a guardrail. 

“One thing we’ve done in the multifamily space is to utilize real-life human phone calls to build our AI knowledge bank,” said Ellis. “That’s a great guardrail, and it brings that human element into AI.”

The event’s final panel was “The Next Major Wave of AI in Real Estate: What Investors Are Backing.” Mary Stewart Malone, principal and co-founder of Addison Partners, served as moderator, speaking with Jake Fingert, managing partner at venture capital firm Camber Creek; Paul Zeckser, co-founder and CEO of LightTable; Sandy Jacolow, senior vice president and chief technology officer at Empire State Realty Trust; and Jon Carden, chief marketing officer of Fundrise subsidiary RealAI.

The panel noted that AI’s immediate triumph is taking tasks that used to take days or weeks and executing them in minutes.

“On the investment team, you start to look at a deal that’s got over 100 leases and legal documents,” said Jacolow. “You can scan and process those in literally a couple of minutes, have it do a brief analysis for you, pull out key terms, and then have your staff actually working on the analytics.”

But while CRE searches for every possible application of AI, Fingert reminded us that a certain amount of humanity will always be required for successful interactions over the long term. 

“There are a couple things humans are a lot better suited to handle, and one is anything that requires empathy,” said Fingert. “There are a lot of tasks where you have to really understand the human-to-human interaction. I agree that AI is great for identifying discrepancies and understanding where things are off-market. But there’s an element of negotiating a lease that does come down to the human-to-human interaction. At the end of the day, there are certain lease terms that do need a little bit of the human touch.”

Larry Getlen can be reached at lgetlen@commercialobserver.com.