When the broker is a chatbot: How AI will shake up commercial real estate. (2024)

Davies admits it wasn't his credentials, charm, or connections that opened the door to such a potentially lucrative assignment. In fact, it wasn't even him — the executive was dazzled by Davies's chatbot.

Brokers, Davies noted, normally "don't play with this" and the high-placed executive was "just really entertained."

Despite the industry's reputation as a sluggish adopter of new technology, many in the commercial real-estate business have embraced AI tools to help draft marketing materials, dissect leases and legal documents, and comb through reams of data.

A little more than a year after the release of ChatGPT, artificial intelligence is rapidly advancing beyond those helpful but mundane tasks, and brokers such as Davies are beginning to deploy the tech more directly in the hypercompetitive field of selling commercial property.

Their early efforts show how the new software could take on a more client-facing role in real estate, potentially remaking a business that has long hinged on human interactions, relationships, and judgment — and begin to reveal the promise and perils of such a radical shift.

"This is as big as when email or e-commerce started," Davies said. "People who are willing to pick it up and use it are going to go so much farther and faster."

Brokers are already using bots to help sell property

Davies recently enlisted a ChatGPT-powered bot to assist him in his effort to sell a 20-unit apartment building in his home city of Edmonton in Alberta, Canada.

The chatbot followed up with roughly 300 potential buyers Davies had emailed the listing to and it fielded the flood of questions that normally ensue, such as the property's asking price, location, and unit count. Armed with the capabilities of generative AI, which learns from patterns, experience, and instruction, the bot could handle more complex queries as well, such as what returns the property might produce if the building's rents were increased by a certain percentage and the costs of different levels of financing.

While that may sound basic, Davies said he was startled by the software's effectiveness. Normally, he might expect 50 responses after emailing a listing to about 1,000 contacts. The bot was able to steer 76 interested parties to him from the pool of 300 he had targeted. Within three days, the property was in contract — a fraction of the time it normally takes to lock down a deal.

"It lets clients feel empowered," Davies said. He believes contacts had engaged with the bot more readily because they could ask it questions flexibly as they came to mind, whether it was 2 in the afternoon or the morning, and not feel sheepish for inquiring about basic information.

Some were clearly curious about encountering the nascent tech. Davies reviewed the text exchange the bot had with the executive at the large apartment owner and saw he had asked it whether it could buy him groceries. Davies called the man the next day and though the deal was too small for his company to consider, the executive expressed admiration that Davies was willing to push the boundaries.

The next time the company had a large property to sell, "he would make sure I got a shot at it," Davies said.

Rod Santomassimo, the founder of a real-estate coaching business called the Massimo Group, has worked with Davies and other brokers to use AI and said he's convinced it will soon become ubiquitous and increasingly capable.

Santomassimo expects that within a few months, brokers he works with will use bots that can call and have conversations with clients and customers rather than just text them.

"We will do it in probably 60 days," he said.

While Santomassimo's bot, for now, sets up meetings with potential buyers who simply continue to converse with it, it could be used to filter stronger bidders from those who are less qualified "before the end of the first quarter," Santomassimo said.

He believes AI assistants will become a common tool for salespeople and executives across real estate and the wider business world.

Armed with troves of data and the ability to speak in a convincingly human manner, bots may soon be able to describe to clients and customers any physical or financial detail of a property in real time. Bots could be used to call hundreds, even thousands or tens of thousands, of property owners to gauge their interest in buying or selling, collecting valuable leads in the span of minutes.

The software may eventually even help choose the winning bidder in a property sale — and help negotiate terms of the deal.

For now, outside of AI's own limitations, there appear to be few laws or industry restrictions to encumber its adoption. A spokesman for the National Association of Realtors, for instance, said that beyond the organization's code of ethics, "NAR doesn't have any specific rules or policies about AI."

Some of the industry's biggest players are hesitant

There are those who are skeptical — including some of the largest and most powerful real-estate companies in the country.

CBRE released its own chatbot called Ellis AI – after the E in CBRE – in June 2023 to give its brokers and employees access to a bot similar to ChatGPT. Among its motivations, said Sandeep Davé, CBRE's chief digital and technology officer, was to create a more secure alternative that would allow its professionals to use the tech without fear of breaching confidential or proprietary data.

Private documents or privileged information that are accidentally or unknowingly uploaded onto platforms such as ChatGPT enter the public domain.

Davé said that brokers at CBRE haven't expressed interest in bringing AI applications into their conversations with clients.

"If we get that sentiment from our clients, then we would explore that, but we are not getting that," he said.

CBRE has focused instead on using AI behind the scenes, asking it, for instance, to probe its vast repositories of data, relationships, and client requirements to help uncover business opportunities.

James Nelson, the head of Avison Young's tristate investment sales group, said his team uses AI to summarize market data, write proposals and listing descriptions, and boil down quarterly performance reviews for each member of his roughly 24-person team.

Having AI begin to interact with his Rolodex, however, feels like a step too far.

"To have a chatbot call an owner on my behalf, to me, I would not feel comfortable doing that," Nelson said.

Even boosters of the new tech's expanded role in the real-estate business acknowledge that chatbots could become viewed as a nuisance, a gimmick, or worse. There have been instances where AI bots have behaved unpredictably, a phenomenon known as hallucinating.

"Will the bot go rogue? That's the big fear," Santomassimo said. "We can't have a bot asking inappropriate questions."

Another worry is the prospect of unscrupulous parties misusing AI to masquerade as others, spread misinformation, smear rivals, or commit fraud or other criminal activity.

Bob Knakal, a prominent New York City sales broker at JLL and a client of Santomassimo's who plans to use AI in his business, said that he, his wife, and his teenage daughter have memorized code words in case there was ever a situation where they suspected a bot was being used to pose as one of them.

Others embrace AI's inevitability

Some have come to see AI's advance as inevitable.

David Dirkschneider, a broker based in Oklahoma who specializes in selling apartment buildings, employed Santomassimo's ChatGPT-powered bot in his rollout of a $9 million,135-unit apartment building about 90 miles north of Oklahoma City in December. Like Davies, Dirkschneider said the system handled much of the tedium of fielding initial inquiries and funneled a smaller subset of potential buyers to him, accelerating the sales process, which is still ongoing.

The element of speed has already given Dirkschneider a leg up on rivals, he believes.

Successful brokers generally won't go to the trouble of marketing a property unless a seller grants them a window of time to arrange a deal. Dirkschneider said the software would allow him to accept shorter exclusive periods than competing brokers who aren't using it.

Dirkschneider doesn't consider himself a tech visionary but said he has come to grasp that there may also be troubling and unpredictable aspects to an AI-powered world.

Senior brokers may need fewer junior salespeople and administrative and support staff. If the software continues to become more lifelike and sophisticated, even experienced professionals like him could be in jeopardy.

"This is and always will be a relationship driven business," Dirkschneider said. But "the more advanced AI becomes, the less some of our services are needed, and that's what can get scary."

Still, he sees little choice but to embrace it.

"I know that if I don't adopt AI, then some 22-year-old kid right out of college is going to," he said. "He may not have the relationships, he may not have the knowledge, but if he's using AI and he gets a little bit of experience behind him, he's going to blow right past me."