Real estate is starting to embrace AI. Here’s how it could change buying and selling
The data-intensive real estate business seems like an ideal candidate for big changes brought about by artificial intelligence. Smart software could chew through local market trends to spot buying and selling opportunities, pick the best price for a property, and even offer clients the chance to nab their dream homes as easily as they rattle off a wish list of attributes.
At least, that’s the promise.
Online property giant Zillow is already talking up its plans to use AI to put the right properties in front of customers at the right time while automating away the boring parts of the process, such as booking viewings. Rival Redfin also uses AI for everything from querying a listing to virtually “redecorating” properties online.
But so far, many uses of AI in real estate — like being able to digitally repaint the walls of a home in different colors — are more whiz-bang than breakthrough. And sometimes software doesn’t work as well as good old-fashioned human agents.
Amy Robeson, a real estate agent with Vanguard Properties who focuses on the East Bay, has had to wrangle with Zillow’s sometimes recalcitrant AI technology. Listing a home in Oakland’s Montclair hills, she noted that the property, like many others nearby, had started out as a small cabin decades ago but now included two extensive additions.
Zillow’s technology read the marketing copy and mistakenly listed the property as a cabin. Eventually, said Robeson’s colleague Dana Jongewaard, they had to remove the word “cabin” entirely. On top of that, there was “no one to report it to that we could find,” Jongewaard said.
Still, Zillow’s emerging AI has the potential to improve the exchange of information between clients and agents, which is the overarching goal. “It’s all about making humans more efficient,” said Nick Stevens, the company’s vice president of product for AI.
One key AI-powered feature lets users describe in the Zillow app the kind of home they are looking for with natural language, from location to the types of countertops and appliances, Stevens said. The company takes that information, and other data it has about a people’s preferences based on their saved properties and previous searches, to offer suggestions more likely to catch a buyer’s fancy.
Redfin is also embracing AI with a feature that lets customers ask questions about a listing and get fluent responses, said Ariel Dos Santos, the company’s vice president of product and design. One of the company’s tools allows users to redesign a room in a particular style, or paint it a different color using AI-powered technology. “We get a tremendous amount of engagement with the bot and people returning to the site multiple times,” Dos Santos said.
It’s not about cutting the people out of the equation, said Zillow’s Stevens. Rather, the AI provides agents as much information as possible ahead of time about what a person is looking for to make a speedier match.
Yet Vanguard agent Robeson cautions that such capabilities don’t always work in the Bay Area market, which has a lot of older properties, and can result in buyers missing out on the right house. For instance, a client looking for hardwood floors might not be served houses with hardwood floors hidden under carpeting.
Engaging buyers and offering them choices is relatively safe territory for AI today. But overreliance on the nascent technology can produce costly problems, as Zillow learned in 2021. For a feature called Zillow Offers, the company used AI not only to generate its “Zestimates” price estimates but also to generate initial offers from Zillow to buy some properties from sellers. That resulted in the company picking up homes for prices too high for it to profitably recover via later sales.
The company’s stock swooned and it announced plans to lay off 25% of its staff, or about 2,000 people. More recently, a San Francisco real estate firm called Avenue 8 shut down its brokerage services and laid off staff to focus solely on software. Avenue 8 says it will produce an AI app to simplify buying and selling for consumers and agents, including showing people homes they are more likely to want.
Still, that falls short of AI’s potential to dramatically alter property sales. Venture capitalist Mike Jones, who works for investment firm Science and splits his time between Colorado and Southern California, sees many more opportunities.
He experienced frustration dealing with Zillow this summer when trying to schedule a tour of a Los Angeles property he had his eye on.
What Jones said he really wants is access to the mountains of information companies such as Zillow and Redfin have in their databases to determine when is the best time to buy a house, or what the best mortgage option is.
Jones is no stranger to technology or real estate, having bought and sold properties before. He said that if an app were good enough to track his interest in a certain neighborhood and pair that with the data to alert him to a price dip, it would make it easier to seal the deal.
“As far as an effective search directory, it’s super effective,” Jones said of Zillow. But “I don’t find that their insight is effective.”