When your broker is a bot: Artificial intelligence boosts NYC apartment searches

By Jennifer White Karp  |
January 11, 2021 - 2:30PM

Luke, a chatbot developed by Israeli company RealFriend, had its first sale: A $1.15 million apartment at 537 Court St. in Carroll Gardens, Brooklyn.


Does your New York City real estate agent respond to texts immediately at any hour, day or night? Ask you detailed questions about what you are looking for? Sift through data about truck routes and construction projects and help analyze whether an apartment is a good deal for you? If you answered yes to all three—your agent may drink a lot of espresso—or (more likely) you are engaging with a chatbot.

A chatbot is a new way to boost your apartment search and New Yorkers can currently work with two different, AI-powered bots: There’s Luke, developed by the Israeli company RealFriend, for renters and buyers, and Robin, from Localize, also an Israeli company, which is primarily aimed at buyers.

You could think of working with a bot as the first step towards finding an apartment. Instead of sifting through listings yourself—you communicate via text to Luke or Robin (both avatars are white males with spiffy haircuts—Luke also has facial hair and glasses) and they send you listings that match your parameters, along with encouraging tidbits of information about the listings like this note from Robin, “It’s got low monthlies and hardwood floors.”

After you fine-tune what you are looking for and narrow down your selection, you get the option to be passed on to an affiliate broker. If you’re using Luke, you will get an offer to be paired with an agent vetted by the company. Localize’s Robin is working with agents from Compass, Living New York, and Brown Harris Stevens.

Tech solutions for the pandemic

The use of technology in New York City real estate has been patchy, but the pandemic and the need for social distancing helped force the industry into the 21st century and embrace video tours and online interviews. Chatbots—which never get sick or need to sleep—may be a good fit for renters and buyers who want to do as much of the legwork of finding an apartment online to limit the number of places they trek to see in person.

With the NYC real estate market coming back to life after the shutdown “we saw the demand for people to get advice 24/7,” says Omri Klinger, chief technology officer and co-founder of RealFriend. The company launched Luke for buyers in beta mode in June. “We are automating the best real estate agent in town,” Klinger says.

Chatbots juggle multiple clients

The advantage of a chatbot is its ability to handle multiple requests with the same attention to detail. With Luke, you get quick access to tailored listings.

“You can ask a question at 1 a.m. and Luke doesn’t mind sending another 10 listings,” Klinger says. The bot will scan real estate feeds and come back with suggestions that match your preferences. You can ask questions about the listing, and if he can’t answer it, “a human expert will teach it so Luke learns how to respond,” he explains.

One of the ways these bots differentiate themselves from a typical online search form is that they not only allow you to customize a search but weigh your options—Luke will ask whether a preference is a must have or “a deal breaker,” Klinger says.

For example, you might text, "I’m looking for a two bedroom with a large balcony," and Luke would want to know how important that feature is to you. You can adjust your search and add other requirements, such as outdoor space, natural light, and a short commute. Luke can answer questions about listings and neighborhoods as well—communicating very much like a broker. Luke can also run an analysis and tell you whether you are getting a good deal based on your budget and preferences.

Of course, if you were so inclined, you could pore over real estate sites yourself, then cross check public databases, but it would take thousands of hours, Klinger says.

The service is free to renters. Luke gets a referral fee from buyer’s brokers, who typically get a 3 percent commission. Luke gets 25 percent of that commission.

Klinger says currently they are working with more than 1,000 active buyers and about 2,000 renters per month.

What it's like to buy with a chatbot

Luke’s very first sale was a $1.15 million apartment at 537 Court St. in Carroll Gardens, Brooklyn. When the buyer, Azucena Olvera, was hunting for a place, she clicked on an Instagram ad for Luke and decided to put the search in his hands.

“I was trying to do it myself on the internet and felt I didn’t need an agent to show me what was available,” she told the New York Post.

Same level of service

When you work with a bot, you get the same level of attention whether you are buying on a budget or in the market for a luxury apartment, says Omer Granot, president and COO of Localize.

“When you want to buy a home, the $500,000 and the $5 million buyer still needs the same amount of guidance and support. We help you get it and we help the agent give it,” he says. “It’s a much more efficient way of searching. We think it is a better home-buying experience.”

Robin’s focus is in giving buyers the full picture on every property. It launched two months ago as part of a beta test and has matched 100 buyers to buyer agents.

“In a sophisticated market, everyone has different preferences,” Granot says. “We believe people should have the full data.” That means delivering the bad news along with the good. As Ganot says, “We can tell you the crime rate is high, but a new police station is coming. We can say there is no dog park, but in two years a dog park is coming.”

That sort of granular detail may be familiar to anyone who has ever plugged in a search or on Localize's online apartment search platform, which launched in June 2019. There you can set parameters, such as near parks, low crime, quiet street, low monthlies and more. An interactive map shows you options that fit your criteria, and once you make a selection, you get a report about the address with details about, for example, whether there are any building violations or complaints.

How to get connected

I couldn’t resist plugging in my own address to see that it faces a highway, there are complaints about idling vehicles in the area, that it gets plenty of natural light, and that crime is low for the area—all true.

In doing so, I was automatically connected to Robin, who texted me to get more details about my search. It’s easy to see the appeal of turning over a search to someone (or something) willing to do the work for you—without the pressure of dealing with a real person. (I felt no qualms about ignoring Robin, then testing “Stop,” when it was time to end our brief relationship. Other people get more attached to their bots and chat with them like a real friend.)

For some users, Granot says, they’re not initially aware they are dealing with a bot because the interaction “feels very natural.”

The company will be pairing up with more brokerages in the future, continuing to fuse technology with service from real humans. In other words, the robots will not be replacing humans anytime soon.

“This entire industry is very manual and can gain from introducing new technology,” he says. “This is a huge financial decision. No one is just going to communicate by text and buy an apartment. They need a human at the end—there will always be humans in the process,” he says.



Jennifer White Karp

Managing Editor

Jennifer steers Brick Underground’s editorial coverage of New York City residential real estate and writes articles on market trends and strategies for buyers, sellers, and renters. Jennifer’s 15-year career in New York City real estate journalism includes stints as a writer and editor at The Real Deal and its spinoff publication, Luxury Listings NYC.

Brick Underground articles occasionally include the expertise of, or information about, advertising partners when relevant to the story. We will never promote an advertiser's product without making the relationship clear to our readers.