As futuristic as it sounds, artificial intelligence is here. Agents and homeowners are using it, and many more are influenced by its invisible calculations.

As futuristic as it sounds, artificial intelligence is already here. Thousands of agents and homeowners are already using it, and many many more are influenced by the invisible calculations AI makes behind the scenes. Bots, in some form or another, are literally determining everything from home prices to the color of paint would-be buyers see in listing photos.

Numerous companies have deployed AI, but below we’ve gathered together some of the most exciting examples. The list isn’t comprehensive, and it’s also worth noting here that there are different definitions of artificial intelligence and what really “counts” as AI. Is machine learning AI? Are algorithms?

This one post isn’t here to settle that debate, and even the companies featured below had varying answers when asked to define what, exactly, AI means. In any case, the point here isn’t to wrestle with the semantics surrounding these technologies. Instead, it’s to note how technologies that fall broadly within the AI family are impacting the real estate industry.

So, at last, here are some of the companies deploying AI in impactful and exciting ways.

The chatbots and digital assistants

OJO

OJO, which launched in 2017, interacts with homebuyers as they navigate their first steps into the market. After would-be buyers sign up, a chatbot reaches out to them via text message to inquire about their background and home preferences. OJO gathers data on things such as lifestyle, preferred locations, and desired proximity to work and school.

Credit: OJO

 

The technology will also ask users to choose between images of homes and home-related activities as it builds a profile on what they might be looking for. Finally, it uses all the data it gathers to suggest homes and neighborhoods.

The idea was to give buyers resources at a stage when they may not yet be ready to speak to an actual agent.

“In the discovery phase of the home buying and selling process, consumers often are not ready to speak to someone yet and don’t want to be sold by third parties looking to inundate the client as a lead,” John Berkowitz, co-founder of OJO Labs, said last year.

The technology also uses the data it gathers to connect buyers to agents. After a buyer has worked through the process of selecting home preferences, the AI will ask for permission then transfer that person’s information to an actual human agent. The company believes this method keeps buyers engaged and provides agents with more information about their clients than they would normally have.

OJO currently operates in 12 US cities.

Keller Williams’ Kelle

Keller Williams’ artificial intelligence-based virtual assistant Kelle debuted earlier this year and is designed to lower the barriers agents face when interacting with data.

An image showing a market snapshot produced by Kelle. Credit: Keller Williams

Right now, Kelle can fetch contact information, create calendar events, and was recently updated with the capability to pull “Snaps,” cards that feature sales numbers for a local market and an agent or their team. Keller Williams is training Kelle to do things such as predicting the likely outcome of a transaction as well.

The tool also allows agents manage their business from a mobile app. The app is the only way to interact with Kelle, and has been downloaded more than 125,000 times, according to Keller Williams.

It is only available to Keller Williams agents.

The technology is a work in progress, and Adi Pavlovic, the company’s director of innovation, told Inman they are currently training Kelle to perform other tasks, such as reading real estate contracts. And earlier this year, Keller Williams demonstrated its ability to scan video of home interiors then translate that information into searchable tags on listings.

Kelle isn’t designed to replace agents, Pavlovic said, but instead is meant to make them more efficient and bring costs down.

“Our goal is to really just arm the agent with as many tools as possible,” he added.

IMRE

IMRE — which stands for Instant Messaging Real Estate Corp. — provides agents with a chatbot that interacts with prospective buyers. The bot can converse with buyers who land on an agent’s website or Facebook page, or via text message. 

Stephen Jagger, the company’s CEO, said IMRE offers a free version of the bot, as well as versions that cost $10 and $20 per month. The benefit, he said, is that it frees up agents’ time.

“It’s pretty compelling because it’s a low price point, it’s very functional,” Jagger added. “It works 24 hours a day. It never gets tired.”

Agents can monitor conversations between would-be buyers and the bot, and jump in when needed. The bot will also alert agents when it believes human intervention is needed.

IMRE builds some of its software in house, and uses AI tech from Google and IBM’s Watson. The company has thousands of users, Jagger said.

“The main purpose of all of this is to help people do things faster,” he continued. “People won’t wait. They’ll just move on to some other website, some other realtor.”

Structurely 

Structurely is the company behind Aisa Holmes, a chatbot that nurtures and qualifies leads. “Aisa” stands for Artificial Intelligence Sales Agent, and the tool is meant to be a kind of assistant for agents.

An image showing Structurely’s bot interacting with a human. Credit: Structurely

Aisa Holmes can interact with potential clients on a variety of platforms. If someone lands on an agent’s Facebook page, for example, the bot can ping them using Facebook Messenger. When people find properties through an aggregator like Zillow, the bot can text them. Aisa Holmes can also chat with potential clients via realtor.com or email.

The bot uses these interactions to gather information about a buyer’s preferences, and agents are free to jump in when needed.

Structurely CEO Nate Joens told Inman that one of the technology’s greatest strengths is the ability to have “human-like” conversations that touch on things other than real estate. The bot can express condolences if someone mentions a divorce, for example, and will thank veterans for serving their country.

“The most exciting part about our product we hear from our users is our AI’s ability to ’empathize’ with leads in a very human-like way,” Joens said.

The photo scanners

Zillow

Online search heavyweight Zillow has deployed AI across multiple departments, including — perhaps most notably — its Zestimate home valuation feature. The company is training its software to take into account “unstructured data,” or things like countertops and appliances, when it estimates the value of a property. Though the feature has only been deployed in a small part of Washington state, Zillow found that using AI made its Zestimates 15 percent more accurate.

Jeremy Wacksman, president of Zillow brand, told Inman that AI is also learning how to take into account things like the size of rooms or the quality of a home’s view. That data then gets cross referenced with more traditional information about a listing to produce more accurate valuations.

Zillow also uses AI to refine the recommendations it provides users. Wacksman said that AI tracks how users interact with the site, looking at “everything from the homes you share and save and contact on, to even the places where you dwell, where you linger.”

“AI and machine learning do a lot of workload under the hood,” Wacksman said.

Restb.ai

Restb.ai scans listing photos, then figures out what’s actually in them. So for example, it can tell the different between interior and exterior shots, or which room in a house it’s looking at. It can analyze hundreds of images a second, differentiate between 16 different architectural styles, and classify images according to more than 30 different scenes.

The technology can also make what might seem like judgement calls: When scanning photos of kitchens and bathrooms, it will figure out and quantify what condition the property is in. In other words, Restb.ai’s bots can tell the difference between a new kitchen with granite countertops and stainless steel appliances, and an old kitchen filled with formica.

The goal is to “really start shortcutting that whole process of looking for a house,” according to Dominik Pogorzelski, the company’s vice president of product and operations.

Images end up in Restb.ai’s system thanks to requests from other companies, such as those that run listing services or real estate websites.

An Oklahoma MLS has already incorporated Restb.ai’s technology. Credit: Restb.ai

The technology is already available in Oklahoma, were would-be buyers can do home searches based on kitchen or bathroom photos. And Pogorzelski said the company is partnering with other listing services to get the technology deployed in other markets in the near future.

The data crunchers

Revaluate

Revaluate uses AI to turn data into leads for agents. Revaluate clients turn over a database of their contacts, and the company then cross references those contacts with internet search, spending, social, and government data. The goal is to build a profile around the contacts and identify when people might be experiencing significant life changes such as marriages or new children that could foreshadow a move.

When Revaluate’s system identifies a contact who might be getting ready to buy or sell a home, that information is then relayed back to the agent.

Chris Drayer, CEO and co-founder of Revaluate, told Inman that the system is currently 36.5 percent accurate, meaning more than a third of the contacts it thinks are preparing to move actually are. He compared that to cold calling, which he said has only a 2 percent success rate.

Tim Segraves, chief technology officer and also a co-founder of the company, said he hopes to push accuracy numbers up to 50 percent in the future.

Revaluate currently serves customers across the US.

“We have hundreds of clients and they range from individual agents to brokerages to very large mortgage companies,” Drayer added.

Cherre

Cherre’s primary business is connecting data sets related to property for large clients such as banks and insurance companies. By crunching thousands of data points, Cherre can help those clients come up with more accurate property prices, better understand risk, or make informed investment decisions, company founder and CEO L.D. Salmanson told Inman.

“It takes intuition and puts some rigor behind it,” Salmanson added.

Salmanson said Cherre is unique for the volume of data it accesses. If the company is doing work for an insurance or mortgage company, for example, Cherre’s AI could analyze information on past natural disasters, 311 calls, and listing data. It’ll look at traffic data, search data, and even Yelp reviews. Salmanson added that the company even ingests images, and in one example said they’ve been able to correlate the presence of single speed bikes, known as “fixies,” with gentrification.

All of this information then allows Cherre to both describe and make predictions about properties in the future.

“We can out predict any model out there,” Salmanson said.

Though these tools are primarily geared toward large clients, the impacts should trickle down to the broader industry.

“If we do our job really well,” Salmanson said, “the price of mortgages will go down, the price of insurance will go down.”

For smaller clients such as individual brokerages, Cherre also offers marketing tools. Those tools give agents information about properties — such as on renovations, zoning codes, and permitting — then compiles listing presentations with information relevant to making a sale.

Reali

Reali is a flat-fee brokerage that as of this week operates across all of California. The company uses artificial intelligence technology to predict the sale price of homes. The idea is to give prospective buyers better odds of placing winning bids, and company founder and CEO Amit Haller told Inman that about half of Reali clients have used the technology.

Reali’s price predictor tool uses artificial intelligence to help buyers win bids. Credit: Reali

The tool pulls from a variety of sources when calculating likely sale prices, including multiple listing services and historical data. The company also believes the tool, which uses machine learning, will get better over time.

Haller said his company is investing in other AI tools as well. One such tool allows sellers to virtually stage the interior of their home with different colors and furnishings in order to appeal to different markets.

Though the tool currently relies on a mix of human and artificial intelligence, Haller said that the company is working on fully automating the process and turning it completely over to the machines.

“We are working on many different tools in the AI space,” Haller said. “Some of them are to increase efficiency for our agents. Some are to be able to better understand our market.”

Update: this story has been updated after publication to clarify the current capabilities of Keller Williams’ Kelle app.

Email Jim Dalrymple II

Show Comments Hide Comments

Comments

Sign up for Inman’s Morning Headlines
What you need to know to start your day with all the latest industry developments
Success!
Thank you for subscribing to Morning Headlines.
Back to top