More and more companies are using AI and machine learning to boost their businesses. Zillow Group is no exception.
“We want to make it easy for people to find homes they love,” said longtime Zillow CTO David Beitel. “And AI can be a huge driver of that.”
GeekWire recently stopped by Zillow’s headquarters in downtown Seattle to meet with Beitel, a member of Zillow’s founding team and the former Expedia CTO.
Zillow is no stranger to AI. In 2006 the online real estate giant debuted the “Zestimate,” an algorithm that uses statistical and machine learning models to estimate the value of homes. It now has a nationwide median error rate of 2.4% for on-market homes, and 7.49% for off-market homes.
Beitel sees opportunity with the recent wave of generative AI advancements, especially given how much data the company manages. It’s also changing the way Zillow developers write code.
“The capabilities are beyond what we’ve been able to do before,” said Beitel.
Speaking on Zillow’s earnings call last week, CEO Rich Barton said “we already see how AI will be fundamental in accelerating our business.”
Read on for key takeaways from our interview with Beitel.
Customer-facing
Zillow is experimenting with AI to reimagine how the company talks with customers to have a more “natural dialogue,” said Beitel.
“It’s about all the great content we have, and how we present it in a way that allows you and Zillow to have a dialogue that helps find the home that’s best for you,” he said.
Zillow is also looking at generative AI as a way to stretch marketing dollars and automate some of the research, planning, and content production that goes into marketing campaigns.
“We have great people that can do this, but now we can enable them to do it faster, get through tests more efficiently, and do them at a higher scale in a more automated fashion,” Beitel said.
Beitel also called out Zillow’s recently launched Showcase 3D tour product that uses AI to create immersive listings.
Engineering
Zillow started testing an AI-powered “copilot” coding assistant tool with about 200 engineers. “Within the first week, others were asking for it,” Beitel said. “You knew this was working.”
Zillow’s engineers are accepting code suggestions from the assistant 40-to-50% of the time. “We get feedback even from our more senior and principal engineers that these suggestions are helpful for them to write their code,” Beitel said.
Real estate agents
Automation and AI can help Zillow customers on their path to buying or selling a home, but Beitel said humans are still crucial to the process.
“We’re not looking to replace agents,” he said.
A majority of Zillow’s revenue comes from Premier Agent, its advertising service for real estate agents.
Beitel said the company wants to give agents AI tools to improve their interactions with Zillow users.
“How do we help them respond really efficiently so that they’re not spending as much time on the tedious tasks and they’re spending more of their time on where their expertise lies?” Beitel said. “That’s the promise of the tech.”
Zillow Offers
Zillow found out the hard way that relying on AI and statistical modeling doesn’t always produce intended results.
The company invested heavily in Zillow Offers, its “iBuying” home-flipping program billed as a way for consumers to avoid the hassle, time commitment and uncertainty of a traditional sale.
But unpredictability in forecasting home prices caused Zillow to shut down the business in 2021 and lay off about 2,000 people. Former AI2 CEO Oren Etzioni referred to “adversarial machine learning” — when a machine-learned model is trained and tested on a historical distribution of data, but used on a different distribution — as a factor in Zillow Offers’ demise.
Zillow still offers a iBuying option via a partnership with Opendoor but it is now focused on building a housing “super app” that addresses various aspects of real estate including buying, selling, and renting.
“Having a long career in tech, you are definitely taking big bets,” Beitel said. “We’re going to have some really big successes and we’re going to have some learnings and some challenges through that.”
“It’s just part of the process. We’re going to keep innovating,” he added. “We’re going to keep focusing on building what we think is best for our customers.”
Editor’s note: Details about Zillow Offers were added to this story.
