What’s your dream home?
A house near the beach? A home with a primary suite that has north-facing windows? Do you fantasize about a home library with a bay window? Or easy access to a nature trail so you can run a few miles before work in the morning?
Mickey Neuberger
Realtor.com’s new natural language home search feature aims to help consumers find exactly what they want, using everyday phrases and questions. The feature currently supports more than 300 terms; however, Realtor.com Chief Marketing Officer Mickey Neuberger said the goal is to support an infinite number of phrases — so consumers can easily spot the listing of their dreams, no matter how specific their needs are.
“The future is more conversational and allows for a high degree of personalization, which really helps the consumer,” he told Inman in an exclusive interview last week. “It’s about making it easier for people to find homes and meeting the new consumer expectation, which is not just in real estate, but in e-commerce, travel, and research, for conversational search. And there’ll be many, many iterations following this.”
Neuberger said the feature is already providing invaluable feedback on what consumers want — with listings featuring in-law suites in high demand — and exposing areas of opportunity for Realtor.com to expand the feature’s algorithm and indexing capabilities.
“We are the real estate platform most trusted by professionals. We stand for that authority, trust, and professionalism,” he said. “And AI is part of making our platform easier to use so users can access the best listing content, data and insights.”
What follows is a version of Inman’s conversation with Neuberger that has been edited for length and clarity.
You kicked off the month by launching a new AI-powered feature that enables consumers to search using natural language. What was the catalyst for creating this feature?
Generally, and not just in real estate, consumers’ expectations for search have gone way, way up. Everyone is using ChatGPT and other AI interfaces in their day-to-day lives, and there’s an expectation that you can engage with a platform using natural language. It’s almost the new normal now, and we’ve been all over [keeping up with consumers’ needs].
I mean, about a year ago, we launched a Dynamic Maps experience that allowed users to view a neighborhood or street and quickly see lot sizes or dollars per square foot. The old way of searching before Dynamic Maps was that you literally had to click through property by property to understand the lot size on a street. And now with this AI-powered search, it’s just another step in the direction to help people find the property they’re interested in much quicker.
Users don’t have to use a bunch of filters to get to the home that is perfect for them, and they don’t have to click through a bunch of pictures because we’ve tagged all the pictures. So, if they’re looking for high ceilings, all they need to do is type high ceilings into the search bar, and the homes with high ceilings will instantly pop up.
As you said, natural language search is now an expectation. My mom, who is 66, and my nephew, who is 10, both use AI-powered platforms to search. One thing that stands out about this new feature is that it, quote-unquote, learns and remembers a consumer’s search habits, so if they take a break for a week or a month, they’ll be able to pick back up where they started.
I imagine that solves a major pain point for consumers.
Yeah. The power of personalization with these new search tools is extremely strong because you pick up so many clues about what the user is interested in. In our feature, once you type in a search, you can refine it over time. It remembers what the search was and allows you to build upon it. When you return to the platform, we keep some of those filters so you don’t have to start over.
We’ve experimented with this feature and the right time frame to retain some of those initial search elements you were interested in. If you stop searching, let’s say for six months, you’re likely interested in something different than what you started with. So, we’re working on that.
Gotcha. That makes sense. So, you launched this feature on the ninth, and we’re now a couple of weeks out from that. What’s the initial data and feedback on how well this feature works? Have you identified any areas for improvement?
Consumers who use the feature are higher intent and tend to be much deeper in the search process. And the most popular search phrases are exactly what you’d expect, focusing on the number of bedrooms and bathrooms and whether a property has a pool. The most common searches are two-bed and two-bath, or three-bed and two-bath, or a house with a pool. Lot size is also a common search, like a mobile home with two acres of land. We also see a lot of people search for homes with an in-law suite, which is interesting.
Obviously, we’re also finding that consumers are using some terms for which we don’t have good results yet. So, we have a team keeping track of common search terms and making sure we can support them, if we don’t already. I would say this is a very fast-moving space, and we’re making new releases every week.
Outside of expanding the search terms the feature has indexed, what are some of the other ways you see natural language search evolving on Realtor.com? Homebuyer demographics are diversifying, so are there plans to support consumers who speak English as a second language?
Right now, we support about 300 terms, and of course, our goal is to support an infinite number of terms. And that requires our team to improve the algorithm to understand users’ language patterns and understand all of the features of every listing. For example, is a listing near a specific park or school? What’s the listing’s architectural style or exterior color? What specific amenities does it have in the kitchen? What is the view? Does it have north or south-facing windows? Those are things that are important to consumers.
We are absolutely working on supporting other languages. I mean, we have a Spanish-language site as well. That segment is the fastest-growing, first-time homebuyer segment, so we will absolutely support them and other groups.
You know, the future is more conversational and allows for a high degree of personalization, which really helps the consumer.
Alongside this feature, Realtor.com also released the results of an AI survey, which revealed that Americans are increasingly looking to AI to start their home search journey. And I think that speaks to this tension that the industry has with AI.
Yes, it can help streamline and optimize, but there’s also a fear of replacement. How do you frame these AI advancements to the agents who leverage your platform, who might be nervous that you’re bolstering something that can cut into their influence? How do you let them know that they’re still an integral part of the search process?
Oh yeah. I understand that. I mean, agents I’ve spoken to love this product because it brings in very qualified consumers. They understand exactly what the consumer is looking for and can help them in that journey. So this is not, in any way, a replacement. This is to complement and deliver high-quality leads to our agents. The best agents are embracing AI and using it throughout the process, from engaging a lead all the way to closing.
Lastly, how does this natural language search feature fit into Realtor.com’s broader technology goals? What does this set the stage for?
We are the real estate platform most trusted by professionals. We stand for that authority, trust, and professionalism. And we’ve been doing this longer than anyone else. We actually invented this space, there’s a lot we’re doing to continue our lead, and AI is part of making our platform easier to use so users can access the best listing content, data, and insights.
And we’re the number one source of consumer real estate news in the country, and we’re continuing to invest there as well through our relationship with News Corp, The Wall Street Journal and Dow Jones.
I think in this new AI-first era, our lead on trust and authority is going to be a big advantage for us.
