The housing market is a mess. If you’ve tried to buy or sell a house recently, you know the feeling. It’s not just you. Zillow’s own CEO, Jeremy Wacksman, candidly described it as “bouncing along the bottom”. In an interview with Wired, he pointed out that only 4.1 million existing homes were sold last year in the US. A healthy market? That’s closer to 5.5 or 6 million. When the core of your business depends on people moving, and people aren’t moving, you have a serious problem.
Zillow’s valuation is a stark reminder of this, sitting at a quarter of its 2021 peak. So, in the face of a stagnant market and existential threats from every corner of the tech world, what’s their big move? They’re going all in on artificial intelligence. This isn’t just about tweaking a few lines of code; it’s a full-blown real estate AI transformation that Zillow believes is an “ingredient rather than a threat”. But is it a life raft or just a shinier anchor?
The New Digital Kerb Appeal
For years, Zillow’s game was aggregation. It collected all the publicly available property listings and put them in one place with a slick interface. That was its moat. But in today’s world, that’s not enough. The company is now weaving AI into every part of its platform, attempting to create an experience so rich that you wouldn’t dream of starting your property search anywhere else.
It’s about evolving beyond a simple list of houses. Think of it this way: the old Zillow gave you the blueprints of a house. The new Zillow wants to give you a holodeck simulation.
– Virtual Staging: Selling an empty house is tough. It lacks soul. Zillow’s AI can now digitally furnish an empty room, transforming a sterile box into a cosy living room or a chic home office. As Wacksman notes, these images must be clearly watermarked, so no one thinks the Eames chair comes with the house. It’s a clever piece of property tech innovation that solves a real-world problem for sellers.
– Immersive Tours with Gaussian Splatting: Forget clicking through a gallery of static photos. Technologies like Zillow’s SkyTour, which uses an AI technique called Gaussian Splatting, create fluid, three-dimensional models of a home from a handful of images. It’s the next step in making you feel like you’re actually there. While Wired reports that only a small percentage of users are currently using these features, it’s a clear signal of where the housing market AI is headed.
This constant Zillow algorithm evolution is designed to build a stickier ecosystem, making the platform more than just a search portal; they want it to be an indispensable tool for dreaming, planning, and transacting.
The Ghost in the Machine: Prediction Models
Let’s not forget Zillow’s spectacular failure with iBuying, where it tried to flip houses at scale. The company lost hundreds of millions of pounds because its market prediction models couldn’t keep up with a volatile market. It was a painful, public lesson in the limits of algorithms when faced with real-world chaos.
So, has Zillow learned its lesson? It seems so. The company has shifted its strategy. Instead of using AI to buy houses itself, it’s now using AI to provide better information to the humans who do. The new and improved Zestimates and market analytics are the result of that very expensive education. The goal is no longer to be the market, but to provide the best possible map of the market for buyers, sellers, and agents. This refined focus on data analysis is at the heart of their new strategy.
The Barbarians at the Gate
Here’s where the strategy gets really interesting, and frankly, a bit desperate. Zillow built its empire by being the front door to real estate data. But what happens when someone builds a teleporter?
That teleporter is generative AI. Why would you open the Zillow app when you can simply ask an AI assistant from Google or OpenAI, “Show me all the four-bedroom flats in Islington with a south-facing balcony under £1 million”? These AI chatbots threaten to bypass Zillow entirely, pulling data from across the web and serving it directly to the user. They become the new aggregator, relegating Zillow to just another data source.
Google is already experimenting with this, partnering with HouseCanary to show listings directly in search results. This isn’t a hypothetical threat; it’s a live fire exercise on Zillow’s turf. The public nature of real estate listings, once Zillow’s greatest asset, has now become a vulnerability. Anyone can access the data; the real battle is for control of the user interface.
Imagine Zillow spent a decade building the world’s most magnificent and comprehensive shopping centre for houses. Now, Google has just handed everyone a personal shopper that can instantly pull any item from any shop without them ever needing to walk through Zillow’s main entrance. That’s the strategic challenge they face.
What’s Next on the Horizon?
The real estate AI transformation is only just beginning. We’re moving towards a future where AI doesn’t just show you the house but helps you through the entire gruelling process.
Think about AI agents that can schedule viewings, help you compare mortgage rates from different lenders, and even review initial legal documents for red flags. For real estate agents, AI could automate the drudgery of paperwork and lead generation, freeing them up to do what they do best: negotiate deals and advise clients.
Zillow’s endgame seems to be positioning itself as the central operating system for this new reality. By offering a suite of AI-powered tools for everyone involved—from the casual browser to the serious buyer and the professional agent—it hopes to embed itself so deeply into the transaction that bypassing it becomes unthinkable. Their Zillow Immerse app for the Vision Pro is a perfect example of this. It’s a bet that the most immersive and integrated platform will ultimately win.
The question is whether they can build a moat of user experience that’s strong enough to fend off the sheer scale and power of giants like Google. Zillow is in a race against time, trying to innovate its way out of a pincer movement from a slow market and fast-moving competitors.
They are making smart, necessary moves. But are they enough? Is this flurry of AI activity a sign of a company confidently building the future, or is it the frantic thrashing of a company trying to stay relevant? Let me know your thoughts in the comments below.
