It won’t will be a banner year for real estate app Zillow. “We describe the domestic market as bottoming out,” CEO Jeremy Wacksman said in our conversation this week. Last year was disastrous for the real estate market and the situation is expected to improve only slightly in 2026. (If January historic decline home sales rate is an indicator and is even too sanguine). “One way to think about it is that 4.1 million existing homes were sold last year – the normal market is 5.5 to 6 million,” Wacksman says. He hastens to add that Zillow itself is doing better than the entire real estate industry. Still, its valuation is one-quarter of its high in 2021. Hours after our conversation, Wacksman announced that Zillow’s profits increased last quarter. Nevertheless, the next day, Zillow’s stock price fell by almost 5 percent.
Wacksman sees a brilliant spot – artificial intelligence. Like every other company in the world, generative AI poses both opportunity and risk for Zillow’s business. Wacksman prefers to focus on the positives. “We believe that AI is actually an asset rather than a threat,” he said on the earnings call. “Over the last few years, the LLM revolution has really opened all of our eyes to what is possible,” he tells me. Zillow integrates artificial intelligence into every aspect of its business, from how it lists homes to how it hires agents to automate workflows. Wacksman marvels that with Gen AI you can search for “houses near my child’s new school, with a fenced yard, for less than $3,000 a month.” On the other hand, his customers could end up asking the same queries on chatbots powered by OpenAI and Google, and Wacksman has to figure out how to take the next step and go to Zillow.
Throughout its 20-year history – Zillow celebrated its anniversary this week – the company has always used artificial intelligence. Wacksman, who joined the company in 2009 and became CEO in 2024, notes that machine learning is the power behind “Zestimates,” which estimate a home’s value at any given point in time. Zestimates became a viral sensation, making apps and sites like irresistible Zillow has gone crazy— which is what it is TV show on the HGTV network – they’ve built a company around highlighting the most intriguing and quirky offerings.
Recently, Zillow has spent billions aggressively pursuing modern technologies. One of the ongoing efforts is to improve the presentation of homes for sale. The so-called function SkyTour uses an artificial intelligence technology called Gaussian Splatting to transform drone footage into a 3D rendering of the property. (I love writing the words “Gassian Splatting” and can’t believe an indie band hasn’t adapted it yet.) The AI also powers a feature in Zillow’s Showcase component called Virtual stagingthat supplies homes with furniture that doesn’t actually exist. There’s a risky undercurrent here: once you leave aside the authenticity of the actual photo, the question becomes whether you’re actually seeing a credible representation of the property. “It is important for both buyers and sellers to understand the line between virtual staging and the reality of a photo,” Wacksman says. “The virtually staged image must be clearly watermarked and disclosed.” He says he has confidence that licensed professionals will follow the rules, but as artificial intelligence comes to dominate, “we need to evolve these rules,” he says.
Currently, Zillow estimates that only a single-digit percentage of users exploit these exotic viewing features. Particularly disappointing is an outing called Zillow Immerse, which runs on Apple Vision Pro. Following its February 2024 rollout, Zillow called it “the future of home tours” Please note that this is not claimed to be the near future. “This platform has not yet gained widespread consumer traction,” Wacksman says of Apple’s destitute innovation performance. “I think VR and AR are coming.”
Zillow is on a more solid footing, using artificial intelligence to increase the productivity of its employees. “It helps us do our jobs better,” says Wacksman, adding that developers are writing more code, customer service tasks have been automated, and design teams have reduced the time it takes to implement new products. As a result, he said, Zillow has managed to keep employment “relatively stable.” (Zillow did this release some work recently, but Wacksman says it involved “a handful of people who didn’t meet performance requirements”).