AI Meets Real Estate: Zillow's Engineering Chief on the New Home Hunt
When Seattle-based Zillow talks about the future of real estate, the conversation increasingly runs through its engineering organization. Toby Roberts, the company's senior vice president of engineering, has emerged as a leading voice on how artificial intelligence is quietly rewriting the rules of the home transaction — not with flashy robots, but with smarter layers beneath the familiar listing pages. For buyers and sellers across Washington State, the shift is less about novelty and more about how decisions get made.
The most visible change is in search itself. Traditional filters — beds, baths, price range — are giving way to natural-language queries and image recognition that let a shopper describe a "sunny craftsman with a big yard" and get results that match intent, not just keywords. Behind the scenes, machine-learning models are also refining valuation estimates, flagging likely repair costs, and predicting how long a home might sit on the market. The result is a marketplace that behaves less like a static catalog and more like a live, reasoning assistant.
From Listings to Intelligence
Yet the deeper implication is structural. If AI can absorb listing data, public records, and neighborhood signals faster than any human agent, the traditional gatekeeping role of the industry begins to erode. Zillow's own history — including its earlier missteps in iBuying — has taught the company that prediction is powerful but fallible. Roberts' engineering focus suggests a more cautious path: using AI to empower consumers and agents with better information rather than replacing the human judgment that still anchors trust in a high-stakes purchase.
The open question for Washington's housing market is whether these tools will widen or narrow the gap between informed and uninformed participants. Early indications point to a democratizing effect — more transparency, faster comparisons, clearer trade-offs. But the technology is only as good as the data it trains on, and biases embedded in historical records can quietly persist in algorithmic outputs. As Zillow pushes forward, the real test will be whether its AI earns the confidence of a public that has learned to be skeptical of both hype and hidden fees. For now, the home hunt is becoming a conversation — and the industry is still learning how to listen.