Well, well, well. Zillow went and dropped a whole new AI-powered home search experience this week, and the real estate internet has been buzzing about it ever since. Meanwhile, Inman went out and asked actual agents how they're really using AI day-to-day (spoiler: it's less dramatic than the headlines suggest but more useful than the skeptics admit), and one agent-slash-AI-expert wrote a sharp breakdown of exactly where AI still falls flat in a real estate transaction. Big week. Let's unpack it.
1. Zillow Launches "AI Mode" — a Conversational Home Search That Remembers What You Want
Zillow officially unveiled Zillow AI mode this week, a new conversational search experience that lets buyers and renters ditch the old filter-and-scroll approach and instead just... talk. Users can ask things like "Can I afford this apartment if I move in June?" or "Find similar homes within my budget that are closer to light rail," and the platform will serve up listings, affordability estimates, neighborhood insights, and even let you schedule tours — all within the same conversation.
The feature remembers preferences across sessions, adapts over time, and pulls from Zillow's proprietary data and custom AI models. It also includes a built-in Fair Housing Classifier designed to flag discriminatory steering in both user questions and AI responses. The tool is currently in beta with a limited group of users, with a full rollout planned throughout 2026. Zillow CEO Jeremy Wacksman framed it as a fundamental platform shift, saying the company can now connect the entire housing journey — from discovery to keys in hand — in a way that wasn't possible before.
Why It Matters: This is the biggest consumer-facing AI launch in real estate so far. With 70% of U.S. buyers and sellers already on Zillow's apps and sites, this isn't some startup experiment — it's happening at scale on the most-visited real estate platform in the country. For agents, the question isn't whether to panic — Zillow says the goal is to create more informed, higher-intent clients, not to replace professionals. But it does mean the "basic education" part of an agent's job is about to get absorbed by AI. Agents who already lead with strategy, negotiation, and local expertise will be fine. Agents whose value proposition starts and ends with "let me show you some listings" should be paying close attention.
2. What Agents Actually Think About AI — It's Less Dramatic Than the Headlines
Inman went straight to the source this week, talking to agents and industry professionals across the country about how they're really using AI on the ground. The verdict? It's not the revolution the vendors are selling, and it's not the threat the doomsayers are warning about. It's somewhere in the middle — and that middle is actually pretty useful.
Most agents described using tools like ChatGPT and Claude for writing, editing, idea generation, and market research. The time savings are real but modest: less time on routine tasks, more time for clients. One agent called it "one of the greatest opportunities for high-volume teams and solo agents alike to free up time currently spent on repeatable, non-client-facing tasks." Another was more measured, saying he uses AI daily but isn't sure he's using it in the "game-changing way the vendors are promising." The consensus: AI isn't changing what makes a successful agent — it's giving good agents more time to lean into what they already do well.
Why It Matters: This is the reality check the industry needs right now. Between the "AI will replace all agents" crowd and the "AI is just a fad" crowd, there's a growing middle of professionals who are quietly using AI to work smarter without making it their whole personality. If you're an agent who hasn't started experimenting yet, this is your sign — not because the sky is falling, but because your competitors are buying back hours in their day and spending them on clients. Start with the basics: listing descriptions, email drafts, market research. You don't need to overhaul your business. You just need to start.
3. An Agent-Slash-AI-Expert Breaks Down Exactly Where AI Falls Short in a Deal
Tatiana Zagorovski, a St. Louis Realtor who also works in AI, wrote a sharp column in Inman this week laying out the specific scenarios where AI can derail a real estate transaction. Her thesis is simple: AI is great at structured answers in predictable environments, but real estate is full of unpredictability — and that's where human expertise lives.
Her examples are practical and pointed. AI can pull comps but can't interpret why two similar homes sold $75,000 apart. It can't read the body language of a seller who's emotionally attached to their price. It doesn't know that a neighborhood's vibe shifted after a new development went in two blocks over. And it definitely can't navigate the messy middle of a negotiation where incomplete information, human emotion, and timing pressure all collide. Zagorovski's argument isn't anti-AI — it's that agents need to be able to articulate this gap to prospects who think ChatGPT can handle the whole transaction after reading about that Florida man who sold his house without an agent.
Why It Matters: This piece is practically a script for your next listing presentation. As consumers get more comfortable using AI for research and even transaction tasks, agents need a clear, confident answer to the question: "Why do I need you?" The answer isn't "because AI is bad." The answer is "because AI handles information, and I handle judgment." If you can walk a prospect through specific scenarios where AI would fail — missed context on comps, misread negotiation dynamics, overlooked local nuance — you're not fighting AI. You're positioning yourself as the thing AI can't be. That's a much better sales conversation.
That's the wrap, folks. Zillow's making moves, agents are finding their groove, and the smartest people in the room are figuring out where the AI ends and the human begins. That line is where the money is. See you next time.