Artificial Portal Wars (Transmission #304)

With every company on the planet seemingly infusing AI into everything, it’s no surprise we’re seeing natural language models infiltrate real estate search.

Zillow entered the ring in 2023, and this week, the upgraded experience references “additional criteria when searching for homes using natural language, including commute time, schools, nearby points of interest, and affordability.”

According to CTO David Beitel, buyers can “specify details like location or style,” including:

Points of interest (POI): Look for homes near your favorite places. For example, queries like, “Apartments near Denver Union Station,” will yield results that keep you close to the places that matter to you.Commute time: Find homes based on travel time to key locations—not just to work. For instance, you can search for: “Homes 30 min drive from Millennium Park.”Affordability: Set your budget and let natural language search handle the rest. Just type “Homes for sale under $400,000 in Austin” or search by monthly payment, such as “Seattle homes under $4,000 monthly,” to discover options in your price range.Schools: Find homes in a specific school district by searching for terms like “3-bedroom houses near Roosevelt High School” or simply “Roosevelt High School.”

Zillow’s move follows Flyhomes’s acquisition of ZeroDown and subsequent launch of an AI-powered search experience, as well as Tomo’s prior release of an experience where “home buyers can simply describe their ideal home just as they would in a conversation with a friend or real estate agent.”

Let’s get the obvious out of the way: The consumer experience is not there.