AI Can Sell Your Home. Genius Move? Or Colossal Blunder Waiting to Happen?

Did you catch that New York Times tech reporter’s article? Yes, the one detailing his attempt—successful, he assures us—to sell a home with a chatbot instead of an agent.

Human real estate agents, we can imagine, are not exactly thrilled.

Home sales, agents quickly point out, are high-risk transactions. There’s a lot of money at stake. And there’s more. All sorts of emotions come into play whenever someone transfers or receives title to a home. AI doesn’t do emotion. Granted, it does put on a pretty good show of mimicking empathetic language.

Deed transactions involve trust. Authentic trust. That’s critical. AI can gather and analyze figures. AI can report all the facts about a community. But unlike a local agent, AI cannot explain the human elements of a home or a neighborhood.

That said, let’s look at what this New York Times journalist did with AI.

Tech Journalist Takes Public Plunge Into an AI-Assisted Deed Transaction

The New York Times tech writer decided to apply AI to “the intricate world of Hudson Valley real estate.” The idea was irresistible. What a lot of money there was to be saved by avoiding a full-service human agent! After all, wrote the journalist, “rising home prices and soaring agent commissions mean the math is becoming undeniable.”

He gave AI the basic facts about the home. The chatbot produced a listing description. It called the home “flooded with light” and “perfect for bird lovers.” So far, so good.

How would AI handle the all-important step of setting the listing price? A brief look at Zillow showed a valuation of $550K. The writer called a few agents, who largely agreed without offering much analysis. AI’s advice (go higher) turned out to be the right move.

The journalist began by quizzing the Google chatbot, Gemini. He received leads for local photographers and staging help, along with advice on arranging images online effectively. He used Homecoin, which lets sellers list for-sale-by-owner homes on the Multiple Listing Service for a couple hundred bucks. And voilà! DIY listing achieved.

Agents’ requests to see the home rolled in. The journalist put the buyers’ agents’ questions to the chatbot. Each time, the journalist copied and pasted AI’s suggested answer to the agents. It worked. He acknowledges that he writes about tech for a living, which gives him that much of an edge. He used multiple AI platforms, depending on the task. He also recognizes that the context for his particular deed transaction was “a thriving market, with no special circumstances.”

He did purchase an add-on service, which allowed him to access an agent as needed. At the end, he hired a closing lawyer. But without an agent representing him as an in-person negotiator, the journalist felt bolder. He asked the buyers to pay their own agent a commission, and thereby wriggled out of a $12K charge. When all was said and done, the reporter had saved more than $90K. “That includes the premium over the asking price, plus the roughly $36,000 in fees I didn’t pay.”

Don’t Try This at Home (Without the Expertise to Oversee the Technology)

To sell her Colorado condo, Lorraine Schwarz used Ridley. It’s an AI tool that helps homeowners sell without paying agent commissions.

Notably, Lorraine had valuable experience, too. She’d sold other homes. Plus, she was president of her homeowners’ association and therefore had plenty of knowledge to help her set a price and market the home effectively.

She picked Ridley for its capacity to publish her property through the Multiple Listing Service (MLS), as human agents traditionally do. A property listed on the MLS will show up on the websites—Zillow®, Redfin, and so forth—where many home seekers begin their journeys.  

She used ChatGPT to draft the listing text, and then let Ridley guide the Q&A process to create the actual listing for the MLS. It worked. Within 24 hours, Lorraine received a full-price offer: $425K.  

A traditional 5% commission would have amounted to more than $21K. Yikes! The Ridley platform’s pro plan, which Lorraine bought, costs $4K. A bargain compared with $21K.

Is There Anything AI Can’t Do?

An AI tool can:

• Brief prospective buyers so they know a thing or two before interviewing a real estate agent.

• Estimate a home’s value.

• Tell you “how much house” you can afford. (Look for “home search tools” and you’ll see AI tools that point out affordable neighborhoods.)

• Perform comparative analyses among various neighborhoods.

But AI can’t:

• Tell you what a home-value estimate might be missing.

• Walk through a home, notice its smells, sounds, and mood, and identify risks that aren’t obvious from statistical profiles of homes and neighborhoods.

• Understand why it’s wrong to tell a buyer what they want to hear rather than what they need to know.

In short, today’s AI can orient and prepare prospective buyers and sellers. But it doesn’t replace an agent’s on-the-ground experience and judgment.

Know the Risks and Proceed With Caution

AI for home buyers has become a very big deal. According to a 2026 report from AceableAgent, 85% of buyers now use AI tools in their home search.

But be careful out there. AI can make mistakes—and sometimes does. The errors can be expensive.

We also cannot overlook the value of a real estate agent’s fiduciary duties to clients. A paid professional knows the rules and must adhere to a professional code of conduct and the law. (This is the case even though real estate professionals who are non-lawyers may not give legal advice.)

As always, do your due diligence. The New York Times reporter said it plainly: “The chatbot did make one major mistake. I wasn’t using an agent to sell,” so the chatbot said “I should explain in the listing that I was offering 0%. That’s not allowed, under a 2024 settlement that would have exposed me to fines.”

Whoops.

Important note: This article is not legal or financial advice. Contact your legal or financial professional for guidance. AI is not a legal or financial professional. As the Atlanta-based 11th U.S. Circuit Court of Appeals stated in July 2026, AI tools misinterpret the law and even “generate non-existent legal sources.” The same court stated: “Whatever the merits of artificial intelligence, it is no substitute for actual intelligence.”