It was the kind of trophy asset that’s infamously hard to price because it’s impossible to find comparisons. After what Serhant described as a “contentious” back-and-forth—he likened it to dueling “kings of the world,” with the buyer and the seller each wanting to win—the deal sheet went out at $50 million flat. Then, at the 11th hour, it nearly died.
That’s because the buyer, Serhant said, went to ChatGPT and typed a version of “I’m looking to buy this, is $50 million too much?” The chatbot said yes.
The buyer’s broker then called Serhant to pull out of the deal because AI said it wasn’t worth it. Unsurprisingly, Serhant’s reaction was pretty blunt, telling the broker the move was “dumb” and “stupid.”
Serhant recalled telling the buyer’s broker “your client’s incredibly smart and wealthy, isn’t he using the data? He’s like, ‘I don’t know what to tell you, man. Super intelligence just told him, ‘Don’t do this, it’s not worth it.’”
So then Serhant had to relay the bad news to his client, who did what “anyone would do in that situation,” and turned to ChatGPT too.
The client asked ChatGPT the inverse question: “I have a buyer that no longer wants to spend [$50 million] because you told him not to. Is $50 million too little? And ChatGPT said, ‘You know what, you’re right, it is.’”
To salvage the deal, the fix wasn’t using more AI. It was using old-fashioned research like “off-market context and data that LLMs can’t scrape,” Serhant said.
He also went on to post a video about the debacle on social media, which he said racked up 3 million views in about three hours. Both clients saw it, both came back to the table, and the deal got done.
Like travel agents, realtors were once the “gatekeepers” of information. They had access to MLS listings that consumers couldn’t find on their own, so buyers had to be much more “dependent” on their agents to even start house hunting. But now, he argued, that information is more readily available.
Unsurprisingly, real estate agents beg to differ. Serhant, for example, said real estate agents are even more important to wealthier clients because they want to be told what to do, have someone to defer to, and if something goes wrong, someone to blame. AI can’t absorb that, he said.
“People hate being sold,” Serhant said. “But they love shopping with friends.”
