Celebrity Broker Ryan Serhant Explains Why AI Almost Cost Him a $50 Million Sale
I’ve worked with deals that seemed solid right until the moment someone panicked. But I’ve never seen a $50 million sale nearly collapse because a buyer decided to trust a chatbot over a decade of negotiations.
That’s exactly what happened to Ryan Serhant at Fortune’s Brainstorm Tech conference when he shared how ChatGPT almost nuked one of his firm’s biggest deals.
And the way it went down? It’s the clearest lesson yet about what AI can and cannot do.
The 11th-Hour Question That Almost Killed Everything
Serhant was selling a trophy New York City penthouse. It was the kind of asset where finding a comparable is basically impossible.
The buyer and seller had been locked in what he called a “contentious” back-and-forth. He described it like two “kings of the world,” each trying to win the negotiation.
After weeks of this, a deal sheet went out: $50 million flat.
Then, at the 11th hour, everything almost fell apart.
The buyer opened ChatGPT and typed: “I’m looking to buy this. Is $50 million too much?” The AI responded yes. Within hours, the buyer’s broker called Serhant to pull out. The deal was dead because an algorithm said so.
Serhant’s response was blunt. He told the broker the move was “dumb” and “stupid.” But here’s where it gets worse.
The Inverse Question That Proved AI’s Real Problem
When Serhant had to break the news to his client (the seller), his client did exactly what anyone panicked would do. They opened ChatGPT themselves.
They asked the opposite question: “I have a buyer who no longer wants to spend $50 million because AI told him not to. Is $50 million too little?”
ChatGPT said yes.
Same tool. Opposite clients. Opposite answers. This wasn’t a bug. It was the system working exactly as designed, which is to say, it was broken for high-stakes decisions.
As Serhant put it: “AI models know the history of the internet. They don’t know the path forward, and they don’t know what the internet, Reddit, and Zillow don’t know.”
How Serhant Fixed It (And What It Says About Real Estate)
The solution wasn’t more AI. It was old-school intel. “Off-market context and data that LLMs can’t scrape,” Serhant said. He also posted a video about the debacle on social media. It racked up 3 million views in three hours.
Both clients saw it. Both came back to the table. The deal closed.
This story sits at the heart of a much bigger debate: Are real estate agents becoming obsolete, like travel agents? Andrew C. Spieler, a distinguished professor at Hofstra University, told Fortune that agents were once “gatekeepers” of information. Now that data is public, their value is shrinking.
But Serhant pushes back hard. Wealthy clients don’t want fewer agents. They want someone to defer to, someone to blame if things go wrong. AI can’t absorb that responsibility.
“People hate being sold,” Serhant said. “But they love shopping with friends.”
The Real Lesson
This wasn’t just about one penthouse. It was about the limits of automation in a world where context, relationships, and trust still move nine-figure deals.
Just like historic properties with stories that algorithms can’t quantify sell on narrative and insider knowledge, so do trophy assets in every market.
The difference between a closed deal and a collapsed one wasn’t smarter technology. It was a human who knew what the internet doesn’t.
What’s your take? Have you seen AI make a deal worse instead of better? Drop your experience in the comments. I’d love to hear it.
Keep the conversation going. Follow us on X and join our Facebook community where we break down real estate trends, AI hype, and deals that actually close. Real estate insights from people who understand the game.
If you’re navigating a property decision or just trying to understand how real estate actually works beyond the buzz, join our WhatsApp community for updates straight to your phone.
Disclaimer: This article is for informational purposes only and based on publicly available statements from Ryan Serhant at Fortune’s Brainstorm Tech conference (June 2026).
Build Like New does not have any affiliation with Ryan Serhant or Serhant brokerage. Individual real estate situations vary, and decisions should always be made with qualified professional guidance.
