ChatGPT Nearly Derails $50 Million Home Deal as Agents Warn AI Misses Pricing Nuance
Updated
Updated · CNBC · Jun 16
ChatGPT Nearly Derails $50 Million Home Deal as Agents Warn AI Misses Pricing Nuance
3 articles · Updated · CNBC · Jun 16
Summary
$50 million was nearly lost after ChatGPT told both sides of a Ryan Serhant-brokered property sale that the agreed price was wrong, prompting buyer and seller to reconsider.
Serhant said he salvaged the deal by arguing the model lacked context on off-market comparables, buyer intent, replacement costs and other deal-specific factors that shape what a property is actually worth.
Coldwell Banker CEO Kamini Lane said clients increasingly use ChatGPT and Claude to price homes and frame offers, but agents still add anecdotal neighborhood knowledge and can deliver unwelcome advice AI may avoid.
Zillow is pushing a more tailored approach with AI tied to floor plans, 3D home data and agent workflows, underscoring the industry's broader effort to use AI as a tool rather than a stand-alone pricing authority.
Is AI's data-driven valuation a path to fairer prices, or a threat to the art of the deal?
When your AI and agent disagree on a home's price, who do you trust with your biggest investment?
How AI Nearly Killed a $50 Million Penthouse Deal: Lessons on the Limits of Artificial Intelligence in Luxury Real Estate
Overview
A recent $50 million penthouse real estate deal nearly collapsed because of artificial intelligence (AI) advice. The AI provided answers that sounded coherent but were not always correct or suited to the unique property and market. This led to misinformed expectations and disagreements, putting the deal at risk. The incident highlights the dangers of relying only on AI for complex financial decisions and shows why human expertise and judgment remain essential in high-stakes real estate transactions.