Booking.com's CEO envisions an AI travel assistant that intervenes before problems happen—revealing both the promise and the trust challenge of generative AI in high-stakes services.

What happened

Booking.com CEO Glenn Fogel, stranded on a Denver tarmac during a flight delay, outlined his vision for an AI travel assistant that would have warned him hours earlier of bad weather and offered alternative options like a four-hour drive or car service to his destination. The assistant would execute the entire rebooking and refund process with a single user confirmation. Fogel emphasized that Booking.com's goal is to own the entire trip-booking process—from the moment someone thinks about a getaway through returning home—rather than just selling individual flights and hotels.

Why it matters

Generative AI in travel booking must solve a trust problem that chatbots currently cannot. Real estate broker Ryan Serhant shared a cautionary tale: a $50 million（約80億円） Manhattan penthouse deal nearly collapsed when both buyer and seller asked ChatGPT about the price, and the chatbot gave contradictory answers to each. Serhant salvaged the deal using off-market data that large language models lack access to. Fogel also flagged rising AI-generated fraud—fake listings, fabricated reviews, and convincing photos of properties that don't exist—warning that "just because it reads well doesn't mean it's truthful." Human accountability, which allows clients to sue a broker but not a chatbot, remains a competitive edge.

What to watch

Fogel has already flagged generative AI as a top priority at Booking.com (valued at $125 billion（約20兆円）), telling investors earlier this year the company is focused on "advancing our use of Generative AI to enhance the value we deliver to both travelers and partners." The critical test will be whether Booking.com can build an AI agent capable of acting "not after it's happened, but before it happens"—and whether travelers will trust it to make autonomous decisions about their itineraries.