Booking Holdings CEO on agentic travel: the AI executive assistant that rebooks your whole trip automatically

Summary

Glenn Fogel has run Booking Holdings since nearly its founding — he joined Priceline a week before the NASDAQ peaked in March 2000, watched the stock fall below $1 and survive a reverse split, and has been there ever since. That history shapes how he reads the current AI moment: as a genuine platform shift, but one that requires balancing investment in new capabilities against the cash flows investors are already counting on.

Booking Holdings sits at roughly $170–175 billion in market cap, which Fogel says is about twice the next-largest travel company — he names Marriott as the likely number two. The biggest airline in the world is around a third of Booking's market cap. Expedia is about a fifth. That scale, and the thousands of engineers behind it, is Fogel's baseline argument for why the company can build its own agentic stack rather than simply distribute through third-party AI platforms.

The agentic travel pitch

Fogel's clearest product vision is a travel experience that behaves like a world-class executive assistant — one that knows your preferences without being told and handles cascading disruptions automatically. His specific scenario: your meeting runs late, you're going to miss your flight, and instead of you scrambling, the system detects the highway delay, switches you to a later flight, adjusts your car pickup, moves your dinner reservation, and — because Booking owns OpenTable — finds a comparable nearby restaurant if the original is full. The hotel gets a heads-up to have your preferred bottle of wine waiting when you arrive.

"The technology is there," Fogel says. "We just have to stitch it together. That's the hard part."

About 65% of Booking's customers already come direct rather than through paid acquisition channels, which Fogel attributes to the platform knowing their preferences and making repeat booking frictionless. He sees LLM memory as a significant accelerant to that loyalty dynamic — the ability to store and act on granular preferences over time is exactly what he wants to build into Booking's own infrastructure.

Distribution question

On whether the agentic travel experience sits inside ChatGPT, Gemini, Siri, or on Booking.com itself, Fogel's answer is deliberately non-exclusive. He says Booking is working with every major frontier lab — "there's not a single one we haven't either signed a deal with, done a deal with, or are making progress with" — while simultaneously building internally. The framing tracks closely with how Shopify's Harley approached the same question: win the transaction layer regardless of where the discovery happens.

Near-term vs. long-term

Fogel is measured about timing. Complex travel itineraries involve enough variables — preferences, real-time logistics, restaurant availability, hotel amenities — that fully automated booking is not imminent. What is already working, he says, is AI-powered customer service that resolves issues almost instantly, which he argues builds loyalty more reliably than anything else because customers don't have to wait on hold.

On voice agents and the trajectory toward agent-to-agent transactions with no human in the loop, he sees it as the direction of travel, but frames today's value as the middle ground: AI that can navigate legacy phone trees and deterministic systems fast enough to actually solve the problem.

Live events and scale

Fogel's take on the Taylor Swift effect and World Cup travel is a useful illustration of what it means to be the market rather than a player in it. Events like the 2026 World Cup in the US, Canada, and Mexico will benefit host cities, he says, but for Booking as a global platform it's largely a zero-sum redistribution — travelers who come for the World Cup would have traveled anyway, so aggregate bookings don't meaningfully change. At Booking's scale, demand shifts between destinations rather than creating new demand.