Booking's CEO says AI doesn't scare Booking. The reason he gives is more interesting than the statement itself

After months of articles in this newsletter about those building AI agents for travel, from Accor's concierge to Meta's Muse, we now hear from the person running the world's largest booking platform, and he says something worth reading carefully. Glenn Fogel, CEO of Booking Holdings, brushed aside in a few words the worry that an AI agent like Instinct could put Booking in difficulty: "if we wanted to do Instinct, how hard do you think that would be? It wouldn't be that hard." According to Fogel, the real problem is not building the technology, which he considers relatively simple, but winning the traveler's trust, particularly when it comes to sensitive data: people prefer to entrust payments and personal information to established companies they already trust, rather than to new platforms, however technologically advanced.

To support this position, Fogel cites a figure worth reporting in full, because it considerably scales down the alarm of recent months: AI-generated traffic today accounts for "significantly less than 1%" of the nights booked on Booking. Over the same period, multi-vertical bookings (flight, hotel and car together) grew at a double-digit rate in the last quarter, more than twice the overall growth rate of transactions. Put differently: while the trade press, this newsletter included, writes every week about new launches of AI travel agents, the real impact on booking volumes remains, for now, marginal.

Hedging isn't naivety, it's a strategy spread across several tables

That doesn't mean Booking is ignoring the topic, quite the opposite. The company is running at least five different AI initiatives in parallel: Priceline's Penny assistant, planning features on Booking.com, Agoda's trip planner, a startup incubated in-house and a not-yet-announced project on the US West Coast, in addition to partnerships with OpenAI and Google for integrating hotel booking. It is a hedging strategy on several fronts, typical of those with enormous resources who prefer to test different paths at the same time rather than bet everything on a single technology or a single partner. It should also be said, in the interest of honesty, that Fogel himself admitted that their "Connected Trip" offering (the attempt to link flights, hotels and other services in a single integrated journey) doesn't yet work the way it should, despite good growth metrics.

Why this voice is worth listening to, with due caution

It makes sense to line up what we have recently written in this newsletter and read it in light of this statement. When we talked about Accor's AI concierge, Sabre's Model Context Protocol or Meta's Muse agent in the piece on Meta and AI for travel, the common thread was always the same: the race of the big players to build the AI interface that will control the discovery and booking of tomorrow's travel. Fogel's voice introduces a useful counterweight, one that doesn't deny the underlying trend but scales down its real speed: the technology to build an AI booking agent is indeed relatively accessible today, so much so that if the CEO of the world's largest platform says it, there must be a reason. The real bottleneck is not technical, it is the trust a traveler has to build before handing their credit card and personal data to a new intermediary, and that is a process that takes time, not just good technology.

For those running an accommodation business, this means something concrete and, in part, reassuring: the transformation of distribution toward AI agents, which we have written about often in recent months, is probably happening more slowly than the pace of press announcements suggests. It is not a reason to ignore the topic, which remains structurally important, but it is an invitation not to chase every single launch with excessive urgency, exactly as we wrote about the volatility of AI visibility on Reddit: better to invest calmly in solid fundamentals (accurate data, multichannel presence, reputation built over time) than to anxiously chase every new platform with the feeling of being left behind.

The limits of this reading

It must be said with the utmost clarity that Glenn Fogel has a direct and rather obvious interest in downplaying the threat posed by AI agents: he leads the company that has the most to lose from a true disintermediation of tourism distribution, so a reassuring public statement is as much part of his job as a neutral forecast. The figure on AI traffic below 1% is self-reported by Booking itself, with no independent external verification, and the definition of what counts as "AI-generated traffic" is not specified in the article, so the number should be taken as indicative rather than as a precise measure comparable over time. That said, the fact that an actor with this visibility into real booking data chooses to stress the current marginality of the phenomenon, instead of riding the alarm to justify defensive investments, is still a signal worth considering alongside, not instead of, the news about competitors' launches.

FAQ

What did Booking's CEO say about AI agents for travel? Glenn Fogel said that building an AI booking agent like Instinct wouldn't be technically difficult, and that the real obstacle is not the technology but the traveler's trust, particularly in handing sensitive and payment data to a new platform rather than to an established company.

How much traffic does AI really generate for Booking.com today? According to Fogel, AI-generated traffic today accounts for significantly less than 1% of the nights booked on Booking, a figure self-reported by the company that considerably scales down the perceived urgency around AI agents in tourism distribution.

What does this statement mean for an independent Italian hotel? It suggests that the transformation of distribution toward AI agents is happening more slowly than the pace of launches and press announcements implies, so it makes sense to keep investing in solid fundamentals such as accurate data and multichannel presence, without chasing every new platform with excessive urgency.

*cover image generated with AI