Hilton's top technology executive says AI assistants have achieved what years of brand advertising could not: loosening online travel agencies' grip on hotel shoppers.

ChatGPT and other large language models have put OTAs "under real threat," Michael Leidinger, Hilton's SVP and Chief Information Officer, said Tuesday at the Destination AI conference in Washington, D.C. His argument is structural rather than rhetorical: OTAs built their position by aggregating a fragmented hotel market, and that function is now replicable inside a chatbot.

"It really, for the first time, puts the OTAs under real threat, because in many ways they were providing those aggregated views in this highly fragmented environment and now, the LLMs have kind of, they're providing that," Leidinger said in an onstage interview, conducted alongside Greg Land, global industry leader for hospitality at Amazon Web Services.

A traveler can now type a stack of requirements into a chatbot and receive a workable itinerary — the exact value proposition that Booking.com and Expedia monetized through search marketing and metasearch placements for two decades. For Hilton, the shift is "much more of an opportunity" to capture bookings directly through its site and app.

The direct-booking math is not clean, though. Leidinger conceded that personal AI agents will likely become commissionable channels in their own right. Meta has already said it will eventually charge for transactions processed through its Muse agent. If personal agents inherit the aggregation role, they may also inherit the commission structure — meaning hotels trade a 15–25 percent OTA levy for a fee set by whoever owns the dominant agent layer.

The cost picture worsens on the technology side. Agents will reshop a booking continuously until departure, canceling and rebooking whenever rates drop. That behavior drives a surge in shopping queries relative to actual purchases, and every query carries infrastructure cost for the hotel chain processing it. Leidinger indicated consumer behavior is already ahead of expectations: travelers are handing agents their payment card details, something he did not anticipate this soon.

The reshopping risk is measurable. A McKinsey and Skift Research study of 1,000 U.S. travelers found that half had changed a confirmed accommodation booking after seeing a better option. If agents automate that behavior at scale, hotels face a future of constant churn in confirmed inventory, compressed rate integrity, and query volumes that grow faster than revenue.

Competitors are calibrating their responses. Booking Holdings CEO Glenn Fogel said at Skift Global Forum that copying an AI agent like Instinct would be easy — trust is the harder part. His comment frames the coming contest as one of brand and reliability rather than technology, which incumbents on both the hotel and OTA sides may find reassuring.

Hilton CEO Chris Nassetta is betting on scale as the hedge. He has cited the company's 25 percent share of quality U.S. hotel supply as leverage with whichever agent wins the distribution battle — an argument that Hilton's inventory is too large for any booking channel to bypass.

The stakes for sellers of travel are straightforward. If LLMs erode the OTAs' aggregation advantage, commission flows and customer-ownership shift — but not necessarily to hotels. The likely outcome is a new intermediation layer of commissionable agents, higher query-to-booking ratios for suppliers, and distribution leverage concentrating in whoever builds the most trusted consumer agent. Leidinger's comments signal that Hilton expects to enter that period with more direct traffic and a quarter of U.S. quality supply as its bargaining chip.