Hilton SVP and Chief Information Officer Michael Leidinger said large language models have put online travel agencies "under real threat," arguing that chatbots now absorb the aggregation role that once made third-party sites indispensable to hotel distribution.

Speaking Tuesday at the Destination AI conference in Washington, D.C., Leidinger framed the shift as a chance to convert more bookings to Hilton's own channels. The OTAs' original edge, he said, was assembling a fragmented market of independent hotels into a single searchable inventory.

"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.

What does Hilton see in the shift?

Hilton's reading is straightforward. A traveler who feeds a chatbot a stack of filters — neighborhood, brand, rate, amenities — no longer needs a third-party site to do the assembly. That compresses the discovery funnel and lifts Hilton's own site and app.

A McKinsey and Skift Research study of 1,000 U.S. travelers found half had changed a confirmed accommodation after spotting a better option. Leidinger cited that behavior as evidence the funnel is already in motion.

Who pays when AI agents reshop the booking?

The opportunity carries a cost line. Leidinger expects personal agents — including the recently launched Instinct tool and Meta's Muse — to become commissionable channels of their own. Meta has said it will eventually charge for transactions through Muse.

Consumers are already surrendering payment card credentials to agentic tools, a step Leidinger said he did not expect this soon. Once a card is on file, an agent can reshop a reservation until departure, cancelling and rebooking when a lower rate surfaces.

That math worries operators. Shopping requests will rise faster than completed bookings, and each request costs Hilton to process. A 50% re-shop rate implies a query-to-stay conversion that could compress operating margin at the distribution layer.

How is the rest of the industry reading it?

Booking Holdings CEO Glenn Fogel, speaking at Skift Global Forum, dismissed the technical moat. Copying Instinct would be easy, he argued; trust is the harder part. His framing put the competitive question on consumer confidence rather than on which company can ship a chatbot first.

Hilton CEO Chris Nassetta has staked the company's response on scale. He points to Hilton's roughly 25% share of quality U.S. hotel supply as leverage with whichever agent ultimately owns the consumer interface. That share, in his telling, gives Hilton negotiating room on commission rates and merchandising.

What changes for operators?

The near-term picture for hotel commercial teams is mixed. Direct booking share gets a tailwind from LLM-driven discovery. At the same time, the cost of serving AI agents — eventual commissions plus infrastructure to handle continuous re-shopping — lands on the operator's P&L.

Leidinger's core point is that the distribution stack may be renegotiated on terms the chain helps set. Whether that translates into a measurable lift in direct share will show up in Hilton's segment-level channel mix when it next reports quarterly results.