The online travel agency will embrace “adaptive content” in what could be a significant evolution in the way people look for hotels. But any changes will be dictated by its customers, rather than hype.
The days of the typical hotel search results page are numbered, thanks to large language models like ChatGPT.
That’s according to Booking.com’s technology chief, who thinks the future lies in “adaptive content” where a uniquely rendered page, based on the intent of the search query, appears.
“People like filtering things at Booking com, but what these (large language) models do well is capture intent quite well,” said Rob Francis, senior vice president and chief technology officer at the online agency, during Skift Future of Lodging Forum on Wednesday.
He cited someone may type in that they want a “cheesy” property. Booking.com wouldn’t necessarily create a filter for that world — but machine learning could find the type of property they’re looking for.
The layout of the results page could also differ based on whether it’s a couple, a family, or a business doing the hunting. “The notion of the page goes away, what best suits this person?” he added.
The page could also reflect the property owner, for example if someone is looking for providers that are “stellar at cleanliness,” Francis said.
Speaking during the “Tech’s Impact on an Evolving Booking Landscape” session, he also said Booking.com had an edge because it has been working in artificial intelligence for a long time.
“At Booking, we have an entire workforce disciplined at doing this at scale,” he told moderator Dennis Schaal, Skift founding editor. Booking.com sold 900 million rooms in 2022, so had an advantage in the amount of data it can analyse and explore.
“That’s where those models get interesting, we think we can augment that. The practice of doing AI well sometimes just comes down to having good operations,” he added.
The tech exec also noted that five years ago it took a pretty special skillset, while now ChatGPT is making it easier, and particularly as new plugins are released. And while the urgency was high to adopt ChatGPT, Francis conceded the agency was rather looking at things for now.
Part of the reason was that it already uses machine learning, for example against fraud and making sure images were correct.
“We think a lot of things can change. We call it adaptive content,” he said.
SALT LAKE CITY — Aundra Morlan is hoping to do a little overseas traveling."Our 29th wedding anniversary, we were going to travel in April. Well, that didn't
Jenny Elia Pfeiffer/Getty Images Taking y
Skift Take Americans are spending a record amount on their travel, and international inbound travel spend finally surpassed it—that's good news for t
I've said it once and I'll say it again: I'd fly strapped to the wing of the plane if it could save me money. Flights are expensive and only getting pricier, an