Hotels Race to Optimize for AI Travel Search as Booking Habits Shift

A traveler opens ChatGPT and types: "Calm hotel with west-facing balcony." Within seconds, five options appear. A year ago, this kind of search would have been impossible. Today, it is becoming routine.

As artificial intelligence reshapes how people plan vacations, hotels are scrambling to understand a new and unfamiliar game. The shift is not merely cosmetic. When someone searches Google for a hotel, they see fifty results. When they ask ChatGPT the same question, they get five—and that is all. The difference is seismic. A top ranking on an AI platform is no longer a nice-to-have; it is survival.

The numbers tell the story. In France alone, thirty-five percent of people used AI to find accommodations last year, according to Nicolas Marette, founder of Custplace, a firm that helps businesses optimize their digital footprint. Globally, research from the Boston Consulting Group found that roughly thirty-seven percent of travelers are already using AI-enabled travel sites to plan and book trips. The hospitality industry has noticed. A quarter of major hotel companies now have AI strategies generating measurable returns across their operations.

But visibility in an AI search engine requires a fundamentally different approach than visibility in Google. "What a hotel needs to do to get well referenced by search engines is not the same thing that they need to do to get referenced by artificial intelligence," explained Johanna Benesty, a consultant at BCG. The problem is that AI models do not all work the same way, and hotels must now learn to speak their language.

At Accor, the French hospitality giant that operates Pullman, Sofitel, Mercure, and Ibis properties, executives have spent the past year grappling with this challenge. Nicolas Maynard, the group's chief of AI and data science, described the core difficulty: AI users often search using vague language. Someone might ask for "a romantic hotel in the south" or "somewhere calm and peaceful." Current hotel systems do not classify properties by such attributes. They organize by location, price, room type, amenities—concrete categories. Teaching machines to understand the emotional texture of a place requires rethinking how data is stored and retrieved. "We need to adapt our systems to take semantics into account," Maynard said.

The challenge extends beyond mere classification. Olivier Cohn, director of Best Western France, sees opportunity in granular detail. Hotels could eventually answer questions like whether a power socket sits on the left side of the bed—useful for a guest accustomed to charging devices on that particular side while sleeping. Such specificity is trivial to describe but difficult for current systems to catalog and retrieve. Some hotels are already deploying AI chatbots to handle routine inquiries, freeing staff to focus on higher-value interactions.

Yet the burden does not rest entirely on hotels. BCG research shows that algorithms favor properties with comprehensive, high-trust information drawn from multiple sources. Client reviews and descriptions matter as much as the hotel's own data. A sparse or inconsistent digital footprint—missing details, outdated information, conflicting reviews—will sink a property in algorithmic rankings.

There is another wrinkle. Just as online travel agencies like Expedia and Booking.com charge commissions and premium placement fees, AI platforms are beginning to adopt similar models. BCG predicts that "the familiar OTA commission model will evolve into AI-era distribution fees, charged for prominence and relevance in algorithmic recommendations." Hotels may soon face a new layer of distribution costs, paying not just for visibility but for the algorithmic prominence that determines whether they appear in that crucial list of five.