When AI Builds the Shortlist, Hotel Marketers Have to Earn a Place on It

A traveler who once typed a few words into a search box and scrolled through ten results now describes the whole trip to an AI assistant and gets back three or four hotels. That shortens the list a marketer has to land on, and it moves the decision to sources the hotel doesn't own. Skift's recap of its Data + AI Summit reports that only 6% of hotels show up in AI search results, and that 82% of AI hotel recommendations draw from OTAs and editorial media.

Devinder Gupta is Global Partner Solutions India Head at Tripadvisor, where he leads India market expansion, strategic alliances, tourism partnerships, and commercial growth. He joined after founding PRYZM International, a consultancy that helps global brands enter the Indian market, and he spent more than two decades in media sales at NDTV, Aaj Tak, Republic Media Network, and Rediff.com. His conversations with hotels, tourism bodies, and experience providers shape his view of what earns a recommendation: how consistently a brand is described, how well it understands the request behind a booking, and whether the experiences travelers ask about can be booked.

“If your hotel isn't among those recommendations, the customer may never even know you existed,” Gupta says. He says the hotels that adjust early to assistants that name only a few properties will have an advantage.

Three or four names

Gupta illustrates the shift with a request a traveler might make to an assistant: “Find me a luxury hotel in Rajasthan for three nights with my family, close to cultural experiences, with a great pool and good vegetarian food.” The response, he says, is that “AI may come back with three or four recommendations.”

Hotels spent the last 15 or 20 years learning how to appear on Google, and the playbook was clear: “SEO, paid search, OTAs, reviews, content and so on.” An assistant answers a long, specific request with a few names, and Gupta says the work of earning one of them can't sit only with the SEO team. It draws on a much broader footprint of brand content, OTA listings, editorial coverage, reviews, destination relevance, and how consistently a property is described across the web. Retail marketers report the same pattern in AI search, which draws heavily on reviews and forum discussions. “I think AI visibility will become as important a conversation for hotel marketers as search visibility became years ago,” Gupta says.

One consistent footprint

Gupta cautions hotels against reading every decision as direct versus OTA. The brand's own website still matters, he says, “because it should be the most authoritative source of information about your property.” A traveler, and increasingly an assistant, forms an opinion from reviews, OTAs, editorial coverage, destination content, and social posts as well.

“You don't completely control the recommendation anymore,” Gupta says. “What you can control is whether the information surrounding your brand is accurate, consistent, useful and gives enough context for an AI system to understand why it should recommend you.” In practice, that starts with checking that the website, the OTA listings, and recent editorial coverage describe the same property in the same way. Brands outside travel already treat that consistency as a discipline, as Bobcat does with one simple story across audiences and platforms.

Gupta says hotels may also rethink the platforms they pay to be on. “Instead of seeing every external platform purely as a distribution cost, some of those platforms may increasingly become signals that influence AI discovery,” he says. He suspects the question marketers ask will shift with that view: “How do I make sure I'm part of the customer's consideration set in the first place?” “That's a very different marketing problem,” he says. “I don't think anybody has completely figured out the playbook yet,” he adds. “We are still quite early.”

The request behind the booking

A recommendation also depends on what the traveler asked for, and Gupta says much of it never reaches the hotel. Historically, a hotel might know the destination, the dates, the room type, and perhaps a loyalty history. Gupta offers a sample of what a traveler might tell an assistant instead: “I'm traveling to Delhi with two children, they love history, my wife is vegetarian, we have Sunday afternoon free, we don't want to travel more than 30 minutes from the hotel and I'd like something the children will remember.”

“That's extraordinary commercial context,” Gupta says. “The hotel may eventually receive a reservation containing almost none of it.” The same blind spot exists in retail, where AI platforms own the discovery and the data about it, so a brand gets the order without the learning.

Gupta describes the fix as “a permission-based intelligence layer capable of understanding intent before, during and after the transaction,” one that turns what it learns into useful recommendations. He sees early pieces in AI concierges, CRM personalization, and conversational commerce, and he says no one has fully solved it. Little Spoon, a children's food brand, offers a retail parallel, using first-party data on a child's age and developmental stage to decide what to send parents and when. For a hotel, Gupta says the question to ask a guest is “What kind of trip are you trying to have?”

Bookable beyond the room

The requests Gupta describes ask for more than a room. The Rajasthan family wanted cultural experiences, and the Delhi family wanted something the children would remember. “The room is important, but very often the memory is created outside the room,” he says.

Phocuswright found that 33% of experiences were booked online in 2025, compared with 64% of travel bookings overall, and Gupta traces the gap to fragmentation. A hotel room comes with a date, a category, an occupancy, a price, and availability. A monument, a wildlife safari, a temple visit, and a sporting event in India can each run on a separate booking system, with its own payment process, cancellation policy, ID requirement, and managing authority. “Being online and being easily discoverable and bookable are not the same thing,” he says.

Marketplaces have solved parts of that, Gupta says, and he sees room for experiences to reach travelers inside a trip they've already started. “How do you bring the right experience into a journey they are already making?” he asks. At PRYZM International, that question is behind PRYZM Trail, which he describes as a way for existing experience inventory to become accessible at relevant points in a traveler's trip, with accommodation as the gateway to it.

“The company that eventually connects identity + intent + inventory + transaction, while earning the traveler's trust and respecting consent, will have a very powerful relationship with that customer,” Gupta says.