Here we go again. Just as the hotel sector starts to figure out how artificial intelligence (AI) search works, along comes another development: the rise of personal AI assistants.
Designed for consumers, the market is booming with recent launches including Booking.com’s Vayo and OpenAI’s dots. Then there is Meta’s Muse, as well as newcomer Instinct, which has just raised $1 billion. Hotels and technology providers are now exploring their potential impact—with one expert warning of “landmines” ahead.
Data distortion
One key concern raised is around a personal AI assistant’s capacity to engage in re-shopping, which is the process of booking a room or flight, then later canceling and automatically rebooking if a better deal comes up.
Booking.com’s Vayo assistant, for example, can collect existing travel confirmations from users’ inboxes and monitor bookings for price drops and availability. Vayo CEO Mike Fridgen also told PhocusWire it is set to roll out transactional booking capabilities and extend automated price-drop capture to hotels. PhocusWire reached out to Booking.com for comment on Vayo's capacity to re-shop.
Instinct, meanwhile, is being used to check restaurant sites every five seconds, informing users about available slots, its founder revealed in an interview. With more than 50% of Instinct’s transactions related to travel, the question is: What happens when room reservations undergo the same treatment?
“It's starting to pop up on everyone's radar,” said Rick Posey, VP of product at Cendyn. “There are a lot of landmines in the process that we're not ready for yet.”
“I can instruct Muse to keep shopping for me for a specific rate in a specific time frame, staying under a price range. That will inflate a hotel’s analytics.”
The result is that a hotel’s look-to-book ratio will rise, and even when bookings are created that will impact “pace”—the term for the rate at which reservations are made for a particular date.
“Pace can look stronger than it actually is because people can instruct the AI assistant to cancel if something else pops up at a lower rate, even somewhere else,” Posey said.
Tracking the volume
Admittedly, there is little evidence so far that re-shopping from these new assistants negatively impacts hotels. Even inbound referral traffic from conventional AI search platforms remains low, as a percentage of total nights being booked. Hotel executives generally quote around 1% for these.
“We are talking low, low percentages of bookings coming through the AI bots,” a revenue management specialist at a global hotel management company, who asked to remain anonymous, told PhocusWire.
“I’ve started looking at this recently, but it’s hard to identify them. We are deep diving into bookings at the moment.”
Safa Rahal, a commercial and AI strategist working in the hospitality sector, agrees.
“We don't see the capacity or the volume quite coming in yet, but we do see it in our Google Analytics referral platforms … references are coming in through AI agents.”
And of course, re-shopping is not exactly new. Platforms like FairFly (rebranded to Oversee) and TripBam (acquired by Emburse) built their business models on this principle. In a similar vein, the likes of Google, Skyscanner and Kayak have offered price tracking for many years.
But it’s the potential scale that lies ahead that worries some industry observers.
“Obviously it is a risk for hoteliers, especially for revenue managers who might need to lower prices because their hotel is empty. Then there's the risk that if the bookings were on a cancelable basis, they would then cancel and get the new price,” said Jordi Serra, CEO and co-founder of marketing and revenue platform 123Compare.me.
But he also believes rising AI usage costs may limit disruptive re-shopping. “A service that checks prices every day is not cheap,” he said. “If the agent is my personal agent, it will process tokens. There will be a moment when you say: This is not a free lunch anymore.”
However, at the Destination AI Summit in Washington D.C., Cendyn’s Posey noted hoteliers were already asking about preparing for AI assistants.
“If you look back 12 months, you start to flag referrals coming from ChatGPT. That was a giant question mark for a lot of people, but now more people are tuned into it, and so we're starting to get some questions—and good ones too,” Posey said.
Those questions include how to tell the difference between an agent and a human—speed to booking was one clue to look out for, Posey said—but more importantly, what can actually be done to counter disruptive re-shopping.
The cancellation question
Re-shopping is for the most part made possible by free cancellation policies, where consumers can cancel an existing booking and then buy a cheaper option without financial penalty.
One response involves hotels tracking the performance of personal AI assistants in the same way they do online travel agencies (OTAs).
“I think this will happen with AI agents, where the agent has credibility in terms of minimal cancellations, where it's booking only when it's confirmed. I think this is how hotels will start to look at it,” Rahal said.
Meanwhile, as revenue management platforms adopt AI, there is more scope to “self-regulate,” according to Gwénaël Merlin, EVP of partnerships and strategic initiatives at hospitality management platform Eviivo.
“There might be a balancing act. There may be a self-regulatory balance. If rooms get canceled, you push the non-refundable rate to counterattack that,” he said.
But with traveler anxiety on the rise, curbing cancellations may not be productive; according to a new global study by Phocuswright, 58% of travelers say they would be prepared to pay a higher rate if the booking came with a free cancellation policy.
‘Don’t fight the personal AI agent’
Further ahead, personal AI assistants could bring more to the table than simply hunting relentlessly for the lowest rate. It’s business after all, and Rahal argues hotels ultimately want to grow direct bookings, no matter where the demand comes from, and that includes AI assistants.
“My prediction is we need to work with this, not against it,” she said.
The “personal” in personal AI assistants points to a bigger opportunity, argues Jason Cincotta, co-founder and CEO at AI demand-channel platform Kismet, a PhocusWire Hot 25 Travel Startup for 2026. Most hotels aim to use personalization to build guest loyalty, increase revenue and stand out from competitors. In theory, consumer AI agents will know the guest better than anyone else.
“Don’t fight the personal AI agent. The opportunity is to offer a brand-new guest experience. A logged-in rate for a known guest, quoted through their AI: It doesn’t break parity, it ties the hold to a person and it gives the agent a reason to stop looking,” said Cincotta.
“The upside is the AI agent knows the guest. Late checkout, a crib, connecting rooms, dinner the first night. Package it and quote it and every rate is bespoke, with nothing to re-shop against,” he added.
Despite being at such an early stage, Rahal thinks the conversations are worth having today. “With every single technology we had, whether it was the OTAs when they came out, whether it was social media, whether it was review sites, when it first comes out, everyone tries to block it. But I think by now we have learned the lesson,” she said.
“I don't think we're even close to the eye of the storm yet. I think it's all building up, and my guess is that it's all going to happen at once where nobody's ready yet.”