AI agents are coming for hotel rates

Hotel search
The hotel search and booking path is about to change. (AndreyPopov/Getty Images)

The era of agentic AI has moved from theory to reality. Instinct and Meta Muse are among the new mainstream AI assistants that do not simply answer questions, but act autonomously on a user's behalf, handling tasks like writing and sending emails, shopping, and booking travel.  

A recent Financial Times editorial: ‘Managing a new world of bots and bargain hunters’ described how these AI assistants might champion the financial interests of consumers. 

“It spent the night haggling over your broadband bill, shifting your savings into a higher interest bank account, and cancelling streaming subscriptions just before their free trials turn into paid plans. By breakfast, it has ordered the anniversary present you almost forgot and switched your home insurance to a cheaper provider.” 

Re-shopping 

What does the growing use of AI agents mean for hoteliers and owners? So far, AI agents have tended to book hotels via an OTA, and re-shopping when prices drop is an emerging capability.  

The manual version of this behaviour is already well established. An Ipsos survey of British consumers found that 25 percent had re-shopped hotel bookings and a further 50 percent said they would do so in the future.  

Louise Tawadrous, director of travel, hospitality, leisure at Ipsos, said: “Amongst those who wouldn’t consider [re-shopping], the barriers were functional and not emotional, so things like the time to do so, the hassle, and concerns that something might go wrong.” 

Rate erosion 

So, a bot that re-shops automatically and constantly in the background will be attractive for bookers, much less so for hoteliers. 

“The main consequence of AI agents booking hotel rooms will be erosion of rate and, through commission, of margin,” said Patrick Angwin, founder of boutique hospitality consultancy, Clanalytix.  

A £200 direct booking would be reduced to £164 by OTA commission and then to £147.60 by the AI agent’s re-shopping activity.     

“The problem is made worse by the fact that neither shows up neatly in reporting. Commission creep sits in a broader cost line, and a cancelled-and-rebooked stay simply appears, in most reports, as a slightly lower rate,” said Angwin. 

Just say no 

There are partial defences that hoteliers can take, he suggested: well-priced non-refundable rates, tighter cancellation windows, and less discounting close to arrival.  

Blocking agents altogether is another option which has already been taken by Amazon and Delta Air Lines. 

“Whereas Amazon has the scale to turn agents away and keep much of the sale, an individual hotel that blocks an agent from booking direct is unlikely, in my view, to persuade bot-inclined guests to book direct themselves,” Angwin commented. “More likely, the AI agent will simply book through an OTA, incurring commission, or, worse, the booking will be lost altogether.” 

Direct bookings 

However, if managed correctly, AI agents might be beneficial for hotels and even improve profitability. 

“What I have highlighted is the risk of price chipping if AI agents are not properly controlled and managed,” said Angwin. “But I think companies will come up with strategies to prevent [price chipping] and to avoid OTAs and take more direct bookings via AI agents.” 

Indeed, some hotel technology partners are already pursuing that goal. In October 2026, the PMS provider Cloudbeds launched a direct integration with the AI agent Meta Muse.  

Adam Harris, CEO, Cloudbeds, commented: “When an agent finds a property that runs on Cloudbeds, the booking lands where it belongs: in the hotelier’s own system, as a direct booking, with the guest relationship intact.” 

Visibility 

However, before the booking stage, a key question, especially for independent hotels, is: will the AI even mention my hotel at all? 

A Stiplo study of general AI use (ChatGPT etc.) found that out of 659 Barcelona accommodation providers, 143 of the properties (22 percent) were invisible to AI.  

Independents were eight times more likely to be invisible than chain hotels. Even if a hotel is recommended, the study found that a direct booking link is rarely provided and AIs link to OTAs more often than hotel websites. 

New GEO ecosystem 

Around 15 years ago, the hotel industry realised it was losing visibility and customer relationships to the OTAs. An entire digital marketing ecosystem sprang up around direct booking websites, SEO, metasearch, and reputation management. 

We're now seeing signs of a very similar ecosystem forming around AI discovery and GEO (Generative Engine Optimisation). Stiplo, founded by Carlo Del Mistro, the former chief digital officer at Ennismore, is among several companies offering GEO audits. Others include UP Hotel Agency, INNsight, Arise Hotel Marketing and Publicus. 

However, it is important to note that nobody has fully cracked GEO yet. The major AI platforms do not publish ranking factors and there is no guaranteed formula for being recommended by ChatGPT, Gemini, Perplexity or Google's AI products.  

Factors such as website content, online reviews, and structured data, all play a role, but nobody can say with certainty exactly how AI weighs each signal.  

Wild west 

“It’s still the wild west out there,” said Angwin. “There are lots of players coming in and saying they can help your hotel leverage AI to gain bookings.” 

As a result, much of what is currently being sold as GEO is an extension of good digital marketing practice, focused on creating accurate, structured and authoritative information that AI systems can easily understand and trust. 

“Hotels and brands need to stand out, which is now harder than ever. They need to be sure that both the human and digital audiences understand what’s special about them,” said Tawadrous. “They also need to communicate that more quickly, succinctly and creatively, especially in a distribution network that is hurtling towards commoditisation.” 

Growing usage 

The proportion of hotel bookings made by AI agents is still tiny, but usage is expected to grow rapidly. One in ten people will routinely use AI agents to shop and pay by 2030, according to a Mastercard report. 

“Once people are more confident about the guardrails, about the security and the control they have over their credit card numbers and personal details, adoption will increase,” said Angwin.  

What hotel owners can and should do now, he advised, is monitor how their hotels are presented across AI platforms and make sure they can see what AI agent bookings cost, which means reporting net rate and contribution by channel rather than revenue alone.  

Some quotes taken from the panel ‘What guests want: consumer behaviour, brand relevance and the future of hotel value’ at the Annual Hospitality Conference 2026 in Manchester.