How hotels should be thinking about their visibility on AI platforms

Travel companies’ ability to gain visibility in the travel discovery process has become more challenging as users across the globe move from traditional search tools like Google to artificial intelligence (AI)-powered search.

Hospitality software platform Cloudbeds is seeking to bring transparency to the process through research revealing how hotels show up on large language models (LLMs).

Adam Harris, co-founder and CEO of Cloudbeds, likened the situation to the early days of the internet when companies were trying to figure out search engine optimization (SEO).

In “The Signals Behind Hotel AI Recommendations” report, the company chose 145 upscale hotels in different markets around the world and used variable searches and phrases to uncover patterns around why they appear in AI search. 

Cloudbeds chose properties that were already showing up in traditional Google results and maintaining a strong presence on online travel platforms such as Tripadvisor, Expedia and Booking.com.

“We had to think through what are consumers doing and try to get the machine to start talking to the machine in a format that would reveal maybe some consistency,” Harris said. “If there’s one thing we all need to be very clear on, it’s that even the researchers at OpenAI don't understand why one specific answer is coming back."

The company conducted its analysis using tools monitoring ChatGPT, Gemini and Perplexity, the three AI engines that together account for 98% of AI-driven traffic to websites. The study was carried out across six destinations, including Bangkok, Barcelona and London, using “hundreds” of automated queries.

Who comes out on top?

The Cloudbeds study revealed branded properties have a significant visibility advantage over independent counterparts in results surfaced by AI. More than two in three hotels recommended by AI to users are hotels that are branded or come from a large group.

The study also found that online travel agencies (OTAs) are dominant in AI-generated answers. Of sources cited, 55.3% were OTAs, and within that pool, Tripadvisor, Expedia and Booking.com—which already have partnerships with OpenAI and other platforms—led the charge. But there are ways for brands to improve their visibility.

“What didn't surprise us is LLMs are only as good as the contextual intelligence that you give them. So the more understanding or the more phrased approach that met an outcome that we would expect, the closer we were going to get to consistency,” Harris said.

The research also showed that a strong reputation is important—properties that were recommended all maintained “excellent guest ratings” and a high volume of reviews on major platforms. A wide digital presence is key, too. Cloudbeds found 98% of properties recommended were shown on YouTube, 97% on travel blogs and 95% on Reddit, showing that AI considers “broad online visibility” as a quality indicator.

“What we found to be the most consistent, and it makes sense, is hotels that often appear regularly have the strongest reviews, had the strongest brand consistency and had the strongest or broadest, trust system around their digital presence. What I love about what LLMs are trying to do for us is they try to instill trust when we ask them anything. They're trying to bring forth the most trusted source of information.”

There is not a lot of variability in how the three LLMs are surfacing the hotels because, according to Harris, they are engineered the same way and reason the same way. Where you do get a lot of variability is, for example, with Google, which might bring in maps more often.

He also described the slight variabilities “as a bit of a game just like back in the day when we were trying to understand SEO.”

“Small tweaks to these agents reveal tweaks to the outcomes, and we're still in this dark period of unknown. There's going to be a little bit of mystery around the behavior of the outcome.”

How to expand your digital footprint

The biggest takeaway of the research, Harris said, is the need for hotels to provide more information online and for the industry to build “verticalized LLM content enablers that speak hotel.”

A hotel's digital footprint needs to be more than just about the property, it needs to points of interest in the local area such as shopping facilities, attractions and transportation, he said. Harris has already experimented with using a prompt for LLMs to build out a hotel knowledge base because he wants hotels rather than intermediaries to take control of their information. He hopes Cloudbeds will offer the prompt, once it has been refined, as a “freemium option” going forward.

Hospitality market intelligence platform Lighthouse today announced its Connect AI solution to bridge the gap between hotels and AI-powered search and travel planning platforms. The technology is aimed at helping hotels get discovered while also leveraging the Model Context Protocol to help LLMs and AI agents access and understand hotels’ real-time rates, availability and information. The AI engine also facilitates direct bookings from users through AI agents.

Following its research, Cloudbeds has devised strategies to help hotels take the next steps and gain visibility in AI results. The platform recommends:

1. Forming a “strategic” OTA portfolio using niche, regional players 
2. Ensuring websites are accurate, contain detailed information and are able to handle conversational language
3. Prioritizing the management of a property’s reputation across reviews
4. Increasing their digital footprint and focusing on storytelling around a brand in a unique way, using platforms such as YouTube, Instagram and Reddit, to help with AI search visibility. 

Leaning in

According to Harris, hotels are hesitant and unsure of how they should approach this situation.

“You can never go wrong by leaning into what makes you special, the stories, reviews, the local experiences, the diverse forms of mutual content. Generative Engine optimization (GEO) rewards of authenticity, just like SEO rewarded authenticity,” he said.

He said that while creating more content might seem daunting for hotels, the one thing he keeps telling customers and prospects is to create that digital footprint.

Cloudbeds is also using the study findings to further develop its technology, including a new product pulling in user-generated reviews to further build trust between hotels and consumers.

While Harris believes this should rank number one on a hotel's priority list, he doesn't think it has even made the list yet. He said that attracting guests is such a huge expense that anything AI can do to help properties acquire them more efficiently is a good thing.

He also said hotels gaining visibility through AI will help drive personalization for consumers and improve how consumers get matched with properties that best suit their needs. 

“For me, this is the biggest opportunity in 12 years of being in this industry to build a better foundation, AI as an infrastructure layer to enable what I think will dramatically improve the guest-facing relationships as well as the property operational efficiencies.”