Cloudbeds has released a new report titled The Signals Behind Hotel AI Recommendations. The study is the first in the hospitality industry to examine in detail how generative AI platforms identify and recommend hotels to travelers.

The report comes as generative AI tools begin to change how people search for accommodations. Rather than reviewing long lists of search results, many travelers are now relying on AI assistants to deliver a single, curated answer. As this shift gains momentum, understanding the mechanics behind AI recommendations is becoming increasingly important for hotels looking to maintain visibility and attract bookings in a more competitive, algorithm-driven environment.

To better understand these dynamics, Cloudbeds researchers conducted hundreds of AI-generated queries across six international destinations. The analysis focused on 145 hotels that consistently ranked in top recommendations from leading platforms, including ChatGPT, Perplexity, and Gemini—together accounting for an estimated 98% of AI-driven web traffic. Researchers evaluated factors such as source citations, sentiment scores, brand affiliation, digital visibility, and OTA strategy.

The findings suggest that OTAs are a primary source of content feeding AI responses, accounting for 55% of citations. Among these, Booking.com was the most frequently cited, followed by Expedia and Tripadvisor. Hotels with strong OTA presence and well-maintained listings were more likely to be surfaced in AI recommendations.

Brand affiliation emerged as another significant factor. Of the properties identified in the study, 72% were affiliated with major hotel groups or chains. These branded hotels had, on average, a 4.43 percentage point advantage in visibility compared to independents. This advantage appears to be linked not only to scale but also to more consistent digital content and guest feedback management.

Reputation was consistently strong across all top-ranked hotels. Every property analyzed had high review volumes and strong guest ratings across platforms like Google, Tripadvisor, and OTA review sections. The average sentiment score among recommended hotels was 75 out of 100, suggesting that AI platforms heavily weight both volume and tone of guest feedback.

A broad digital footprint also correlated strongly with AI visibility. Nearly all recommended properties showed up across multiple content channels: 98% appeared on YouTube, 97% were mentioned in travel blogs, and 95% had some presence on Reddit. This suggests that AI models draw not just from structured data sources like OTAs, but also from broader web content, including social, video, and user-generated commentary.

In addition to identifying key visibility drivers, the report (available for download) outlines five strategies hotels can implement to improve their presence in AI-driven recommendations: maintain a strong and diverse OTA portfolio, ensure hotel websites are accurate and up to date, actively manage online reviews, expand reach across content platforms, and develop a clear and differentiated brand narrative.

Related