We analyzed 20,000+ AI hotel recommendations. Here's what we learned.

About this research

This analysis examined 20,000+ AI prompts across 30 major European cities, analyzing 100 separate AI response runs using Perplexity and chatGPT. The scope was rigorous: we focused exclusively on entities that generate booking revenue (hotels and OTAs), excluding review platforms, meta-search engines, and information aggregators.

But here's what separates this from simple mention-counting: We didn't just look at how many times a hotel or OTA appeared. We built a comprehensive scoring system that accounts for context, emphasis, position in responses, sentence location, and citation intensity.

A mention in an opening sentence of a response carries different weight than one buried at the end. A hotel mentioned as "my top recommendation" vs. "you can also try" creates vastly different booking outcomes. This wasn't a spreadsheet exercise—it was a gargantuan effort to reflect actual AI system mechanics and real booking revenue capture.

This methodology reveals what simple counting misses: the actual economic impact of visibility.

The core finding

Hotels capture only 47% of booking authority. OTAs appear in just 16% of mentions but capture 53% of booking authority. Why? Because OTA mentions are booking instructions while hotel mentions are descriptions. When AI says, "try the Savoy," a traveler search for it. When AI says, "book through Expedia," a traveler goes to Expedia. The difference is directional—one leads to a choice, the other leads to a booking transaction.

But here's the opportunity: Individual hotels that build sufficient booking authority can capture 50-70% direct bookings. Mid-range chains without differentiation will capture less than 15%.

What determines booking authority

The study identified six factors that predict whether AI systems will recommend you (and how prominently):

Iconic/Heritage status (28% of booking authority)

Hotels with 50+ year history or cultural significance get mentioned 4x more frequently and get 4x higher authority per mention. This matters. A 193-year-old hotel generates 60x more authority than a hotel in a big chain.

Design/Architectural Distinction (19%)

If you're new, this is your path forward. Virgin Hotels Edinburgh opened recently but achieves equal authority to century-old hotels through award-winning contemporary design. Design-focused travelers specifically search for properties they saw on Instagram.

Location/City Symbolism (16%)

Being near a landmark isn't enough, you need to become synonymous with that landmark. Hotel Sacher Wien doesn't just sit near the Vienna State Opera; it is the Opera district in travelers' minds.

Category-Specific Positioning (15%)

"Nice 4-star hotel" is invisible. "Best romantic hotel in Prague" offers 55-65% direct bookings. Own a specific travel occasion or traveler type. The Witchery by the Castle succeeds because travelers specifically search for romantic Edinburgh hotels.

OTA Integration/Market Availability (14%)

This seems counterintuitive, shouldn't strong direct bookings mean avoiding OTAs? No. Complete profiles on Booking.com, Expedia, and Hotels.com with current pricing and real-time inventory signal to AI that you're legitimate and available.

Media Presence/Content Authority (8%)

Every travel publication mentions that feeds AI training data. Travel writers shape future AI recommendations. Build relationships with travel journalists. Pursue design publication features. Every article compounds your authority.

Where Hotels Win and Where They Lose

The study analyzed six different query types. Your strategy should differ by type:

Where you can compete on:

• Location-specific queries ("Hotels near Colosseum"): 50-60% direct capture
• Occasion-based queries ("Best romantic hotels in Prague"): 55-65% direct capture
• Feature-specific queries ("Rooftop bar hotels"): 45-55% direct capture (if distinctive)
• Luxury positioning: 30-50% direct capture

Where OTAs win:

• Budget queries: over 95% OTA capture
• Generic positioning: over 80% OTA capture

Brutal reality for different property types

The study reveals distinct tiers of different paths to revenue:

Ultra-luxury heritage hotels (Hotel Sacher Wien, The Balmoral): 60-70% direct bookings. You're already winning. Protect your position through impeccable reputation and continued media engagement.

Luxury design/boutique hotels (Virgin Hotels Edinburgh, The Dylan Amsterdam): 45-55% direct bookings. Double down on design distinctiveness. Get featured in design publications. Build influencer partnerships.

Luxury international chains (Four Seasons, Ritz-Carlton): 30-40% direct bookings. Problem: Your brand helps, but individual properties get diluted by scale. Solution: Create location-specific branding and property distinctiveness within the chain. "Four Seasons Prague" with local design story beats generic "Four Seasons."

Independent regional hotels: 25-35% direct bookings. Own your neighborhood. Build local media relationships. Emphasize what makes you unique.

Mid-range chains (Marriott, Hilton, Radisson): less than 15% direct bookings. This is structural. You can't compete with distinctiveness at scale. Either specialize into a niche (Marriott Autograph Collection) or accept you're primarily OTA distribution. Your competitive advantage is operational excellence and loyalty programs, not booking authority.

Budget chains: less than 5% direct bookings. Accept OTA dominance. Compete on value and optimization for OTA algorithms.

Actionable takeaways

Stop competing everywhere. Own a specific vertical: location, occasion (romantic, family, business), or positioning (design-forward, culinary-focused, cultural). Build authority there first.

Invest in what makes you distinct. If you're not heritage, become design-distinctive or occasion-specific. If you're not central, own a neighborhood or unique experience. Repositioning takes time—start now.

Maintain complete, current OTA profiles. It signals legitimacy to AI.

Build travel media relationships like revenue-generation. Every article feeds AI training data. This is not marketing; this is booking authority infrastructure.

Accept that mid-range chains face structural disadvantages. If you're a generic mid-range property, direct bookings under 15% are realistic. Optimize for OTA rather than fighting it.

For new hotels: Design distinction is your path to booking authority. Make spaces worth mentioning in design publications.

The real insight

The industry is shifting from an OTA-dominated era to an AI-dominated one. Hotels that understand that they're now being discovered by AI rather than by Google or OTA algorithms can compete effectively. Those that don't - will become increasingly invisible to the systems that now drive booking decisions.

Your strategy must change. Your visibility architecture must change. The time to start is now.