Independents 8x More Likely Than Chains to Be Invisible to AI Travel Assistants

Independent hotels are 8 times more likely than chains to be completely unknown to AI travel assistants. That's the headline finding of a Barcelona study presented at Skift's Data + AI Summit in London this month, based on 27,360 AI answers covering 659 accommodations, four AI assistants and three languages.

Carlo Del Mistro and the team at Stiplo ran the study this autumn, asking every question 20 times because AI answers shift between runs. The scale matters: this is the most rigorous look yet at how AI actually recommends hotels.

The distribution is brutal. About 1 in 5 Barcelona hotels was never recommended at all — not in any language, by any assistant. Meanwhile, roughly 20% of hotels captured more than 80% of all AI mentions.

Where does AI get its information?

Here the study hits operators' direct-booking economics directly. About 68% of what AI says about hotels comes from guides and listicles. Only about 7% comes from hotel websites. And only about 7% of AI recommendations link back to the hotel's own site — most point to OTAs (Booking.com most often), TripAdvisor and listicles.

So the direct booking engine you've spent years funding supplies roughly 7% of the picture AI paints of your property. The rest of the funnel runs through channels that take commission.

There's a twist, though. Once an AI knows an independent hotel, it recommends it about as often as a chain. The barrier isn't the product or the brand — it's that the model has never heard of you.

What about AI booking?

AI assistants are moving from recommending to transacting. Google's AI Mode now lets US travelers book hotels without leaving the conversation, and its launch partners are the big chains and OTAs — not independents. Meta's Muse, built into WhatsApp, Instagram and Facebook, is rolling out booking too. It completes hotel reservations through Expedia and Hotels.com, not through hotel systems. When Skift tested Muse on Hilton, Marriott and Hyatt websites, it hit errors before reaching a rate.

Actual AI-driven volume is still small: Booking Holdings says less than 1% of its room nights come from AI, and Bernstein estimates about 1.6% at Marriott. But shortlists form long before bookings do. About a third of AI hotel recommendations come from what the model already "remembers" from training — and that memory is being written now, while volumes are low and few hotels are watching. Hotels that ignored Google in 2008 and OTAs in 2012 spent the next decade buying visibility back at 15 to 25% commission.

What doesn't work — and what does

Hotels are being sold LLMs.txt files, crawler settings, FAQ markup and schema packages as "AI optimization." The Barcelona data found no measurable independent effect from any of them.

What actually moved visibility:

• Google review volume — the strongest single driver
• Backlinks and existing Google rankings
• Placement in the guides and listicles AI actually reads
• Clear content answering real guest questions

AI didn't create this gap. It exposed fundamentals many independents outsourced to the OTAs, and it's widening it.

The practical playbook for an operator follows directly from the data: test your property across ChatGPT, Gemini, Grok, Perplexity, Meta AI and Google's AI Mode using real guest questions, several times each, to see your true competitive set. Make review volume a weekly KPI built into checkout and front-desk routines. Pitch the local listicles the assistants keep citing. Publish plain-language answers on parking, accessibility, quiet rooms, distances and late arrival. Only then add AI tools, once there's visibility for them to amplify.

The question worth asking this week isn't "what's our AI strategy?" It's simpler: when a guest asks AI where to stay in your city, is your hotel in the answer? Hotels that get into AI's memory in 2026 will be the defaults in 2028.