Real insights from our research into AI travel recommendations
10 case studies found
Across 2,000+ hotels and 30,000 AI mentions, this case study shows how Google Reviews and Ratings modestly, but meaningfully, influence whether properties get surfaced in ChatGPT, Gemini, and Perplexity — and what hotels can do about it.
Gemini consistently cites OTAs in 37–45% of its travel-related answers across all model versions. Booking.com leads by a wide margin (80%) — with Expedia and other OTAs trailing far behind.
Evidence from Hotelrank’s 2025 dataset shows GPT-5 mini and nano models systematically amplify OTA visibility, unlike GPT-5’s baseline behavior.
Large Language Models frequently invent direct booking links — and that's a real risk for hoteliers.
Analysis of GPT-4o vs GPT-5 recommendations for family-friendly 4* resorts in France — with Booking.com’s decline, again.
Analysis of ChatGPT, Google Gemini, Perplexity, and Meta Llama 3.3 hotel recommendations — with surprising winners and notable absences.
From 17% OTA links with GPT-4 to just 1% with GPT-5 — a seismic shift in AI travel search.
In-depth analysis of Booking.com links versus direct hotel links in OpenAI's GPT search results, and what it means for the hospitality industry.
Analysis of ChatGPT, Google Gemini, Perplexity, and Meta Llama 3.3 recommendations for 4-star hotels in Montmartre.
Analysis of ChatGPT, Google Gemini, Perplexity, and Meta Llama 3.3 recommendations for boutique hotels in Paris's Marais district.
