Measurement

The direct answer: AI hotel rankings churn so hard that a single week’s position is close to meaningless. Across 11 consecutive weeks (May 11 – July 20, 2026), a research pipeline fired the same 616 hotel-discovery prompts at ChatGPT, Gemini, Perplexity, Copilot and Google AI Mode every Monday — and found that, on average, 70.8% of each engine’s top 10 is replaced from one week to the next. The composite top-10 keeps only 29.2% of its members week-over-week; the top-100 keeps 49.7%. The median hotel that stays on a 500-deep list still moves 55–80 places between Mondays depending on the engine (Nicolas Sitter, AI Hotel Volatility 2026, July 22, 2026). If your marketing report says “we entered ChatGPT’s top 10” based on one week of checks, it will usually un-happen by itself. The honest way to measure AI visibility is a two-layer read: a stable pool membership (are the engines reaching for you at all?) tracked weekly, and a volatile position judged only on a smoothed 4–8 week rolling window. ...

The short answer Most AI-influenced direct bookings are already happening on your books — your analytics just can’t see them. When a traveler asks ChatGPT or Gemini for a hotel in your city and your property gets named, the majority never click a link at all: they copy your hotel’s name into Google, type it into their browser, or open your app directly. A 62-day single-hotel study that surveyed arriving visitors and cross-checked their raw referrer data found that 68.6% of guests who said an AI assistant sent them arrived carrying no trace of it — no referrer, no campaign tag, nothing any attribution model could use (Nicolas Sitter, Hotel Ranque study, September 2, 2026). ...