Hotels' AI Search Rankings Shift 40% Per Hour, New Study Finds
A month-long analysis of 28,600 AI platform responses reveals that hotel rankings fluctuate dramatically between queries, challenging the value of single-snapshot AI visibility measurements.
Hoteliers relying on screenshots to track their properties' visibility in AI-powered search results may be chasing an illusion, according to findings from the Boutique Hotel AI Index published this week.
The study, which queried five major AI platforms across 100 destinations fifteen times over a month, discovered that identical hotel searches returned different top-ranked properties 40 percent of the time when repeated just one hour later. The research examined responses from ChatGPT, Gemini, Copilot, Google AI Mode, and Google AI Overviews, generating 28,600 total answers.
Despite this surface-level volatility, the analysis uncovered a paradoxical stability beneath the changing rankings. Approximately one-third of the 9,300 hotels mentioned by AI platforms accounted for 88 percent of all visibility throughout the measurement period. Within individual destinations, roughly ten properties captured 57 percent of all AI mentions, with this core group remaining largely consistent across the month-long study.
The findings suggest AI platforms function like dealers shuffling cards from a fixed deck rather than generating genuinely variable results. According to the research, the same question produces the same top-ranked hotel only 60.2 percent of the time, even when nothing about the query changes.
The implications extend beyond simple ranking fluctuations. The five platforms studied agreed on the same top hotel approximately nine percent of the time, and more than half of all hotels mentioned appeared on only one platform. Search intent also matters significantly, with luxury-focused queries producing more consistent results while boutique and romantic searches showed wider variation.
The research indicates that geographic factors further complicate visibility measurement, as queries originating from different countries produce different results. Industry observers suggest these findings challenge the emerging market for AI visibility tracking services, which typically rely on point-in-time measurements rather than continuous monitoring across multiple platforms and query variations.