How this score is calculated
Every destination's score is an unweighted average of four metrics, each measuring a different aspect of how AI models talk about that destination. A destination missing data for one metric has that metric excluded from its average — it's never treated as a zero.
Model agreement
How often ChatGPT, Claude, and Gemini agree on the #1 recommended pick for the same question. Higher agreement means AI models converge on the same answer for a destination; lower agreement means the models diverge.
Confirmed-open rate
Of the places AI models recommend as a top pick, the share that our follow-up checks have confirmed are still open and operating — a measure of how current and reliable the AI-generated recommendations are.
Distinctiveness
How much a destination's most-mentioned themes differ from those of other tracked destinations in the same country. A destination that AI describes the same way as all its peers scores low; one AI associates with a unique set of themes scores high.
Presence rate
How often a destination comes up in regional and national AI questions that aren't specifically about it. A recommendation counts as full credit; a mention (without being recommended) counts as half credit. Destinations without any regional or national question data don't have a presence rate — the metric is excluded from their average rather than counted as zero.
Putting it together
The composite score is the plain average of whichever of these four metrics have enough data to compute — no metric is weighted more heavily than another.