How the GeoScore is calculated
A score is only useful if you can see how it was built. This page sets out what we measure, how it is weighted, and what the number does and does not mean.
Two scores, not one
AI Visibility
What the engines actually returned. We combine your per-engine scores across ChatGPT, Gemini, Perplexity and Claude (60%) with your mention rate across the guest-style prompts we test (40%). It answers: are you named today?
AI Readiness
Whether the public evidence supports being named tomorrow. Website and structured data (35%), citations and listings (35%), social presence (30%). Readiness usually moves first; visibility follows as engines re-crawl.
What goes into readiness
Website & structured data
Whether an AI engine can read what you are, where you are, what you sell and what makes you different — structured data, page coverage, clarity of category and location, and question-level content.
Citations & listings
Consistency and completeness of the third-party records engines lean on: Google Business Profile, review platforms, booking sites, industry directories and local listings.
Social presence
Whether your public social footprint gives engines usable, recent, specific language about your experiences — coverage, recency, descriptive quality and consistency with your other records.
Each component is scored 0–100 from the public evidence available. Per-platform scores are reported separately so you can see where the weakness sits, and each result is locked for 24 hours so day-to-day model wording doesn't read as real movement.
How prompts are chosen
Prompts are built from real hospitality decision language rather than keyword lists — the questions people actually ask when deciding where to stay, eat, book a treatment or play.
Each pack mixes category intent ("boutique hotel with a spa"), location intent, occasion intent ("romantic weekend", "family Sunday lunch") and comparison intent ("X or Y"). Pro accounts can add and track their own prompts and see mention rate, rank and share of voice per prompt over time.
Score bands
Named early and described accurately across most tested prompts.
Visible, but usually behind competitors or described generically.
Occasional mentions, with clear evidence gaps behind them.
Rarely or never named when your ideal guest asks.
Limitations we are open about
- AI answers are probabilistic. The same prompt can return different wording on different days, so we lock each business's result for 24 hours and report movement with an explanation rather than presenting a single number as absolute truth.
- We measure public evidence only. We cannot see your booking data, your PMS or private channels, and we do not attempt to.
- Scores are comparative and directional. A GeoScore is designed to show where you stand against your competitive set and what to change — it is not an official rating issued by any AI provider.
- We cover ChatGPT, Gemini, Perplexity and Claude. We do not report on platforms we do not test.
- Competitor scores are calculated from the same public-evidence method and are estimates, not audited figures supplied by those businesses.
See the method applied
The sample report shows every one of these components scored on a worked example.