Four signals. One score.
Every model that matters.
The GEO Score scores, from 0 to 100, how prepared your brand is to be discovered and recommended by AI — across ChatGPT, Gemini, Perplexity and Claude, tracked weekly.
What the GEO Score measures
GEO score measures eligibility, not traffic. It asks one question of every model: if a buyer in your category asked for a recommendation, is there enough clear, corroborated, extractable information about you for the model to name you?
The score evaluates four things: how clearly AI understands your brand as an entity, how often trusted sources mention you, how easily your content can be extracted and cited, and how consistently you are named across the web. Each is scored independently, then combined into a single 0–100 score.
The four signals
| Signal | Where it lives | What it measures |
|---|---|---|
| Entity Clarity | On-site | How clearly AI understands your brand as a defined entity in knowledge graphs — consistent name, description, structured data, resolvable identity. |
| Citation Density | Off-site | How many authoritative third-party sources reference your brand by name. |
| Structured Proof | On-site | How extractable your content is — self-contained passages, tables, statistics, schema markup a model can lift and quote. |
| Corroborating Mentions | Off-site | Whether you are named in editorial coverage, "best of" lists and roundups AI reads. |
Two signals are about what your site says. Two are about what everyone else says about you. A brand that scores well on one pair and badly on the other is usually seen but not selected.
Where the signals come from
GEO score is built on The Signal Hierarchy, Visigent's published methodology for AI brand visibility. The hierarchy orders the signals that move recommendation rate, from foundational (entity definition) to amplifying (community presence), and is derived from our analysis of more than 40,000 websites and 1,500 sources.
The full method — prompt sets, sampling, scoring and its limits — is in How to measure AI visibility.
How often it moves
Weekly. Each week the platform re-queries the models with your category's prompt set, re-scores the four signals and records the change.
Because every execution step maps to a signal — entity submissions to Entity Clarity, editorial outreach to Citation Density, passage rewrites to Structured Proof, list inclusion to Corroborating Mentions — you can see which change moved which number.
Does a higher GEO score mean more recommendations?
In the categories we monitor, stronger GEO score scores consistently align with higher recommendation frequency: brands with stronger authority, citation and trust signals appear more often inside AI-generated answers.
We track both GEO score and recommendation rate weekly, side by side, so the relationship is visible in your dashboard rather than asserted on this page.
GEO score is a leading indicator. Recommendation rate is the outcome. Traffic and pipeline from AI are the result.
What GEO score is not
Not a ranking.
AI does not rank; it recommends. GEO score scores your eligibility to be recommended, not a position.
Not a traffic forecast.
It predicts whether you can be named, not how many people will click.
Not a single-model score.
Each model is queried separately; the score reflects all of them, and the dashboard shows each.
Questions about the score
See where you stand
Get your GEO Score across ChatGPT, Gemini, Perplexity and Claude — where you're understood, where you're missing, and what to fix first.
Start with the score.
Enter your website. In 24 hours: your GEO Score, how each model describes you, where you're missing, and the first three fixes.