The GEO Score measures technical readiness, not actual AI visibility. A brand can score low on GEO Score but still be highly visible in AI answers due to brand authority. For actual visibility measurement, request a GEO Audit.
The 3-Pillar Framework
Accessibility (33 points)
Can AI crawlers find and access your content?| Signal | Max Points | What It Checks |
|---|---|---|
| robots.txt AI Crawler Access | 10 | Whether the major AI crawlers (top-tier bots from OpenAI, Anthropic, Perplexity, Google, and a wider set of secondary crawlers) are allowed in your robots.txt |
| Sitemap Accessibility | 8 | Whether a valid sitemap.xml exists with lastmod dates |
| Page Crawlability | 7 | Whether pages return HTTP 200 without requiring JavaScript to render |
| llms.txt Presence | 8 | Whether a structured /llms.txt file exists describing your brand and content |
Readability (37 points)
Can AI parse and understand your content’s structure?| Signal | Max Points | What It Checks |
|---|---|---|
| Heading Hierarchy | 7 | Whether pages use one H1 and nest headings logically |
| Answer-Block Formatting | 7 | Whether FAQ sections, Q&A patterns, and question-style headings exist |
| Content Depth | 6 | Whether key pages have substantive body content |
| Internal Linking Quality | 5 | Whether pages have contextual internal links beyond nav/footer |
| Meta Description Quality | 6 | Whether meta descriptions are present and length-appropriate on every page |
| Open Graph Tags | 6 | Whether og:title, og:description, og:image, and og:type are present |
Understandability (30 points)
Can AI extract meaning, context, and authority?| Signal | Max Points | What It Checks |
|---|---|---|
| JSON-LD Structured Data | 8 | Whether Organization, Product/Article, and FAQPage schemas exist in JSON-LD |
| Author Attribution | 5 | Whether author names and Person schema are on content pages |
| Trust & Credential Indicators | 5 | Whether an About page with team info and contact information exists |
| Content Freshness | 5 | Whether content shows recent updates and last-modified dates |
| AI-Optimised Content Signals | 7 | Whether comparison pages, buying guides, and “best of” lists exist |
Scoring
The GEO Score is calculated by summing all 15 signal scores. Each signal is scored independently — the total is additive, not weighted or gated.Score Bands
| Score | Band | Meaning |
|---|---|---|
| 80–100 | Fully Optimised | Strong technical signals across all pillars |
| 60–79 | Well Optimised | Solid foundation with some gaps |
| 40–59 | Partially Optimised | Notable gaps in technical readiness |
| 0–39 | Not Optimised | Foundational work needed across multiple areas |
What the GEO Score does NOT measure
- Actual AI visibility — whether your brand is mentioned in AI-generated answers
- Brand authority — how well-known your brand is to AI models
- Content quality — whether your content is accurate, helpful, or well-written
- Competitive position — how you compare to competitors in AI citations
How scanning works
The scanner runs in three stages:- Static fetch. robots.txt, sitemap.xml, llms.txt, and the homepage are fetched directly via HTTP.
- Browser crawl. A headless browser renders up to 5 pages and extracts DOM-level signals — headings, internal links, author info, structured-data blocks, content patterns.
- Signal evaluation. Each of the 15 signal analysers runs against the collected page data, and the scores are summed.