RankGen's AI Visibility Score (0–100) measures how well a brand's website and content are optimized for discovery by AI models. Here is the complete scoring methodology.
The AI Visibility Score is a heuristic composite score calculated by crawling and analyzing a brand's website. It evaluates eight dimensions that research and empirical testing have shown to correlate with AI brand recommendation frequency. Each dimension is scored independently and weighted by its impact on AI discoverability.
| Dimension | Max Points | What It Measures |
|---|---|---|
| Category Clarity | 15 | How specifically the brand's category is communicated in content and metadata |
| Educational Depth | 20 | Comprehensiveness of educational content about the brand's category |
| Comparison Content | 15 | Presence of competitive comparison and differentiation content |
| Authority Tone | 10 | Quality and professionalism of language signaling authority and expertise |
| FAQ Presence | 10 | Number and quality of structured FAQ question-answer pairs |
| Brand Repetition | 10 | Frequency and naturalness of brand name usage in content |
| Structured Content | 10 | Quality and completeness of JSON-LD schema markup |
| Geographic Clarity | 10 | Explicitness of geographic scope in content and structured data |
| Total | 100 |
The scraper extracts the page title, H1, meta description, and first 500 words of body content. It checks for specific category language — the presence of the brand's defined category name in prominent positions. A title that includes the category name scores 5 points. An H1 with category language scores 4 points. Meta description inclusion scores 3 points. Body text category density contributes the remaining 3 points.
Educational depth is measured by word count, paragraph count, H2/H3 count, and content complexity. Pages below 500 words score 0. 500–1000 words scores up to 10. 1000+ words with multiple structured sections scores up to 20. The GPT-4o AI enhancement layer also scores semantic content relevance — content that directly educates about the category scores higher than content that merely mentions it.
The scraper looks for comparison indicators: "vs.", "alternative", "compare", "difference", "vs [competitor name]" patterns in headings and body text. Tables with comparison structure score an additional bonus. The presence of a dedicated comparison section (H2 level or higher) scores the maximum.
Authority language detection scans for a dictionary of authority-signaling terms ("leading", "trusted", "expert", "recognized", "professional", "experienced", "industry-leading", "established"). Each unique authority term in context scores 1–2 points up to the maximum. Vague superlatives without context score lower than specific authority claims backed by evidence.
FAQ detection identifies question-answer pairs using several heuristics: explicit FAQ sections, heading-followed-by-paragraph patterns, "?" in headings, and FAQPage JSON-LD schema. 1–3 FAQs: 3 points. 4–7 FAQs: 6 points. 8+ FAQs: 10 points. FAQPage JSON-LD presence adds a bonus.
Brand repetition counts occurrences of the brand name in body text, normalized by content length. Optimal brand name frequency is 3–5 occurrences per 1000 words. Too few (under 2/1000) scores low; optimal (3–5/1000) scores maximum; excessive (over 8/1000) is penalized as unnatural.
Structured content scoring checks for JSON-LD schema presence (Organization: 4 points, FAQPage: 3 points, other schema types: 1–2 points) and HTML structure quality (H1 presence: 1 point, H2 count: 1 point, semantic HTML: 1 point).
Geographic clarity scans for explicit geographic language in the page title, meta description, H1, Organization schema's areaServed field, and body text. At least one clear geographic reference scores 5 points; multiple geographic signals including structured data scores up to 10.
When AI enhancement is enabled (Pro and above plans), RankGen uses GPT-4o to semantically analyze the scraped content. The AI evaluates content quality beyond heuristics — identifying gaps between claimed category position and actual content depth, specific improvement recommendations, and semantic authority signals. The AI-enhanced score and heuristic score are blended for the final result.
Based on audits of thousands of brand websites, typical score distributions are: