What We Measure — 147 Factors, 12 Categories, One AI Visibility Score

Before AI can cite your content, it needs to access it. Most AI platforms — ChatGPT, Perplexity, Google AI Overviews — use web crawlers to discover and index content. If your robots.txt blocks these bots, if your content requires JavaScript to render, or if you haven't signalled AI-readiness through standards like llms.txt, AI platforms will never see your content. This category checks whether the front door is open.

AI platforms need to know who you are before they can recommend you. Entity identity measures whether your brand exists in the knowledge systems AI relies on — Google's Knowledge Graph, Wikidata, Wikipedia. Research shows brand search volume has the strongest correlation (0.334) with AI mentions — stronger than backlinks or domain authority.

AI platforms don't just read your content — they need to extract specific passages to cite in their responses. Research on Google AI Overviews shows the ideal citable passage is 134-167 words, self-contained (understandable without context), and directly answers a question. Pages with extractable summary blocks get cited more.

AI systems parse your content structurally — heading hierarchy, paragraph length, section organisation. Well-structured content with clear heading hierarchy, readable sentence lengths, and logical flow helps AI understand your topic and extract relevant information. Poor structure forces AI to guess at relationships between ideas.

When someone asks ChatGPT a question, the platform decomposes it into 3-8 sub-queries and looks for sources that answer multiple sub-queries. Pages that cover the complete semantic space of a topic get selected more. Google AI Overviews sources with 15+ recognised entities per 1,000 words have 4.8x higher selection rates.

AI platforms cross-reference claims before citing them. Content that cites authoritative institutions (.gov, .edu, peer-reviewed journals) is more trusted. Original data, statistical claims, and properly attributed research increase citation probability by 22%. AI systems check whether what you say is supported.

Schema markup (JSON-LD structured data) is a direct communication channel between your website and AI systems. It explicitly tells AI: this is our organisation name, this is the article's author, these are our FAQ answers. Pages with complete schema markup have significantly higher AI visibility because there's no guesswork.

AI systems are increasingly multimodal — they process text, images, and video together. But they still rely heavily on text descriptions to understand media. Images without alt text are invisible to AI. Videos without transcripts are skipped. Text-based descriptions allow AI to comprehend your full content.

AI platforms strongly prefer fresh content. Perplexity's content decay rate is 2-3 days for time-sensitive topics. Google AI Overviews prioritise recently updated sources. Stale content with outdated statistics or no visible update dates gets deprioritised. Multiple date signals need to be consistent.

Google's E-E-A-T framework (Experience, Expertise, Authoritativeness, Trustworthiness) directly influences AI citation decisions. Research shows 96% of Google AI Overview citations come from sources with strong E-E-A-T signals. This means identifiable authors with credentials and clear organisational identity.

This is the ultimate measure. Everything else contributes to one outcome: does AI actually mention and cite your brand? This category uses live testing — we send real queries to ChatGPT and Perplexity and analyse the responses. We check if your brand is mentioned, how prominently, and whether your URL is cited.

While AI crawlers care less about performance than human visitors, basic technical health still matters. Extremely slow sites may time out before AI crawlers finish. Poor search visibility means AI training data may not include your content. This category carries the lowest weight but shouldn't be ignored.