How We Measure AI Visibility
The Cited Methodology is a six-dimension framework for assessing how likely a website is to be cited by AI search engines — ChatGPT, Perplexity, Google AI Mode, Claude, and Copilot. It produces a single Citation Readiness Score that tells you where you stand and exactly what to fix.
Get your Citation Readiness Score freeResults in 3–5 minutes · No account required · 100% free
AI search doesn't rank websites. It cites them.
For twenty-five years, search visibility meant a position in a ranked list. Getting to position one required competing on signals that search engine algorithms had developed to approximate relevance: backlinks, keyword presence, page speed, and dozens of technical factors.
AI search engines operate differently. When someone asks ChatGPT or Perplexity a question, the system does not return a ranked list — it generates an answer, and within that answer it names a handful of sources. The decision of which sources to name is made before the answer is written, not after. It depends on whether the AI system has already determined your site to be a trustworthy, legible, citable authority on the topic being asked about.
Research published by Princeton University and Columbia University confirms that AI citation behaviour follows patterns distinct from traditional search ranking, with structural and authority signals playing a disproportionate role relative to keyword-match signals. The GEO study (2023) identified content structure, sourcing transparency, and domain authority as primary drivers of AI citation. Pew Research reports that a majority of adults under 30 now use AI assistants for information-seeking tasks previously served by search engines — making AI citation a first-order commercial concern, not a secondary channel.
The Cited Methodology was designed to measure this. Where traditional SEO tools optimise for ranking signals, Cited measures citation signals — the factors that determine whether an AI assistant names your website when your customers ask it a question you should be answering.
What makes up the Citation Readiness Score
Every website is assessed across six dimensions. Each contributes a weighted proportion of the overall score, reflecting its relative importance to AI citation likelihood.
Weights reflect the relative impact each dimension has on AI citation likelihood, based on observed citation patterns across major AI platforms.
AI Citability
Whether the content structure signals a citable, authoritative source.
AI assistants select sources that surface clear, quotable claims with an identifiable author or organisation. Content that lacks this structure is bypassed, however accurate it may be.
Brand Authority
Whether external signals of expertise and trust back the site.
AI retrieval systems incorporate signals that indicate a site is recognised as a legitimate source in its field. These signals extend beyond traditional link-based metrics to include entity recognition, coverage in authoritative publications, and consistency of the brand's claimed expertise.
Content Quality
Whether depth, accuracy, and topical completeness match what AI systems reward.
Large language models are trained on vast corpora and have developed implicit standards for what constitutes useful, reliable content. Surface-level content that covers a topic in name only is consistently passed over in favour of pages that demonstrate genuine expertise and answer questions completely.
Technical Health
Whether the site's infrastructure allows AI crawlers to discover and process it.
A site that cannot be reliably crawled and indexed by AI systems cannot be cited by them, regardless of content quality. This dimension covers the technical foundations that enable AI platforms to access, interpret, and update their understanding of a site.
Schema Markup
Whether structured data makes entities, claims, and context machine-readable.
Schema.org markup provides AI systems with an unambiguous interpretation of a page's content — its author, subject, date, and relationship to other entities. Sites that speak the language of structured data are materially easier for AI systems to reason about and cite with confidence.
Platform Performance
Whether the site's citation profile is consistent across major AI platforms.
Different AI platforms weight signals differently and draw on different underlying data sources. A site may be well-cited by one platform and invisible to another. This dimension tracks cross-platform consistency, because visibility on a single platform is a fragile position.
From URL to Citation Readiness Score in three stages
Discovery
The assessment begins by examining the website as AI systems see it — not as a browser user sees it. This means evaluating the site across all six dimensions from the perspective of AI crawlers, retrieval systems, and language model inference pipelines. The discovery stage establishes a baseline state for each dimension before any scoring is applied.
Scoring
Each dimension is scored on a 0–100 scale, then combined into a single Citation Readiness Score using the weighted model described above. The composite score represents the site's overall readiness to be cited by AI assistants. Dimension-level scores identify exactly where performance is strongest and where the largest gaps exist.
Prioritisation
Recommendations are ranked by estimated impact on citation likelihood — not by technical difficulty or implementation effort. The fastest path to improved AI visibility is not always the most technically interesting change. Cited's prioritisation model is designed to surface the changes that will move the Citation Readiness Score the most, in the least time.
The assessment is automated. The judgement is human.
Every Cited audit is reviewed by a certified AI technician before the client sees it. The automated assessment measures the objective signals — structured data, crawl accessibility, content depth, authority indicators. The human review layer applies the contextual judgement that no scoring model can replicate.
That includes understanding the client's competitive landscape: which topics their sector is actively discussed in AI responses, which competitors are currently cited most frequently, and whether the site's brand is being correctly understood or misclassified by AI systems. It also includes identifying anomalies in the automated scores — cases where a technically high-scoring site is underperforming in practice, or a lower-scoring site is nonetheless appearing in AI answers due to domain-specific authority signals.
This combination of systematic measurement and experienced human review is what distinguishes a Cited audit from a data dashboard. The score is the starting point. The expert's interpretation is what makes it actionable.
20+ years
of search and digital strategy experience
5 platforms
audited: ChatGPT, Perplexity, Gemini, Claude, Copilot
100% reviewed
every audit reviewed before delivery
Built on published research, refined in practice
The Cited Methodology is grounded in published academic and industry research on how AI language models select and cite sources, how retrieval-augmented generation systems weight trust signals, and how large-scale AI citation behaviour is changing consumer information habits.
Aggarwal et al. — Princeton University & Columbia University (2023)
Identified nine content modifications that measurably increase citation likelihood in AI-generated responses, establishing that AI citations are predictably influenced by content structure.
Pew Research Center (2024)
Found that over half of adults under 30 regularly use AI assistants for research and information-seeking, confirming the commercial urgency of AI visibility as a distinct channel from traditional search.
SE Ranking Industry Study (2024)
Documented that AI Overviews and AI Mode cite sources based on domain authority and content freshness signals distinct from traditional SERP ranking factors — reinforcing the need for a dedicated measurement framework.
Frequently asked questions
How is AI visibility different from SEO?
Traditional SEO competes for positions in a ranked list of ten blue links. AI visibility competes for selection as one of two or three named sources in a generated answer. The signals that influence position in a keyword-ranked list — backlink volume, keyword density, click-through rate — are different from the signals that determine whether a large language model treats a site as a trustworthy, citable source. Cited measures the latter.
What is a Citation Readiness Score?
The Citation Readiness Score is Cited's proprietary composite metric. It expresses a website's overall AI visibility as a single number from 0 to 100, computed as a weighted average across six dimensions. A score above 70 indicates a site is well-positioned to be cited by AI assistants. A score below 40 suggests significant barriers to AI citation exist.
How often do AI platforms change what they cite?
The major AI platforms — ChatGPT, Perplexity, Google AI Mode, Claude, and Copilot — update their retrieval and citation behaviour continuously, both through model updates and changes to their underlying search infrastructure. Cited's methodology is reviewed and updated to reflect these changes, ensuring that the dimensions we measure remain predictive of real-world citation likelihood.
Can I improve my AI visibility without a developer?
Many of the highest-impact improvements to AI visibility are content and structural changes that require no code: rewriting key pages to lead with clear, quotable claims; adding an FAQ that answers questions AI assistants are commonly asked; structuring content so that the site's expertise on a topic is unambiguous. Technical changes — schema markup, crawl configuration, llms.txt — deliver a further lift, and Cited's managed service handles both layers end to end.
Ready for your
AI score?
Get an AI visibility audit in under 5 minutes.
No account needed, just enter your URL.
Results in minutes · Sent to your email · 100% free