How Scores Are Calculated
- 1
Each dimension produces a raw score (0–100)
Every audit dimension is evaluated independently and scored on a 0 to 100 scale, where 100 means the dimension is fully optimized for AI citability.
- 2
Raw scores are weighted and summed
Each dimension has a defined weight (percentage). The weighted sum of all 22 dimension scores produces your overall AI Visibility Score, also on a 0–100 scale.
- 3
The overall score maps to a letter grade
A = 90+, B = 80–89, C = 70–79, D = 60–69, F = below 60. The grade gives you an at-a-glance sense of your AI citability relative to the industry benchmark.
Why these weights? The dimension weights reflect empirical correlation with AI citation behavior — primarily derived from analyzing patterns across sites that appear in Perplexity, ChatGPT, Gemini, and DeepSeek responses. AI Citation Score (25%) and Answer Readiness (20%) together account for nearly half the score because passage-level retrieval and direct-answer structure are the most consistent differentiators between cited and non-cited content.
The 22 Dimensions
Ranked by weight — the dimensions at the top have the biggest impact on your overall score.
AI Citation Score
Measures how likely AI systems are to retrieve and cite your content in response to user queries. Factors include passage-level retrieval signals, quote uniqueness across the web, and authority signals such as inbound citation patterns from other widely-cited sources.
Good: Content appears verbatim or near-verbatim in AI responses and is referenced as a primary source for factual claims.
Answer Readiness / LLMO
Measures how well your content is structured for direct extraction as an AI answer. Evaluates entity precision (named entities are clearly identified), synthesis readiness (content breaks down complex topics into digestible pieces), and contextual density (sufficient background is provided for the LLM to ground its response).
Good: Pages answer a clearly scoped question in 2–4 sentences at the top, with supporting detail below — a pattern AI systems consistently pull from.
Brand Signals / AI Trust
Measures how strongly AI systems associate your brand with relevant topic areas. Analyzes unlinked brand mentions across the web, the quality of citation contexts in which your brand appears, and consistency of entity naming across your site and external references.
Good: Your brand is mentioned authoritatively by third-party sources in contexts directly related to your core topics — with or without a hyperlink.
Entity Clarity / Knowledge Graph
Measures how clearly your site communicates who and what it describes to knowledge systems. Checks Schema.org markup depth and correctness, presence in Wikipedia or Wikidata, and E-E-A-T signals (Experience, Expertise, Authoritativeness, Trustworthiness) as inferred from content and structural signals.
Good: Your organization and key people are represented in Wikipedia/Wikidata, Schema.org markup is comprehensive and error-free, and author pages demonstrate clear expertise.
Content Depth
Measures how thoroughly a topic is covered relative to what search and AI systems expect for a given query intent. Evaluates topic coverage breadth, alignment with the questions users actually ask (Q&A alignment), and paragraph depth — whether sections go beyond surface-level definitions to provide genuine insight.
Good: Content covers a topic end-to-end with original analysis, data, or perspective — not just what competitors already state.
Platform Readiness / AI Access
Measures how easily AI crawlers and platforms can access and process your content. Audits the presence and quality of an llms.txt file, robots.txt directives that may block AI crawlers, and headless render fidelity — whether the content seen by JavaScript-disabled crawlers matches the rendered page.
Good: An llms.txt exists and points crawlers to the right pages; robots.txt does not block GPTBot or CCBot; content renders identically with and without JavaScript.
Technical Foundation / Core Infrastructure
Measures the health of your site's core technical stack as it affects AI indexing. Evaluates Core Web Vitals (LCP, INP, CLS), HTTPS implementation, and the validity of structured data across all pages.
Good: All Core Web Vitals in the "Good" range, every page served over HTTPS, and zero Schema.org validation errors in structured data.
Source Authority / Outbound Citations
Measures the quality and intentionality of outbound citations from your pages. Looks at whether you link to authoritative external sources, cite data sources with links, and include expert quotes or references that signal your content is grounded in real-world knowledge.
Good: Pages cite peer-reviewed sources, link to authoritative organizations, and include at least one outbound reference that signals depth rather than just linking for SEO.
Media Readiness / Accessibility
Measures how accessible and descriptive your media assets are to AI systems. Evaluates image alt text coverage, ARIA labels on interactive elements, and the ratio of descriptive media (images with meaningful captions or alt text) to decorative media.
Good: Every image that conveys information has descriptive alt text; all interactive elements have ARIA labels; charts and infographics have text fallbacks.
Readability
Measures how easily both human readers and AI systems can parse and understand your content. Evaluates Flesch-Kincaid reading ease, passive voice ratio (high passive voice reduces clarity), and sentence rhythm — whether sentences vary in length to maintain engagement without sacrificing comprehension.
Good: Content scored "Good" on Flesch-Kincaid (60–80 for general audiences), passive voice below 15% of sentences, and natural variation in sentence length.
Image Optimization
Measures how well images are optimized for both user experience and AI interpretation. Checks for descriptive alt text, appropriate sizing and format (WebP preferred), and cumulative layout shift (CLS) contribution — whether images cause unexpected layout shifts that affect Core Web Vitals.
Good: All informational images have descriptive alt text, images are served in next-gen formats, and no image causes more than a negligible CLS contribution.
Hreflang
Measures the correctness and completeness of your hreflang configuration for multilingual sites. Checks that language codes are valid (e.g., en-US, fr-FR), self-referencing hreflang tags exist on every page, and x-default is set appropriately for pages serving multiple regions.
Good: Every localized page has valid hreflang pointing to itself and all alternate language/region versions; x-default is set on the generic version.
Local SEO
Measures how well your site is optimized for local search and geo-specific queries. Checks NAP (Name, Address, Phone) consistency across your site, LocalBusiness schema completeness, and Google Business Profile signals where applicable.
Good: NAP is identical across your website and all external directories; LocalBusiness schema is complete and matches your Google Business Profile exactly.
Video Schema
Measures whether video content on your site is properly structured for AI and search discovery. Checks for VideoObject schema with required fields (name, description, thumbnailUrl, uploadDate, duration), Clip schema for video segments, and SeekToAction markup for jump-to timestamps.
Good: All videos have complete VideoObject schema; key moments are marked with Clip schema; SeekToAction is present for instructional or reference videos longer than 5 minutes.
SEO Quality
Measures traditional on-page SEO signals that still correlate with AI visibility, since AI systems are trained on content that ranks well. Checks title tag relevance and uniqueness, meta description quality, canonical tag correctness, internal linking structure, and the absence of indexable duplicate content.
Good: Each page has a unique, descriptive title; meta descriptions mirror the page's value proposition; canonical tags are self-referencing or correctly cross-referenced; no significant duplicate content issues.
Keyword Opportunity
Measures the gap between the queries your content targets and the queries where AI systems currently cite competitors but not you. Identifies high-volume, high-citation-potential keywords where your content could be the preferred source if optimized for citability.
Good: Your site appears in AI citation opportunities for your core keywords — not just in traditional search rankings.
Competitor Gap
Measures how your content stacks up against competitor content for the same target queries. Identifies dimensions where competitors score higher on AI citability factors — giving you a prioritized list of where to out-optimize rather than what to generically improve.
Good: You score higher than direct competitors on at least 3 of the top 5 weighted dimensions for your primary keyword set.
Content Quality
Measures holistic content quality signals including originality, factual accuracy indicators, structure (headings, lists, tables), and whether content provides value beyond what is already widely published. Also evaluates thin content (pages with fewer than 150 words of unique text).
Good: No thin content pages; every page has a clear purpose and delivers information not found verbatim elsewhere on the web.
Blog Passage Readiness
Measures how well individual passages within blog posts can be retrieved as standalone answers. Evaluates whether key claims are stated clearly in self-contained paragraphs, whether Q&A-style headers match how users actually phrase queries, and whether passages can stand alone without surrounding context.
Good: Each blog post contains at least 3 passages that answer a specific question directly and completely — even if read in isolation.
Internal Linking & Topic Architecture
Measures how well your site's internal link structure communicates topic hierarchy and related content to AI systems. Evaluates anchor text descriptive quality, topic cluster coverage, and whether pillar pages are clearly linked from supporting content.
Good: Every page is reachable within 3 clicks from the homepage; topic clusters are clearly defined with a recognized pillar page; anchor text is descriptive rather than generic ("learn more" avoided in favor of specific topic names).
Schema Markup Breadth
Measures the extent and correctness of structured data beyond what individual dimension audits cover. Checks for coverage of Article, BreadcrumbList, FAQPage, HowTo, and Review schemas where appropriate — and validates that all markup is error-free using Google's Rich Results Test format rules.
Good: All pages that describe articles, FAQs, how-to processes, or reviews use the corresponding schema type; zero validation errors in structured data.
Page Speed & Resource Efficiency
Measures resource load efficiency as it affects AI crawler budget and user experience. Evaluates total page weight, render-blocking resource count, image compression quality, and CDN usage. AI crawlers may abandon crawling pages that are excessively slow or resource-heavy.
Good: Pages load in under 2 seconds on a simulated mobile connection; fewer than 5 render-blocking resources; images served from a CDN in WebP or AVIF format.
What Each Grade Means
Your overall AI Visibility Score and what to expect — and prioritize — at each level.
Excellent
Score 90–100
Your site is well-structured for AI citation. Content is retrievable, authoritative, and platform-accessible. Minor improvements in lower-weighted dimensions can push you toward the top tier.
Good
Score 80–89
Your site has solid fundamentals with clear room to grow. AI citability is competitive but not yet dominant in your category. Focus on the top 3 weighted dimensions to move into A territory.
Average
Score 70–79
Your site has identifiable gaps that are costing you AI citations. Issues in AI Citation Score and LLMO readiness are likely the biggest drags. Targeted fixes in the top 5 dimensions will move the needle most.
Below Average
Score 60–69
AI systems are not yet treating your content as a primary source for your topic area. Significant foundational work is needed. Start with platform accessibility and content depth before pursuing advanced citation strategies.
Poor
Score Below 60
Your site faces structural barriers that prevent AI systems from accessing, trusting, or citing your content. A systematic approach addressing the top 9 dimensions will be required before meaningful AI citability improvements are possible.
How We Validate Our Scoring
An honest note on methodology
GeoXylia's scores are derived from structural analysis and content pattern recognition. We do not have access to proprietary AI model ranking data, citation logs, or any internal signals from Perplexity, OpenAI, Google, or any other AI platform.
What we do have is a set of signals that correlate, across thousands of audited sites, with whether a site actually appears in AI-generated responses. These signals are grounded in how retrieval-augmented generation (RAG) systems work: what they look for, what they ignore, and what patterns in source content consistently produce higher citation rates.
We continuously update our dimension weights and scoring rules as AI platforms evolve — particularly as new standards like llms.txt gain adoption and as AI companies publish more about how their retrieval systems evaluate content. GeoXylia is not affiliated with or endorsed by any AI platform.
Ready to see your score?
Run a free audit on your site and get your full 22-dimension breakdown in minutes.