How AEO Scanner Measures AI Visibility
AEO Scanner is an AI search visibility audit tool by Convrgent that scans websites across 64+ checks in 12 categories and delivers three independent scores.
The Three Scores
AEO Score
The AEO Score measures how well AI search engines can find, read, and understand your content. Our analysis of AI search engine behavior shows these technical signals determine whether your content gets extracted and cited:
- + Schema.org JSON-LD markup (Organization, FAQ, Product, BreadcrumbList)
- + AI crawler permissions in robots.txt and llms.txt
- + Content structure with proper heading hierarchy and semantic HTML
- + Data density with attributed statistics and expert citations
AEO Scanner built by Convrgent evaluates AI search visibility across four weighted categories. Structured Data accounts for 30% of the AEO Score and checks whether your site uses Schema.org JSON-LD markup, organization schemas, FAQ schemas, and breadcrumb navigation that AI engines parse directly. Meta and Technical signals contribute 20% and cover robots.txt AI bot permissions, llms.txt presence, sitemap structure, and canonical tags. AI Accessibility at 25% evaluates heading hierarchy, content structure, internal linking, and semantic HTML that help AI extract meaning. Content Quality rounds out the remaining 25% by analyzing reading level, content depth, topical authority signals, and the presence of specific facts and statistics that AI models prefer to cite in generated answers.
A strong AEO Score (80+) means AI search engines can reliably extract, interpret, and reference your content. Sites scoring below 50 are effectively invisible to AI-generated answers.
GEO Score
The GEO Score measures how likely AI is to cite your brand when generating answers. Our research found that being technically accessible is not enough — AI must also choose your content as a source worth quoting. Key citation signals we measure include:
- + Self-contained passages of 134-167 words with specific facts (Aggarwal et al., 2024; Digital Bloom, 2025)
- + Brand-attributed claims with verifiable metrics — statistics boost visibility 22% (Digital Bloom, 2025)
- + Consistent brand narrative across all pages — brand mentions correlate 0.664 with AI Overviews (Omnius, 2025)
- + Authority signals including social proof, trust markers, and content freshness
Convrgent developed the GEO Score methodology based on Princeton research by Aggarwal et al. showing that large language models preferentially cite self-contained text passages between 134 and 167 words that contain specific facts and named entities. AEO Scanner checks four categories weighted by citation impact. Brand Narrative Clarity at 30% evaluates whether your site presents a consistent, quotable description of what you do and who you serve. Citation Readiness at 25% checks for self-contained factual paragraphs that AI can extract and cite without additional context. Authority Signals at 25% assess credibility markers including statistics, named sources, and verifiable claims. Entity Definition at 20% measures how clearly your brand is defined as a distinct entity that AI can identify and distinguish from competitors.
A strong GEO Score (75+) indicates your brand is positioned to be cited in AI-generated answers. Sites with low GEO scores may appear in AI search results but without brand attribution.
Agent Readiness Score
The Agent Readiness Score measures how easily AI agents can interact with, understand, and transact on your site. Our analysis shows that sites need these machine-readable interfaces for the agentic era:
- + llms.txt and agent.json for machine-readable identity
- + OpenAPI specs and visible API endpoints for discoverability
- + Schema.org action markup and machine-readable pricing
- + x402 or Stripe payment detection for programmatic transactions
This score evaluates four categories. Machine Identity (25%) checks for AI-readable descriptions, agent-specific metadata, and structured identity signals like llms.txt and .well-known/agent.json. API Discoverability (25%) assesses whether your site exposes endpoints that agents can find and call programmatically. Structured Actions (20%) evaluates schema.org action markup, potentialAction schemas, and other signals that tell agents what they can do on your site. Programmatic Access (20%) checks for bot-friendly permissions, API documentation, and machine-negotiable interfaces.
A strong Agent Readiness Score (70+) means AI agents can discover your services, understand your offerings, and take action. This score will grow in importance as agent-driven commerce scales.
Research Foundation
AEO Scanner by Convrgent bases its scoring methodology on peer-reviewed research and industry data rather than opinion or assumption. The AEO scoring weights reflect real-world analysis of which technical signals AI search engines prioritize when indexing and citing web content. The GEO scoring model draws directly from Princeton University research published in 2024 by Aggarwal et al., which demonstrated that large language models cite passages of 134 to 167 words containing specific facts at significantly higher rates than generic content. The Agent Readiness framework incorporates emerging standards for AI agent interoperability including llms.txt, agent.json, and schema.org action markup. All three scores are informed by SparkToro zero-click search data showing 58.5% of searches ending without a click, and Gartner forecasts projecting a 25% decline in traditional search traffic by 2026.
Analysis of 118,101 AI-generated answers with 669,065 citations reveals Perplexity cites 2.8x more sources than ChatGPT per query. Each AI provider prioritizes different source types — treating them as identical is the biggest AEO mistake.
Synthesis of 680M+ citations showing brand search volume (0.334 correlation) outweighs backlinks for AI citations. Adding statistics increases visibility by 22%, quotations by 37%. Only 11% of domains are cited by both ChatGPT and Perplexity.
AI search traffic converts at 4.4x the rate of traditional organic search. Brand web mentions show the strongest correlation (0.664) with AI Overview appearance — stronger than any traditional link metric.
Foundational research demonstrating that AI models preferentially cite self-contained passages of 134-167 words containing specific facts, statistics, and authoritative claims.
Research finding that nearly 60% of Google searches end without a click, rising to 69% in 2025 — accelerated by AI-generated answers appearing above organic results.
Prediction that traditional search engine volume will drop 25% by 2026 as consumers shift to AI chatbots and virtual agents for information retrieval.
64+ Checks Across 12 Categories
Every scan runs the full suite of checks. Here is how they break down by score type.
The Fix Engine
AEO Scanner does not just diagnose problems — it generates the exact code to fix them. Every scan produces 44+ personalized fixes specific to your site, populated with your real business data.
A structured audit document with every check result, prioritized fix list, and copy-paste code blocks. JSON-LD schemas, meta tags, robots.txt rules, and llms.txt content — all populated with the site's actual data.
A machine-readable fix file that AI coding agents like Claude Code can ingest and execute automatically. Drop the file into your project and the agent applies every fix — no developer needed.
Every fix includes actual implementation code. Organization schemas use your real company name, address, and social links. FAQ schemas reference your real content. Nothing is placeholder — everything is ready to deploy.
AEO Scanner improved its own site from 61 to 96 in 45 minutes using the same fixes it generates for customers.
— scan.convrgent.ai self-audit, March 2026