Most businesses assume that if they rank on Google, AI can find them. That assumption is wrong — and it's costing them citations they don't even know they're missing.
I spent a few weeks building a server-side AEO analyzer: a tool that actually fetches your page, reads your robots.txt, checks your schema, and returns a scored report. Not a quiz or a self-assessment. A real fetch, the same way an AI crawler sees your site.
After running it on [BISYRI: insert number] sites across [BISYRI: insert industries], the patterns were hard to miss.
Why I Built It
I kept running into the same client question: "Is my site ready for ChatGPT?" I had no fast, honest answer. Existing AEO tools were either quiz-based (you answer yes/no questions about your own site) or they scraped surface-level metadata. Neither shows what an AI crawler actually sees.
I wanted a tool that did what the crawler does: fetch the page, parse the HTML, read robots.txt, check llms.txt, and run a real analysis. So I built one.
The tool is live at bisyri.co/tools/aeo-check/. It checks 10 AEO factors and returns a 0-100 score with specific fix instructions for every gap it finds.
What the Tool Checks
- JSON-LD schema markup — which types are present, whether high-value types like FAQPage and Organization exist
- FAQ content — FAQPage schema or question-format headings AI can extract
- Entity clarity — Organization schema with name, contact, and sameAs links to authoritative profiles
- AI crawler access — whether 7 major bots (GPTBot, ClaudeBot, PerplexityBot, and others) are allowed in robots.txt, plus llms.txt presence
- Meta tags — title, description, canonical, OG, and Twitter tags
- Content extractability — text-to-code ratio, semantic HTML, paragraph count
- Heading structure — one H1, logical H2/H3 hierarchy
- Content freshness — datePublished or dateModified in schema or meta
- Sitemap — /sitemap.xml accessible and referenced in robots.txt
- Page weight — HTML response size as a crawlability proxy
What I Found Across Real Sites
[BISYRI: Add 3-4 specific findings from running the tool. Examples: "X% had no schema markup at all", "Y of Z sites were blocking at least one major AI crawler", "The average score was N/100". Use real numbers from your own testing.]
The finding that surprised me most: [BISYRI: insert the most striking specific result — e.g., "a site with a DR of 40+ scoring 18/100 because it was blocking GPTBot and had no schema"].
The Three Failures That Appeared on Almost Every Site
1. Missing entity schema. Most sites had no Organization block in their JSON-LD. AI engines need a structured record of who a business is before they will cite it for brand queries. No Organization schema means AI has to infer your identity from unstructured text — and often won't.
2. AI crawlers blocked. [BISYRI: insert your stat — e.g., "About a third of sites I tested were blocking at least one major AI crawler."] In many cases this was unintentional — a blanket Disallow: / added years ago for SEO reasons, now silently excluding the site from AI citation.
3. No FAQ schema. FAQPage JSON-LD is the most extractable format for AI engines. It explicitly labels questions and answers for machine reading. Almost no small business sites had it.
What the Score Distribution Looked Like
[BISYRI: Add a breakdown — e.g., "Of the X sites tested: N scored Excellent (80+), N scored Good (60-79), N scored Fair (40-59), N scored Poor (under 40). Add any notable outlier stories.]
What Good Looks Like
The highest-scoring sites had three things in common: comprehensive JSON-LD schema, all major AI crawlers allowed in robots.txt, and direct-answer content structure where the first sentence of each section states the answer clearly. Not necessarily the biggest or best-known sites — just ones built with AI extractability in mind.
Try It on Your Own Site
The tool is free, no login required, and returns results in under 10 seconds. Enter your URL at bisyri.co/tools/aeo-check/ and get a scored breakdown of what's blocking your AI citations — with specific fix instructions for each gap.