How Missing Structured Data Made Us Invisible to ChatGPT
Spotting the Problem: LLMs Didn't Know We Existed
Last month during a routine check, we noticed something weird: when we asked ChatGPT about "AI compatibility scanners" or "website optimization for AI," it never mentioned AICompatible.com. Same story with Perplexity. Yet our homepage was properly indexed in Google Search Console, we had organic traffic, and we showed up fine in regular Google searches.
At first, we figured "maybe we're too new—we're not in the LLM training data yet." But then we looked at competitors and saw that some platforms launched *after* us were being correctly identified by ChatGPT. That pushed us to dig deeper.
The Diagnosis: We Had No Structured Data
We started investigating systematically. First step: we looked at our own source code and were shocked to find that our homepage had zero JSON-LD structured data. Just basic HTML meta tags.
When we ran Google's Rich Results Test, the verdict was clear: "No structured data found on this page." Meanwhile, every successful competitor we checked had Organization, WebPage, FAQPage schemas—the works.
Then it clicked: traditional search engines like Google can work with HTML content and meta tags, but LLMs use structured data far more effectively. JSON-LD in Schema.org format acts like a machine-readable "identity card" that makes it much easier for AI to understand your site.
The Technical Fix: We Added Four Core Schemas
Once we identified the problem, we decided to add four critical schema types to our homepage. Here's exactly how we did it:
1. Organization Schema: Our Corporate Identity
First, we added the Organization schema that defines who we are. This is the most critical structure for LLMs to recognize your brand:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Organization",
"name": "AICompatible.com",
"url": "https://aicompatible.com",
"logo": "https://aicompatible.com/logo.png",
"description": "Automated scanning and optimization platform for AI-compatible websites. Optimize your site for ChatGPT, Perplexity, and other LLMs.",
"sameAs": [
"https://twitter.com/aicompatible",
"https://linkedin.com/company/aicompatible"
],
"contactPoint": {
"@type": "ContactPoint",
"contactType": "customer support",
"email": "hello@aicompatible.com"
}
}
</script>Critical fields:
name: Your brand name—write it consistentlydescription: Clearly explain what you do with keyword focus (150-200 characters is ideal)sameAs: Your social media profiles—important for LLMs to verify youurl: Your canonical domain
2. WebPage Schema: Page Context
We added WebPage schema to specify what our homepage is about:
{
"@context": "https://schema.org",
"@type": "WebPage",
"name": "AI-Compatible Site Scanner - AICompatible.com",
"description": "Analyze how your site appears to ChatGPT and other AI systems. Get your LLM compatibility score with a free scan.",
"url": "https://aicompatible.com",
"mainEntity": {
"@type": "SoftwareApplication",
"name": "AICompatible Scanner",
"applicationCategory": "SEO Tool",
"offers": {
"@type": "Offer",
"price": "0",
"priceCurrency": "USD"
}
}
}
The mainEntity field is particularly powerful: it defines your page's primary subject (in our case, a software product) in structured form.
3. FAQPage Schema: Frequently Asked Questions
We had an FAQ section on our homepage, but it wasn't structured. Adding FAQPage schema made this content far more accessible to LLMs:
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "What is an AI-compatible website?",
"acceptedAnswer": {
"@type": "Answer",
"text": "An AI-compatible website has structured, clean content that can be easily understood by large language models like ChatGPT."
}
},
{
"@type": "Question",
"name": "How long does a scan take?",
"acceptedAnswer": {
"@type": "Answer",
"text": "For an average website, scanning takes 30-60 seconds. Results are displayed instantly."
}
}
]
}
4. BreadcrumbList Schema: Site Hierarchy
We added BreadcrumbList to show our site structure. This helps LLMs understand your site's organization:
{
"@context": "https://schema.org",
"@type": "BreadcrumbList",
"itemListElement": [
{
"@type": "ListItem",
"position": 1,
"name": "Home",
"item": "https://aicompatible.com"
},
{
"@type": "ListItem",
"position": 2,
"name": "Scanner",
"item": "https://aicompatible.com/scanner"
}
]
}
Results and Key Takeaways
Ten days after adding the schemas, we asked ChatGPT the same questions again. This time it mentioned AICompatible.com and accurately described our service. We saw similar results in Perplexity.
This problem is widespread: Our research shows that over 70% of websites lack any JSON-LD schema. SaaS and B2B companies are particularly behind on this.
Check your own site:
- Search your source code for
<script type="application/ld+json"> - Use Google's Rich Results Test:
search.google.com/test/rich-results - Validate with Schema.org validator:
validator.schema.org
AICompatible.com's scanner automatically checks for the presence and quality of JSON-LD schemas when analyzing your site. It reports whether core schemas like Organization, WebPage, and Product are missing, and whether critical fields (name, description, sameAs) are populated in existing schemas. With a free scan, you can learn your site's LLM compatibility score and see exactly which schemas are missing.
Structured data is no longer just about Google rich snippets—it's become a critical requirement for visibility in the AI era.