11 Schema Markup Mistakes That Invalidate Your AI Citation Signals

Why Schema mistakes reduce citation rates

AI retrieval systems treat Schema markup as a trust signal — but only valid, correctly-structured Schema. Invalid Schema is either silently ignored or, in some cases, can create conflicting signals that reduce a page's citation probability relative to pages with no Schema at all.

Our analysis of 500 cited pages found that 94% had valid Schema — but we also found that among pages with Schema, roughly 34% had at least one structural error that reduced its effectiveness. These are the most common errors.

Mistake 1: Using a plain string for acceptedAnswer

This is the most common FAQPage Schema error. The acceptedAnswer property must be an Answer type object — not a plain string. Using a string won't cause a validation error in all tools, but it will often cause the Schema to be silently ignored by AI systems.

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is project management?",
      "acceptedAnswer": "Project management is the process..."
    }
  ]
}

{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is project management?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Project management is the process of leading work to achieve goals and meet success criteria within a specified time frame."
      }
    }
  ]
}

Mistake 2: Schema that doesn't match visible content

Schema markup must reflect content that's actually visible on the page. Adding FAQPage Schema for questions that only exist in the JSON-LD (not in the HTML) violates Google's spam policies and can result in penalties. More immediately, it reduces AI citation rates because AI models cross-reference Schema against visible content.

Mistake 3: Missing required properties on Article Schema

Article Schema has several properties that, while not strictly required by schema.org's definition, are practically required for AI citation effectiveness:

Mistake 4: Author as a flat string instead of Person type

Writing "author": "Jane Smith" instead of the proper Person type object with a sameAs array loses significant E-E-A-T signal. The flat string tells AI systems the author's name. The Person type with sameAs links tells them who the author is — their verifiable identity across the web.

Mistake 5: No dateModified on Article Schema

Our research found that 73% of cited pages had a dateModified within the past 12 months, and that the presence of dateModified in Schema correlated with citations independently of actual content freshness. Many sites set datePublished and never update dateModified — this is a missed signal that's trivially easy to fix.

As part of any content update workflow — even minor edits — update the dateModified field in your Article Schema.

Mistake 6: Generic, underpopulated Organization Schema

Many sites have Organization Schema with just a name and logo. This minimal implementation provides little value. Your Organization Schema should be a rich entity declaration that helps AI systems unambiguously identify your brand:

• ▸Include sameAs links to LinkedIn, Crunchbase, Wikipedia (if applicable), G2, Capterra, and major social profiles
• ▸Include description with a clear, keyword-rich description of what your organization does
• ▸Include foundingDate and numberOfEmployees for entity richness
• ▸Use knowsAbout to declare your topical expertise domains

Mistake 7: HowTo Schema without step text

HowTo Schema requires each step to have a text property with the actual instruction. Many implementations include only name (the step title) but omit text (the instruction content). This produces technically valid Schema that performs poorly — AI systems use the step text to extract procedural content, not just the step names.

Mistake 8: Duplicate Schema types on the same page

Adding two separate Article Schema blocks, or two FAQPage blocks, creates conflicting signals. Parsers may use either one, potentially choosing the less complete version. Consolidate all Schema of the same type into a single block.

Mistake 9: Placing Schema in the body instead of the head

While schema.org technically allows JSON-LD in both head and body, placing it in the <body> can cause rendering-dependent indexing issues. Crawlers that don't execute JavaScript may miss body-injected Schema. Always place your <script type="application/ld+json"> tags in <head>.

Mistake 10: No sameAs entity linking

The sameAs property is how you tell AI knowledge graphs that your brand entity is the same entity described elsewhere on the web. Without sameAs links, AI systems have no way to connect your Schema entity to external knowledge sources. Add sameAs on both Organization Schema and Person (author) Schema.

Mistake 11: Using Product Schema on non-product pages

Adding Product Schema to blog posts, documentation, or informational pages in an attempt to boost visibility is a spam signal. AI systems and Google's quality systems penalize pages that apply Schema types that don't match the actual content type. Use Article Schema for editorial content, FAQPage for Q&A content, and Product Schema only for actual product pages.

How to validate your Schema