Which Schema.org errors prevent AIs from citing me?

Errors in Schema.org don't just eliminate the benefits of structured data — in some cases, they can actively reduce the confidence AI systems place in the source, because they signal carelessness or inconsistency. The most common errors aren't complex code issues: they're outdated data, missing required attributes, conflicts between the Schema and visible page content, and using incorrect types for the segment. The good news is that all are identifiable and fixable with the right tools.

Error 1: Missing required attributes

Each Schema.org type has required attributes to generate rich results on Google — and to be correctly processed by AI systems. Examples by type:

Article without image: Google requires at least one image to generate article rich results. Without it, the Schema is still read but doesn't generate visual highlighting and carries less weight.

Product without offers: a product without declared price and availability has limited usefulness for AIs responding to purchase queries.

FAQPage with empty answers: questions without answers filled in acceptedAnswer.text have no value — and may be interpreted as defective content.

LocalBusiness without address: a local business without an address declared in the Schema can't be correctly cited in queries with geographic intent ("best physical therapy clinic in Austin").

Error 2: Conflict between Schema and visible content

One of the most serious errors — and one that can lead to a Google manual action — is declaring in the Schema information different from what's visible on the page. Examples:

• Schema declares price $99.90, but the page shows $149.90 (after a promotion ended)
• Schema declares availability: InStock for out-of-stock product
• Schema declares a different number of reviews than displayed on the page
• Business hours in Schema differ from the hours shown in the footer

For generative AIs, this type of inconsistency is especially problematic: if the system detects that the Schema says one thing and the HTML says another, confidence in the source drops — because there's a signal of incorrect or outdated data.

Error 3: Inappropriate Schema type for the segment

Using a generic type when a more specific one exists reduces the precision of signals to AIs:

Wrong	Correct
Organization for restaurant	Restaurant (subtype of FoodEstablishment)
Organization for hospital	Hospital (subtype of MedicalOrganization)
Organization for lawyer	LegalService or Attorney
Article for cooking recipe	Recipe
Article for product	Product

The correct type carries additional semantics that AI systems use to categorize the source in the appropriate context.

Error 4: Schema implemented but not rendered for bots

On sites with heavy JavaScript (CSR), JSON-LD may be in the JavaScript rather than the static HTML. If the AI bot doesn't execute JavaScript, the Schema is never read — even if it's technically correct in the source code. The symptom: Google's Rich Results Test finds the Schema, but PerplexityBot or GPTBot doesn't read it.

How to check: access view-source:yourdomain.com/page and search for <script type="application/ld+json"> in the raw source code. If it doesn't appear, the Schema is being injected via JavaScript.

How to fix: ensure JSON-LD is rendered in the server HTML (SSR/SSG), not injected by client-side JavaScript.

Error 5: Duplicate or contradictory Schema

Multiple JSON-LD blocks on the same page declaring conflicting types confuse processing systems. A classic example: SEO plugins that automatically generate an Organization Schema, combined with a custom Schema declaring LocalBusiness with different information. Google tends to prioritize the more specific one, but the ambiguity reduces source reliability.

How to identify: use Rich Results Test and check if multiple blocks of the same type appear with divergent data.

Error 6: Outdated data

Outdated datePublished and dateModified, outdated prices, old phone numbers and addresses — all are negative signals that the site isn't actively maintained. For AIs that prioritize recent content, an article with dateModified from 2021 in a 2026 data query has a clear disadvantage.

Best practice: establish a periodic process (quarterly) for reviewing Schema data, especially for frequently changing information — prices, hours, phone numbers, publication dates.

How to identify Schema errors on your site

Rich Results Test (search.google.com/test/rich-results): tests a specific URL and lists errors and warnings for each Schema found.

Google Search Console > Enhancements: shows Schema errors at scale for the entire site, grouped by type.

Schema Markup Validator (validator.schema.org): validates conformance with the Schema.org vocabulary independently of Google.

FRT Digital includes a complete Schema.org audit — identifying errors, incorrect types, and outdated data — as part of the AIO Score audit. Learn about the AIO service for continuous implementation and maintenance of structured data.