SCHEMA MARKUP FOR AEO: THE STRUCTURED DATA FOUNDATİON FOR MACHİNES AND ANSWER ENGİNES

When search engines and AI assistants read a page, they are really interpreting raw text. Schema markup, or structured data, is a standard layer that describes your content to these systems in a machine-readable language. In the AEO (Answer Engine Optimization) era, well-built structured data clarifies a piece of content's author, date, and entity relationships. Yet the field is full of myths: some sources present schema as a magic ranking button, while others claim it guarantees AI visibility. Drawing on verified Google sources and controlled studies, we will separate fact from hype one by one.

What Schema Markup Is and Why It Is the Foundation of Structured Data

Schema markup is standardized structured data that describes a page's content to search engines in a machine-readable way using the shared vocabulary called schema.org. Google uses it for two purposes: to understand more clearly what the page is about and to make the page eligible for search features such as rich results. The schema.org vocabulary is jointly supported by Google, Microsoft (Bing), Yandex, and Yahoo!, making it a single universal standard. The vocabulary contains hundreds of types and over a thousand properties and is used by millions of domains. The correct framing is this: structured data is not the strategy itself but the packaging layer that lets machines parse good content correctly.

JSON-LD, Microdata, and RDFa: The Three Formats and Why Google Recommends JSON-LD

There are three valid ways to embed schema.org data in a page: JSON-LD, Microdata, and RDFa. Google's official wording is clear: all three formats are equally fine for Google as long as the markup is valid and properly implemented. Google recommends JSON-LD, but only because it is the easiest to implement and maintain at scale and the least prone to user error. Contrary to a common misconception, JSON-LD as a format does not earn higher rankings or more rich results. JSON-LD is placed inside a script tag and can appear in either the head or the body of the page. What matters is that the markup is present in the page's raw HTML; external .json files are not crawled. For details, Google Search Central documentation is the primary source.

The Most Important Schema Types for Content Sites

For a blog, the most appropriate type is BlogPosting, for news content NewsArticle, and for general articles Article. None of them has required properties, but Google recommends author (Person or Organization), datePublished, dateModified, headline, and image. Organization markup defines the brand identity; name, url, logo, and sameAs (links to social and authoritative profiles) are the most valuable fields, and adding it to the homepage alone is sufficient. BreadcrumbList describes the site hierarchy and is one of the types that still earns a rich result in 2026. The WebSite type defines the site name. For businesses serving a local area, the LocalBusiness type provides a strong local signal with name, address, phone, and opening hours. Verifying the recommended properties of Article and other types against Google's official gallery is a good habit.

A typical JSON-LD block for a blog post looks like this:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "BlogPosting",
  "headline": "Schema Markup and Structured Data for AEO",
  "datePublished": "2026-06-15",
  "dateModified": "2026-06-15",
  "author": {
    "@type": "Person",
    "name": "Özkan Göçer",
    "url": "https://www.ogocer.com/en/"
  },
  "publisher": {
    "@type": "Organization",
    "name": "Ogocer"
  }
}
</script>

Which Types Still Earn Rich Results in 2026: The End of FAQ and HowTo

Many outdated sources still present FAQ and HowTo markup as a source of rich results, yet the situation has changed completely. The table below summarizes what has been removed and what is still valid.

FAQPage is still a valid schema.org type and leaving it on the page does no harm, but it no longer produces a rich result in Google Search. You can verify the dates from the primary sources, the FAQ document and the HowTo announcement.

Structured Data Is Not a Ranking Factor

A common and costly misconception is that adding schema directly pushes a page higher. Google clearly rejects this: structured data is not a general ranking factor. Even a manual action for a structured data violation does not affect ranking; it only removes rich result eligibility. Google spokespeople including John Mueller and Danny Sullivan have confirmed this repeatedly since 2019. The benefits are indirect: a potential click-through-rate increase via rich results, better entity understanding, and eligibility for AI search features. Google's structured data policies clarify the distinction. In other words, schema does not replace classic SEO work; it is a technical complement to it.

How AI Answer Engines Actually Use Schema

The most critical AEO question is this: do systems like ChatGPT, Perplexity, Gemini, Claude, and Google AI Mode read structured data? Controlled tests by Ahrefs and SearchVIU in October 2025 produced a surprising result. The engines in question extracted only the visible HTML content when fetching a page live; none of them found data placed solely inside JSON-LD, for example a price present only in the schema. The practical reason is this: as of 2026, the major AI crawlers (GPTBot, ClaudeBot, PerplexityBot, OAI-SearchBot, and others) do not execute JavaScript and only read the raw HTML returned by the server. Therefore schema injected later via JavaScript is never seen by these bots. Bot behavior may change over time, but this is the picture today.

A broader model shows that schema plays different roles across the training, indexing, index-based search, and live-fetch stages. Structured data can contribute indirectly especially in index-based systems such as Google AI Overviews and Bing with Copilot, because Google AI Mode uses the same index and markup as organic search. Another important fact is this: there is no special schema type or separate AI optimization for AI Overviews or ChatGPT.

Does Schema Increase AI Citations? Examining the Evidence

The claim that adding schema gets you cited more in AI is the most frequently repeated AEO myth. The most rigorous evidence is Ahrefs' difference-in-differences study: between August 2025 and March 2026, 1,885 pages that added schema were compared against 4,000 matched control pages. The result was striking, because adding schema produced no meaningful citation increase on any platform. A 2.4 percent increase for Google AI Mode and 2.2 percent for ChatGPT were at the level of statistical noise; for Google AI Overviews a small but significant decline of 4.6 percent was measured. Kurt Fischman's academic preprint reached a similar conclusion: generic schema has no independent effect on citations, and the main determinant is organic ranking position. Ahrefs' study shares the findings in detail.

Frequently circulated figures such as 73 percent, 3.2x, 317 percent, or 90 percent of AI answers rely on structured data have no verifiable primary source; most are uncontrolled agency observations, and that 90 percent figure actually belongs to a completely different topic, the accuracy rate of AI Overviews. The frequent presence of schema on cited pages is a correlation, not causation: authoritative publishers already use schema and get cited a lot. Likewise, the narrative that only 12 percent of sites use structured data so early movers win is false; that figure refers to registered domains, while about 54 percent of active sites already use JSON-LD.

Entity SEO, @id, and E-E-A-T Signals

The most valuable use of structured data is clarifying entities. Within JSON-LD, @id gives a node a unique identifier and lets you reference the same entity from different places, while schema.org's url property indicates the entity's actual web address, and the two should be used together. The recommended @id format is the canonical URL plus an identifier, for example the homepage address and an organization tag. An important limit is this: search engines process structured data page by page and do not automatically merge entities sharing the same @id across pages, so it must be reinforced with HTML internal links.

On the author authority side, Person markup clarifies who wrote the content via name, url, jobTitle, and especially sameAs, that is, by linking to authoritative profiles. Still, markup alone does not increase E-E-A-T. Google states that E-E-A-T is not a direct ranking score and cannot be raised simply by adding markup. John Mueller calls structured data an extremely light signal. Real authority is determined by user-visible author information, a consistent publishing history, and off-site recognition.

Implementation, Consistency, and the Most Common Errors

In implementation, the most common mistake is syntax: a single missing comma or unclosed bracket makes the entire schema block invisible to Google's parser. The second common mistake is inconsistency. The data in JSON-LD must match the content visible on the page exactly; Google can disqualify markup that does not represent visible content from rich results, and AI systems understand the content worse when the markup deviates from the visible text. For example, if dateModified in the schema says 2023 while the text references 2026 data, trust is eroded.

For WordPress users, plugins such as Yoast or Rank Math generate basic schema automatically. The right approach is not to turn this off and write by hand, but to audit the generated markup and keep it aligned with the visible content. The critical technical rule is clear: because AI bots do not execute JavaScript, structured data should be embedded in the raw HTML on the server side, not injected only in the browser via JavaScript.

Validation Tools and a Practical AEO Strategy

There are two main validation tools with different purposes. The Rich Results Test tests only Google-specific rich result eligibility and no longer supports FAQ after June 2026. The Schema Markup Validator validates general schema.org syntax and continues to check all types, including FAQPage. The URL Inspection tool in Search Console also shows how Google sees the structured data on the live page.

A practical minimum set for a small business is this: Organization on the homepage (with sameAs), WebSite for the site name, BreadcrumbList for navigation, Article or BlogPosting with a Person author on content pages, and LocalBusiness if you serve a local area. Since FAQ markup no longer earns a rich result, building the question-and-answer structure with visible headings and direct answers delivers the same AEO benefit, because AI engines read the visible text during live retrieval.

Conclusion

In short, schema markup is an indispensable infrastructure for AEO, but not the magic it is hyped to be. Well-built structured data clarifies your content for machines and answer engines, makes it eligible for classic rich results, and strengthens your entity relationships. Ranking or AI citation, however, is won mainly through accuracy, authority, and clear, extractable answers. We have already covered what is AEO and the steps of content optimization for AEO; in the era of AI search, applying all three together is the most solid path. For a broader view, you can explore our AEO services.