Visibility in synthesized responses from Google and AI platforms is no longer a matter of traditional ranking or link authority. In 2026, structured markup—the markup schema—became the deciding factor what determines whether a website's content is cited by ChatGPT, Perplexity, Google AI Overviews, or remains invisible to generative algorithms. Italian publishers continue to focus on basic implementations (FAQPage, Article, Organization), ignoring the real competitive lever: a layered, multi-tiered schema strategy that anticipates Google's fan-out query behavior and maximizes semantic extraction by language models.
This article deciphers the technical landscape of data structuring post-May 2026, moving beyond FAQPage myths as a universal solution. It presents an operational framework based on empirical data and industry research, with specific attention to how query fan-out and knowledge graph positioning require schema implementations that go beyond elementary markup.
Why FAQPage is No Longer Sufficient: The Strategic Shift in May 2026
Google officially retired FAQ rich results on May 7, 2026, closing a nine-year era in which Q&A boxes dominated the search results design. However, this does not mean that FAQPage has become irrelevant.
The reality is more nuanced: If the site already has FAQPage markup deployed, that markup is now inert on Google Search — harmless but producing zero SERP visibility. The discontinuity is not accidental. The schema must match the page's primary topic, not peripheral or supplementary content. Google's March 2026 core update drastically reduced eligibility for rich results for several schema types, including FAQ, Review, and How-To on pages where the markup described non-primary content.
What crucially changed: AI Mode reads structured markup as a trust signal, not a display trigger. Schema that accurately describes content increases the likelihood of AI Mode citing it, even when no traditional rich result is displayed..
How Query Fan-Out Works and Why It Changes Data Structuring
To optimize a schema for the AI era, it is imperative to understand the underlying mechanism of fan-out queries. When a user types a single query, the AI system does not retrieve the best-ranking pages. It launches 8–12 parallel sub-queries—each targeting a different aspect of the user's question—and synthesizes the results into a single answer.
AI Mode uses a fan-out query technique, breaking down a query into subtopics and releasing a multitude of queries simultaneously on behalf of the user. This allows Search to go deeper into the web than a traditional Google search, helping to discover even more hyper-relevant content that matches the query..
The implication for schema markup is direct: SEO specialists need to move beyond optimizing for individual keywords and instead focus on understanding the entire user journey and the many questions someone might ask about a topic. Instead of ranking individual pages, content strategies must aim for topical authority. This means covering a topic exhaustively, addressing all relevant sub-queries and facets, and linking them semantically..
This shift transforms data structuring from a rich snippet tactic to an epistemic foundation. The schema must facilitate entity extraction, semantic relationship mapping, and claim verification—not just formatted display.
The Three Dimensions of Effective Schema Markup for AI: Intent-Match, Entity Clarity, and Semantic Layering
Dimension 1: Intent-Match Schema
The schema should match the primary theme of the page. This is not a minor detail—it's the differentiator between signal and noise for AI systems. An e-commerce product page with an supplementary FAQPage that addresses secondary questions will have a lower citation priority than a page completely focused on that Q&A.
Operational Test: perform an audit of each schema on the site and verify that the type (@type) describes the primary visible content, not peripheral content or widget sections. Pages with intent-schema misalignment not only lose AI Overviews visibility but risk manual flags for «hidden or misaligned data.».
Dimension 2: Entity Clarity
Websites with clean entity schemas are cited more frequently by AI responses because AI can confidently resolve who or what the source is.. This means that the implementation of the plan regarding the organization, key people, and institutional relationships becomes crucial.
For example, Schema markup is one of the few tools that SEOs have to make entities and relationships explicit and understandable for an AI: this is a person, they work for this organization, this product is offered at this price, this article was written by that person, etc.
Practical Implementation: Ensure each asynchronous page (Team, About, Product) contains Organization schema with properties sameAs that link to Wikipedia, Wikidata, and social profiles. AI structured schema contributes to knowledge graph positioning. When AI systems build understanding of entities and relationships, structured schema provides the explicit connections. A strong presence in the knowledge graph increases the likelihood that AI will recognize and consider your brand a trusted, authoritative source..
Dimension 3: Semantic Layering
The implementation of a single-layer schema (a single @type per page) is now obsolete. Pages with 3–4 complementary schema types (like Article + FAQPage + BreadcrumbList) received 2x more citations than pages with a single schema type..
Schema layering for maximum AI comprehension means incorporating related schema types within a parent entity. For example, nesting FAQPage schema within an Article schema creates a composite signal that tells AI engines both the content type and the specific Q&A pairs it contains — drastically improving extraction confidence..
The Five Schema Categories for AI Overviews and Fan-Out Queries (2026)
Tier 1: Primary Schema (Required)
The five types of schema that make a difference in 2026: Organization, Article (or BlogPosting), FAQPage (with warnings), Product, and LocalBusiness.
Article/BlogPosting Schema: Essential for any content-based website. AI engines use the Article schema to understand authorship, publication date, and thematic focus. This helps determine content freshness and authority—both critical for citation decisions..
Organization Schema: The Organization schema communicates the company name, address, logo, contact details, and areas served to AI systems. This helps build your brand identity and feeds into features like Google's Knowledge Panel..
FAQPage Schema (Current Use): Although rich results have been withdrawn, The reason is simple: AI platforms present information in a question-and-answer format. When your content already exists in that format and you explicitly flag it through schema, AI systems can extract, verify, and cite it with confidence..
Tier 2: Secondary Schemas (Highly Recommended)
HowTo Schema If you’re creating instructional content, the How-To format is a must. AI assistants love step-by-step processes because they align perfectly with the way users ask questions. “How do I…” queries trigger AI Overviews 73.1% of the time.
Product Schema: For e-commerce sites, Product schema is necessary on every product detail page. It supports direct surfacing of the product in shopping queries and AI-powered comparison responses..
Tier 3: Specialized Schemas for Entity Mapping
BreadcrumbList Schema: Structure site navigation not only for human users, but for AI semantic understanding.
Speakable Schema: This is a type of schema specifically designed for AI and voice assistants. The Speakable schema identifies the most citable passage within a long document, improving AI extraction accuracy.
Practical Implementation: Stratified JSON-LD and Cross-Platform Validation
Step 1: Format Selection — JSON-LD is the Only Viable Standard
JSON-LD in the head remains the preferred format after March 2026: Google has not changed its preference for JSON-LD delivered in the document's head. Microdata and RDFa implementations have not increased in effectiveness..
Use JSON-LD. Every AI engine I've tested prefers it because it's cleanly separated from the HTML and is easier to parse programmatically.
Step 2: Accurate Validation of Visible Content
This is critical and where most people make mistakes. Your schema markup must match what is actually visible on the page. AI engines check for this consistency, and discrepancies can get you penalized or ignored..
Example of a common error: your Article schema declares “Published: January 15, 2026” but the page shows a different date. AI systems recognize these inconsistencies and take them into account in citation decisions.
Step 3: Strategic Schema Nesting
Instead of implementing separate, disconnected schema blocks, nest related types to show relationships. A typical implementation for a blog post with FAQs:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "Come Ottimizzare Schema Markup per AI Overviews",
"datePublished": "2026-05-20",
"dateModified": "2026-05-20",
"author": {
"@type": "Organization",
"name": "AI Publisher WP",
"url": "https://aipublisherwp.com"
},
"mainEntity": {
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "Qual è la differenza tra schema markup e rich results?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Lo schema markup è il codice strutturato che aiuta i motori di ricerca a comprendere il contenuto. I rich results sono il modo in cui Google visualizza quel contenuto nei risultati di ricerca. Dopo maggio 2026, lo schema rimane critico per la citazione AI anche se i rich result non sono più visualizzati."
}
}
]
}
}
</script>This nested structure communicates to AI systems both the primary content type (Article) and the specific Q&A pairs it contains (FAQPage), maximizing trust in extraction.
Step 4: Entity Relationship Mapping via @graph
For more complex sites (e-commerce with multiple authors, marketplaces with suppliers), use @graph with @id to build explicit relationships:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@graph": [
{
"@type": "Organization",
"@id": "https://example.com#organization",
"name": "Brand Name",
"url": "https://example.com",
"sameAs": [
"https://www.wikidata.org/wiki/Q123456789",
"https://www.wikipedia.org/wiki/Brand_Name"
]
},
{
"@type": "Product",
"@id": "https://example.com/product-x#product",
"name": "Product X",
"brand": { "@id": "https://example.com#organization" },
"manufacturer": { "@id": "https://example.com#organization" }
}
]
}
</script>This approach builds the relationship graph that AI systems use for knowledge graph positioning.
Measuring Success: Beyond Traditional Rank Tracking
Testing and measuring the performance of AI structured schemes requires tracking signals beyond traditional rank tracking. AI citations don't always correlate with position — a page can rank 5th in organic results but be cited 1st in AI overviews..
Metrics to Monitor
- AI Overview Impression Rate: Percentage of impressions from AI-generated results (via Google Search Console and Semrush AI Toolkit)
- Citation Frequency per Query: How often is your content cited by ChatGPT, Perplexity, and Google AI Mode for target queries
- Schema Error Rate: Schema validation errors reported in Search Console > Enhancements
- Knowledge Graph Presence Tracking Knowledge Panel and Entity Card Visibility in Google Search
Use GEO and Claude Monitoring Systems to track brand visibility in real-time AI responses.
Schema Markup in the Context of GEO and Fan-Out Queries: A Practical Use Case
The combination of effective schema markup with the understanding of fan-out queries amplifies AI visibility in a compounded manner. If an Italian publisher targeting the query “How to choose WordPress hosting” implements:
- Article schema with updated dateModified (indicates AI freshness)
- HowTo schema nested describing the evaluation steps (responds to fan-outs like “best hosting providers” and “hosting choice factors”)
- FAQPage for common questions (“What is disk space?”, “What is bandwidth?”)
- Organization schema with sameAs link to Wikipedia/Wikidata (entity clarity)
Google AI Mode will identify facets such as high-performance hosting, security, technical support, and price — and the site will have opportunities to be cited in multiple parallel sub-queries. Websites with 80% topical coverage maintain 85.4% of their AI visibility despite the instability of the fan-out query at 73%. This finding explains why comprehensive thematic clusters outperform optimized individual pages in AI search..
To further explore visibility and citation strategies, consult GEO and the construction of real citability in AI Mode.
Common Mistakes and How to Avoid Them
Error 1: Schema Mismatch with Primary Content
Adding FAQPage to a product page where the primary content is the product description will penalize its SEO. The supplemental schema must be hierarchically subordinate and semantically related to the primary type.
Error 2: Forgotten datesModified
One mistake that was made was implementing the perfect blog schema markup for a client but forgetting to update the dateModified property when we refreshed the content. The AI engines kept citing the old information because the schema told them the page had not been updated. Always update your schema in force.!
Error 3: No Cross-Tool Validation
Don't skip validation. Use Google's Rich Results Test to confirm your schema is error-free and eligible for rich results. For more in-depth checks, use the Schema.org validator or the structured data testing tools within Search Console. Invalid schema can silently break your eligibility for AI and rich result visibility, even if the content quality is high..
Integration with LLM Crawlbot Management and WordPress Automation 7.0
To maximize the impact of your schema markup, it's essential to control the crawling of AI bots. A robots.txt strategy optimized for GPTbot, Claudebot, and Petalbot Allows relevant AI bots to access the site while blocking irrelevant traffic.
Furthermore, WordPress 7.0 Armstrong introduces API connectors for OpenAI, Claude, and Google Gemini, enabling the automation of AI model-based markup schema generation during content publishing.
FAQ
What is the difference between Schema Markup and Rich Results after May 2026?
Schema markup is structured code (JSON-LD, Microdata, RDFa) that communicates the meaning of content to search engines and AI systems. Rich results are how Google displays that content in search results (rating stars, FAQ boxes, etc.). If your site already has FAQPage markup deployed, that markup is now inert in Google Search — harmless but producing zero SERP lift. However, the schema remains critical for citation in AI Overviews and AI Mode, even without visible rich results.
Should I still implement FAQPage if Rich Results have been withdrawn?
Yes, with one condition: only if the FAQ content is primary or directly supports the main topic of the page. AI platforms present information in a question-and-answer format. When your content already exists in that format and you explicitly signal it through schema, AI systems can extract it, verify it, and cite it with confidence.. FAQPage remains one of the schema categories with the highest citation rate for AI systems.
How Does Query Fan-Out Affect Schema Strategy?
Query fan-out means that a single user search will generate 8–12 parallel sub-queries. This requires your schema to support comprehensive topical coverage, not just the main keyword. Implement nested schema (Article + HowTo + FAQPage) and ensure that the Organization schema clearly establishes the entity's identity to allow AI systems to recognize you in multiple fan-out contexts.
Which Schema Format is Best: Microdata, RDFa, or JSON-LD?
Use JSON-LD. Every AI engine I've tested prefers it because it's cleanly separated from the HTML and is easier to parse programmatically. JSON-LD in the head remains the preferred format after March 2026.
How do I check if my schema markup is working for AI Overviews?
Track three main metrics: (1) AI Overview Impression Rate in Google Search Console, (2) Citation Frequency on ChatGPT and Perplexity for target queries using direct queries, (3) Schema Error Rate in Search Console Enhancements. Do not confuse with traditional rank tracking — A page can rank 5th in organic results but be cited 1st in AI Overviews.
Conclusion: Schema Markup as the Foundation of AI Visibility in 2026
The transition from rich results to generative AI systems has transformed schema markup from a display tactic to an infrastructural foundation for online visibility. In 2026, Content with proper schema markup is 2.5x more likely to appear in AI-generated answers. For Italian publishers, this means that the schema strategy must evolve from single-layer implementations (FAQ, Article, Organization in silos) to layered and multi-contextual stacks that anticipate fan-out query behavior and enable entity mapping in Google's knowledge graph.
Priority actions are: (1) Audit intent-schema alignment on all critical pages, (2) Implement layered JSON-LD with nesting (Article + HowTo + FAQPage + Organization), (3) Build entity clarity via sameAs linking to Wikipedia/Wikidata, (4) Validate cross-tool (Google Rich Results Test + Schema.org + direct tests on ChatGPT/Perplexity), (5) Monitor AI Overview impressions and citation frequency in real-time.
Entity Authority, the new ranking factor replacing Domain Authority in 2026, it is built entirely on a foundation of accurate schema markup and explicit semantic relationships. Websites that invest in structured schema strategies today will be positioned to dominate AI visibility not only in 2026, but for years to come. The question is no longer “Do we need a schema?”, but “Is our schema strategy sophisticated enough to compete in a landscape of fan-out queries and knowledge graph positioning?”
