I've been saying for years that structured data is underinvested in most SEO programs. Now I'm pounding the table even harder.
Here's why: both Google and Microsoft publicly stated this year that they use schema markup for their generative AI features. Google was explicit—structured data is critical for modern search because it's efficient, precise, and easy for machines to process.
ChatGPT then confirmed it uses structured data to determine which products appear in its results.
The companies building AI search are telling us directly: structured data matters. And research shows pages with comprehensive schema markup are 36% more likely to appear in AI-generated summaries.
If you're not investing here, you're leaving visibility on the table.
Why Schema Matters More Now
Traditional SEO benefited from schema through rich results—those enhanced search listings with stars, images, prices, FAQs. That's still valuable.
But AI search adds another layer. AI systems need to understand what your content is about, who created it, and whether it's trustworthy. Schema provides those signals in a machine-readable format.
Think about it from an AI system's perspective. It's processing millions of pages trying to answer a user's question. Which source should it cite?
A page with clear schema markup says: "I'm an article about this specific topic, written by this expert person who works at this organization, published on this date, covering these specific things."
A page without schema is just text that the AI has to interpret. The schema page removes ambiguity. It makes the AI's job easier. So the AI prefers it.
The Citation Connection
BrightEdge research showed higher citation rates in Google's AI Overviews for pages with robust schema markup. This isn't theoretical anymore—it's showing up in the data.
One analysis suggested schema contributes about 10% of ranking factors for Perplexity specifically. That's meaningful for a single tactic.
And here's something that surprised me: schema helps with entity disambiguation. When AI systems are trying to understand whether you're discussing "Apple" the company or "apple" the fruit, schema makes it clear. That clarity increases your chances of being cited for relevant queries.
Which Schema Types Actually Matter
Not all schema is created equal. Here's what I prioritize for AI visibility.
Organization Schema
This establishes your business identity. You need to include your name (exactly as you want to be recognized), your logo, contact information, social media profiles through the sameAs property, founding date, and description.
The sameAs property is particularly important. AI systems cross-reference entities across multiple sources. Linking to your Wikipedia page, Wikidata entry, LinkedIn, and other authoritative profiles dramatically increases entity recognition.
Person Schema for Authors
Author expertise is huge for E-E-A-T and AI citations. For every content author, you should include their name, job title, affiliation through Organization, credentials and qualifications, sameAs links to professional profiles, and an image.
This connects your content to identifiable experts. AI systems can verify that your authors are real people with relevant credentials.
Article/BlogPosting Schema
For every piece of content, include the headline, author linked to Person schema, publisher linked to Organization schema, publication date, last modified date, image, and description.
The dates are particularly important. AI systems weight recency, and explicit date schema signals when content was created and updated.
FAQPage Schema
This is gold for AI search. FAQ schema provides direct answers to questions, is highly extractable for AI responses, can appear in rich results on Google, and matches how users query AI systems.
If your content answers questions, structure at least some of it as FAQ schema.
Product Schema
For e-commerce, you need to include name, image, description, SKU, brand, price and availability, and reviews and ratings.
ChatGPT specifically mentioned using product schema for shopping results. If you're not implementing this, you're invisible to AI shopping features.
LocalBusiness Schema
For businesses with physical locations, include name, address, phone, operating hours, geo-coordinates, price range, and payment methods.
This matters for voice search and AI assistants that handle local queries.
Review/AggregateRating Schema
Social proof signals matter too. If you have reviews, include rating value, review count, and best/worst ratings.
Both traditional search and AI systems use these as trust signals.
Implementation Best Practices
How you implement schema matters as much as what you implement.
Use JSON-LD format. JSON-LD is the preferred format. It's cleaner, easier to maintain, and recommended by Google. Don't use microdata or RDFa unless you have a specific reason.
Place JSON-LD in the head of your pages, or at the end of the body. Don't scatter it throughout the content.
Validate everything. Use Google's Rich Results Test for every page template. Errors in schema can prevent it from being read at all. Common mistakes include missing required properties, wrong data types (string vs number), mismatched content between schema and visible page, and invalid URLs or images. Fix these before deploying.
Keep schema consistent with page content. Schema should describe what's actually on the page. If your Article schema says the author is "John Smith" but the visible byline says "JS," that's a mismatch.
AI systems are increasingly checking for consistency between structured data and page content. Inconsistencies can hurt rather than help.
Connect schema into a graph. Individual schema items are good. Connected schema is better.
Your Article schema should reference your Person schema for the author. Your Person schema should reference your Organization schema for affiliation. Your Organization schema should link to your social profiles via sameAs.
This web of connected schema creates a knowledge graph about your content and your organization. AI systems can traverse these connections to understand relationships.
Keep it updated. Schema with outdated information is worse than no schema. If your Organization schema says your CEO is someone who left three years ago, that's a problem.
Update schema when content is modified, when author information changes, when prices or availability change, and when business information changes.
Automated systems that sync schema with your CMS or database are worth the investment.
The January 2026 Deprecations
Google deprecated some schema types this year: Practice Problem, Dataset, Sitelinks Search Box, and a few others. Some people panicked.
Don't panic. Google is optimizing for commonly used types, not eliminating structured data. The core schema types—Product, Article, Organization, Person, Review—remain fully supported and prioritized.
If anything, the deprecations signal that Google is focusing resources on the schema types that matter most. That's validation, not warning.
Common Mistakes I See
After auditing lots of sites, these are the recurring problems.
Organization schema only on homepage is a big one. Your Organization schema should be on every page, or at least every page that references your company. Don't limit it to the homepage.
Missing author schema on articles is another common issue. Every article should connect to Person schema for the author. Anonymous articles get no author entity benefit.
Outdated lastModified dates are problematic too. If you're auto-generating lastModified as today's date on every page load, you're devaluing the signal. Only update it when content actually changes.
No sameAs properties is the most underused schema property. Cross-referencing your entities across platforms is huge for AI recognition.
Schema that doesn't match page content happens more than you'd think. I see FAQ schema for questions that aren't actually answered on the page. Article schema with wrong publication dates. Product schema with prices that don't match the displayed price. These mismatches hurt you.
The ROI of Getting This Right
Schema implementation has compounding returns.
In traditional search, you get richer results, higher click-through rates, and better feature visibility.
In AI search, you get higher citation rates, better entity recognition, and more accurate brand representation.
For AI features specifically, pages with comprehensive schema are getting cited 36% more. That's a massive edge for a technical implementation.
And once your schema infrastructure is built, maintaining it is relatively low effort. It's not like content creation where you need to keep producing. Get the templates right, and schema scales automatically.
What I'd Prioritize
If you're starting from scratch or auditing your current implementation, here's the order I'd approach things.
Start with Organization schema with sameAs to establish your entity identity first. Then add Person schema for authors to connect content to identifiable experts. Next, implement Article or BlogPosting for content with proper author and date connections. Add FAQPage for Q&A content since it's highly extractable for AI. If you're in e-commerce, Product schema is required for AI shopping features. And if you have physical locations, LocalBusiness matters for local and voice search.
Get these right before worrying about more exotic schema types.
The Bigger Picture
Schema markup is your machine-readable identity. It tells AI systems who you are, what you do, who your experts are, and what your content covers.
In a world where AI systems increasingly mediate between users and information, that machine-readable identity matters enormously. It's not optional anymore—it's infrastructure.
The 85% of enterprises increasing structured data investment aren't doing it for fun. They're doing it because the data shows it works.
If you're in the other 15%, you're falling behind in AI visibility while your competitors pull ahead. That gap gets harder to close over time.
Schema isn't the most exciting topic in SEO. But it might be one of the highest-ROI investments you can make right now.