Google can crawl your page. That doesn’t mean it understands it. Crawling reads the words; understanding assigns meaning to them. Those two things are not the same, and the gap between them is exactly where structured data for SEO does its work. It’s a standardised code layer that sits beneath your visible content and tells search engines, in precise, unambiguous terms, what they’re looking at: this is a product, here is its price, this is the author, this is a question with its accepted answer.
In 2026, that distinction matters more than ever. AI Overviews, Perplexity citations, and Bing’s AI answers no longer simply retrieve pages, they interpret content, attribute it to sources, and decide whether it’s worth citing. Structured data accelerates that process, making your content legible to machine systems in ways that plain HTML simply cannot. Many agencies now apply schema markup as a standard part of every content package they produce, across video transcripts and written articles alike. Getting crawled is no longer enough. You need to be understood.
What structured data is and how it works in search
The gap between what you see and what Google reads
A human visitor reads your page and immediately grasps context: this is a product page, that’s a review, here’s a step-by-step guide. Google’s crawler sees the same HTML but has to infer meaning from patterns and signals. That’s educated guesswork, not confident comprehension. Structured data removes the ambiguity by wrapping your content in a standardised code layer that directly states the facts: this is a Product , this is its price, this is the author. The result is a page Google can interpret with confidence rather than assumption.
How schema.org provides the shared vocabulary
Schema.org is a collaborative vocabulary created by Google, Bing, Yahoo, and Yandex. It defines hundreds of content types, from Article and Product to VideoObject and FAQPage , each with a set of named properties. When you use schema.org types correctly, every major search engine and AI system reads the same structured signals. This standardisation is what makes your markup meaningful across platforms, not just within Google Search. It’s a shared language, and every major search engine has agreed to speak it.
Why JSON-LD is the format to use
JSON-LD vs. Microdata vs. RDFa
Google supports all three structured data formats, JSON-LD, Microdata, and RDFa, but explicitly recommends JSON-LD in most cases, and the reason is straightforward. JSON-LD sits in a separate script block, completely independent of your HTML. Microdata and RDFa embed attributes directly into the HTML elements they describe: functional, but messy to write and easy to break during template updates. For any site that prioritises maintainability, JSON-LD is the clear choice.
Where to place your JSON-LD on the page
Google accepts JSON-LD in either the or the of your page, so placement is flexible. Each block opens with @contextpointing to schema.org, declares an @type, and then lists the relevant properties. Three practical rules apply across all implementations:
- Use absolute URLs for images and page links.
- Keep the JSON syntactically valid, no trailing commas, double quotes only.
- Ensure every value in the markup reflects content that is actually visible on that page.
That last point matters specifically for rich result eligibility. Per Google’s Search Central documentation, markup that describes content not visible to users fails the content-matching policy and blocks rich result eligibility entirely.
Structured data for SEO: schema types that drive rich results
Article and FAQPage schemas
Article (and its subtypes NewsArticle and BlogPosting) is used for editorial content and supports Top Stories features in Google Search. According to Google’s Search documentation, it requires headline, image, author, datePublished, and a publisher block containing the organisation name and logo.
FAQPage marks up question-and-answer content using Question items, each with an acceptedAnswer . It’s worth noting that Google has narrowed FAQPage’s rich result eligibility in recent years and currently limits visible FAQ rich results to certain authoritative sites. The schema still provides useful machine-readable signals and is worth implementing, but you should not rely on it producing a visible rich result for most pages.
HowTo and VideoObject schemas
HowTo structures step-by-step instructions using a step array of HowToStep items, each carrying its own name and text. Like FAQPage, HowTo rich result eligibility has been narrowed by Google and is no longer broadly available for most sites, check Google’s current Search Gallery for up-to-date eligibility guidance before prioritising this type for rich result purposes. That said, the structured signals are still valuable for AI comprehension of instructional content.
VideoObject is the schema for any page hosting or embedding a video. It requires name , thumbnailUrl, and uploadDate as the minimum set, with description and contentUrl or embedUrl recommended to improve visibility further. Google uses VideoObject markup to surface video thumbnails in search results and to index video content for AI systems.
Product and business schemas
Product schema is the foundation for product rich results and merchant listings. At minimum it needs name , image, and an offers block containing price, priceCurrency, and availability . To be competitive, add description , sku, brand, and an aggregateRatingblock where genuine reviews exist. The availability field accepts specific schema.org values: InStock, OutOfStock, and PreOrderare the most commonly used. Keep pricing data current, stale prices are one of the most common reasons product rich results stop appearing.
For service businesses, LocalBusiness and Organisation schemas help search engines populate knowledge panels, and Review and AggregateRating schemas can attach to supported types to surface star ratings in eligible contexts.
Writing JSON-LD for your key page types
Article and FAQPage in practice
The Article schema is straightforward once you have the required properties in place. Here is a minimal but valid example:
<code>{ "@context": "https://schema.org", "@type": "Article", "headline": "How structured data for SEO improves AI search visibility", "image": "https://example.com/images/structured-data-guide.jpg", "author": { "@type": "Person", "name": "Jane Smith" }, "datePublished": "2026-06-01", "publisher": { "@type": "Organisation", "name": "Example Agency", "logo": { "@type": "ImageObject", "url": "https://example.com/logo.png" } } }</code>For FAQPage, the markup centres on mainEntity: an array of Question items, each with a name (the question text) and an acceptedAnswer containing the response. A compact example:
<code>{ "@context": "https://schema.org", "@type": "FAQPage", "mainEntity": [ { "@type": "Question", "name": "What is structured data?", "acceptedAnswer": { "@type": "Answer", "text": "Structured data is a standardised format for annotating web content so search engines can interpret it precisely." } } ] }</code>Remember: FAQPage schema applied to a page that doesn’t display those questions to users will fail Google’s content-matching policy and block rich result eligibility. This isn’t a technicality, Google treats it as a spam signal.
Product schema: the properties that matter most
Product schema’s minimum set is name , image, and an offers block. To be genuinely competitive in product rich results, include description , sku, brand, and an aggregateRating block where real reviews exist. Keep inventory status in sync with your actual stock, mismatches are caught quickly and damage eligibility.
<code>{ "@context": "https://schema.org", "@type": "Product", "name": "Example Product", "image": "https://example.com/images/product.jpg", "description": "A clear, specific description matching visible page content.", "sku": "EX-001", "brand": { "@type": "Brand", "name": "Example Brand" }, "offers": { "@type": "Offer", "price": "49.99", "priceCurrency": "GBP", "availability": "https://schema.org/InStock" } }</code>VideoObject markup: why every video page needs its own schema
VideoObject is often overlooked, yet any page embedding or hosting a video should carry its own dedicated markup. Google uses it to surface video thumbnails directly in search results and to make video content legible to AI indexing systems. The minimum required properties are name , thumbnailUrl, and uploadDate ; add description and contentUrl or embedUrl to improve visibility further.
Production teams that integrate VideoObject markup into their publishing workflow, rather than treating it as a post-publication audit task, give their video content a meaningful head start in AI search visibility. At This Video Works , schema markup is built into the content production pipeline for video transcript pages and associated articles, so content is machine-readable from the moment it goes live.
How to test, validate, and monitor your schema
The right tools for testing and validation
Start with Google’s Rich Results Test. Paste a URL or code snippet and it shows exactly which rich results your markup qualifies for, flagging any errors blocking eligibility. The Schema Markup Validator at schema.org is broader, it checks whether your markup is valid according to schema.org specifications, independent of Google-specific eligibility rules. For generating initial markup without writing from scratch, Merkle’s Schema Markup Generator produces clean JSON-LD snippets quickly and is a practical starting point.
Using Search Console to monitor at scale
Once structured data is live, Google Search Console’s Rich Results reports show which pages are valid, which have warnings, and which have errors blocking eligibility. The URL Inspection tool lets you check how Google interprets any specific page after implementation. When you fix an error, use the Validate Fix button in Search Console to prompt Google to recrawl and restore eligibility. This step is routinely skipped, and it slows recovery unnecessarily. Fix the markup, validate it with the Rich Results Test, then click Validate Fix, in that order.
The most common errors and how to fix them
Google most frequently reports five types of structured data errors:
- Unparsable JSON: syntax problems such as missing commas, unclosed brackets, or single instead of double quotes. Fix by running the markup through a JSON validator and correcting each flagged line.
- Missing required fields: the schema type is present but essential properties are absent. Check Google’s documentation for that specific type and add what’s missing.
- Invalid values: wrong data type, malformed dates, or relative instead of absolute URLs. Match each property to the expected format described in the schema.org documentation.
- Content mismatches: markup describes content not visible to users. Ensure everything in your schema is reflected by visible on-page content.
- Duplicate schema blocks: multiple conflicting descriptions of the same entity, often introduced by CMS plugins. Consolidate to one accurate block per entity and remove duplicates from theme or plugin sources.
Structured data for SEO in the AI search era
How AI Overviews depend on structured signals
AI-powered search features interpret content, attribute it to sources, and determine whether it merits citation. Structured data reduces the friction in that process. When your Article schema carries a clear author, datePublished, and publisher , an AI system can attribute the content with confidence. When your FAQ markup lists direct answers to common questions, those answers become candidates for AI-generated responses. Structured data doesn’t guarantee citation, but it removes the ambiguity that prevents it.
Applying schema at every layer of content production
The businesses gaining ground in AI search in 2026 are treating structured data not as a post-publication technical task, but as a production standard applied from the outset. This Video Works builds this into every client project, schema markup is applied to video transcript pages, written articles, service pages, and business details as part of an integrated content and SEO package. For any business serious about visibility in AI search, the point where content creation and structured data meet is where the competitive advantage now sits.
The practical recap
Structured data for SEO is no longer optional. It’s the layer that tells search engines and AI systems exactly what your content is, who created it, and why it’s relevant. The implementation discipline is consistent regardless of page type: use JSON-LD, choose the correct schema types, include all required properties, and ensure your markup matches what users actually see on the page. Validate with the Rich Results Test before anything goes live. Monitor through Search Console on an ongoing basis.
In the AI search era, the stakes are higher because AI citation depends on machine-readable signals, not keyword relevance alone. An article with clear Article schema, correct author attribution, and a matching publisher block is far easier for an AI system to cite than an identically written article with no structured data at all. The content quality still has to be there, schema makes that quality legible to machines. Think of it as a translation layer: structured data converts your expertise into a format that AI systems can read, assess, and attribute.
Organisations more likely to earn AI visibility at scale are those implementing structured data consistently across all their content, not just high-priority pages. Start with your most important page types, validate every block before it goes live, and treat schema markup as part of your content workflow rather than an afterthought. That shift in approach is what separates the sites that get cited from the ones that merely get crawled.
