Why structured data
matters for AI
Structured data is the most direct way to tell AI what your website contains. No hints. No interpretation. Just an explicit, machine-readable description of your business, products, services and content. 71% of pages cited by ChatGPT use schema markup. Pages with structured data are cited 3.2 times more often than pages without. Yet only 12.4% of websites use it. This guide explains why structured data has become so important, which schema types have the most impact and how to implement it for your UK business.
71%
of ChatGPT-cited pages use schema markup
3.2x
more citations with structured data
12.4%
of websites use schema markup
68%
of AI-cited websites have strong structured data
Structured data: the language AI understands fastest
Structured data is information organised in a fixed format that machines can read directly. In the context of websites, it means schema markup: code you add to your HTML to tell AI and search engines what your page contains. The most used format is JSON-LD (JavaScript Object Notation for Linked Data), a block of code in the head of your page that describes what the page is, what it contains and how it relates to other entities.
Without structured data, AI must interpret your page. It reads your text, tries to understand what your business is, which services you offer, where you are located and what your expertise is. That process is error-prone. AI can misinterpret information or miss important details. With structured data, you tell AI all this information directly. No interpretation needed. It is the difference between someone trying to guess your house number from your description and someone getting the GPS coordinates.
The numbers speak for themselves
The impact is measurable. 71% of all pages cited by ChatGPT have schema markup. 65% of pages in Google AI Mode use structured data. Pages with structured data are cited 3.2 times more often in AI answers. Yet only 12.4% of all websites (45 million domains) use schema markup. That means 87.6% of the internet is missing this advantage. For businesses that do implement it, that is a significant competitive edge. Learn more about how AI selects its sources in how AI selects its sources.
Schema markup is based on the schema.org vocabulary, jointly developed by Google, Bing, Yahoo and Yandex. It defines hundreds of types (Organization, Product, Article, FAQPage, LocalBusiness, etc.) each with dozens of properties. By choosing the right type and filling in the relevant properties, you make your content machine-readable in a standardised format that all AI platforms understand.
JSON-LD is the format you should use. Microdata and RDFa also exist, but JSON-LD is the standard Google recommends and all AI crawlers process most effectively. The advantage of JSON-LD is that it sits separately from your HTML content. It is a standalone block of code you place in the head of your page, without needing to modify your HTML structure. That makes it easier to implement and maintain.
Structured data is no longer an optional optimisation. It is core infrastructure for AI visibility. Without schema markup, you are missing one of the strongest signals you can give AI.
The schema types with the most impact
Organization / LocalBusiness
The foundation. It tells AI what your business is, where it is located, what it does and how to reach it. For UK businesses with a physical location, use LocalBusiness with properties like address, geo, openingHours, telephone and areaServed. Include your Companies House number and VAT registration. This is your digital identity card for AI.
FAQPage
One of the most effective schema types for AI citations. FAQPage schema tells AI your page contains question-answer pairs. AI can extract and cite those pairs directly. Each FAQ is a potential citation. A FAQ section with five questions and answers, combined with FAQPage schema, gives AI five concrete passages to cite.
Article / BlogPosting
Tells AI your page is an article with an author, publication date, modification date and topic. Combined with Person schema for the author, this strengthens E-E-A-T enormously. AI can link the author to external profiles and publications. The dateModified field signals freshness.
Product
For online shops and businesses offering products with prices. Product schema includes name, description, price, availability, brand and aggregateRating. This is the most direct way to tell AI what you sell, for how much and whether it is available. When ChatGPT answers "best cordless drill under £100", it looks for pages with Product schema.
HowTo
For instructional content, guides and step-by-step plans. HowTo schema structures your content into steps AI can present sequentially. When someone asks ChatGPT "how do I register for VAT?", AI can convert a HowTo schema directly into a step-by-step answer with a citation to your source.
Person
Crucial for E-E-A-T. Person schema describes your content's author: name, job title, employer, education, sameAs links to LinkedIn and other profiles. For UK solicitors, include your Law Society number. For accountants, your ICAEW or ACCA membership. Pages with 3+ schema types have 13% higher chance of AI citation.
Does your website have the right schema markup?
VestVale automatically monitors whether ChatGPT, Gemini, Claude and Google AI cite your business. Find out if your structured data is working.
Entity disambiguation: how schema markup helps AI understand your business
One of the biggest challenges for AI is disambiguation: distinguishing between entities with the same name. There are thousands of businesses in the UK called "Smith and Partners." AI needs to know which one you are. Structured data solves that problem. By providing your business name, address, Companies House number and services as schema markup, you create a unique digital identity that AI can match with information on other platforms.
The sameAs property is critical here. With sameAs, you link your website to your Google Business Profile, LinkedIn page, Companies House listing, Trustpilot profile and other online presences. AI uses these links to connect all your online presence to a single entity. Without sameAs, AI might treat your website, your Google Maps listing and your LinkedIn page as three separate entities. With sameAs, AI understands they are the same business.
A practical example
Consider a solicitor's firm "Johnson and Partners" in Manchester. Without schema markup, AI must guess which "Johnson and Partners" it is. With LocalBusiness schema that includes the address, practice areas, team members and sameAs links to the Law Society and Companies House, the firm is immediately recognisable and correctly categorisable. When someone asks ChatGPT "which solicitor in Manchester specialises in employment law?", AI can match the firm to that question via the structured data. Learn more about this process in how AI recommends businesses.
Consistency between schema markup and visible content is essential. If your schema says you are a restaurant in London, but your page text describes a catering business in Birmingham, AI detects that inconsistency. That leads to lower trust scores. Schema markup must be a machine-readable mirror of your human-readable content. No exaggerations, no claims that are not supported on the page.
The aggregateRating property is particularly valuable. By including your average review score and the number of reviews in your schema, you give AI concrete trust signals. A business with "aggregateRating: 4.7, ratingCount: 312" gives AI quantitative evidence of customer satisfaction. That strengthens the trustworthiness of your entire website and increases the chance of citation.
Ensure your business name, address and phone number match exactly across your website, schema, Google Maps and all other platforms. NAP consistency is the foundation of entity disambiguation.
NAP consistency (Name, Address, Phone) is the foundation of entity disambiguation. Ensure your business name, address and phone number match exactly across your website, schema markup, Google Business Profile, Trustpilot, Checkatrade and all other platforms where your business appears. Every inconsistency fragments your digital identity and makes it harder for AI to recognise you as a clear entity.
Implementing schema markup correctly
Always use JSON-LD
JSON-LD is Google's recommended format and is processed most effectively by all AI crawlers. Place it as a script block in the head of your HTML. It sits separately from your content, making it easy to implement and maintain. Most CMS platforms including WordPress have plugins that generate JSON-LD automatically.
Validate with Rich Results Test
Use Google's Rich Results Test to check your schema. 49% of websites with schema markup have errors in their implementation. Errors are not forgiven by AI; they are ignored. Test every page after implementation. A schema with syntax errors is the same as no schema at all.
Combine multiple schema types
Pages with 3+ schema types have 13% higher chance of AI citation. Combine Article + Person + Organization on blog content. Product + Organization + AggregateRating on product pages. FAQPage + Article on knowledge pages. Each additional relevant schema type adds context for AI.
Fill all relevant properties
An Organization schema with only a name is nearly worthless. Fill in address, telephone, email, sameAs, logo, description, founder and areaServed. For UK businesses, include your Companies House number and service areas. The more properties you fill, the more complete the picture AI gets of your business.
Keep schema in sync with content
Schema markup must exactly match the visible content on your page. If your price changes, update the Product schema. If your address changes, update the LocalBusiness schema. Inconsistency leads to lower trust scores. Build schema updates into your content update process.
Use sameAs for linking
Link your website via sameAs to your Google Business Profile, LinkedIn, Companies House, Trustpilot and trade associations. This helps AI recognise all your online presence as a single entity. Each sameAs link strengthens your digital identity. For regulated professions, include your professional body listing.
Common mistakes with structured data
Choosing the wrong schema type. An online shop using Organization schema instead of Product schema on product pages misses the ability to make prices, availability and reviews machine-readable. Choose the most specific type that fits your page type. LocalBusiness instead of Organization if you have a physical location. BlogPosting instead of Article if it is a blog post. Service instead of a generic description if you are describing what you offer.
Incomplete properties. An Article schema without author, datePublished and dateModified misses essential E-E-A-T signals. A LocalBusiness without openingHours and geo misses local signals. Every missing field is a missed opportunity to give AI information. Fill each schema as completely as possible. For UK businesses, this means including your opening hours, service area, accepted payment methods and any relevant certifications.
Schema on the right pages
Schema on the wrong pages. Organization schema belongs on your homepage, not on every page. Product schema belongs on product pages, not on your "about us" page. FAQPage schema belongs on pages with FAQ sections, not on pages without questions. The wrong schema on the wrong page gives AI confusing signals and can actually hurt rather than help your visibility.
Inconsistency between schema and content. If your schema lists a price of £199 but your page shows £249, AI detects that inconsistency. If your schema names five services but your page describes three, that is a negative signal. Schema must be an exact machine-readable equivalent of your visible content. No exaggerations, no claims not supported on the page.
Outdated schema. Schema markup that does not get updated when content changes is a common problem. Opening hours that change around bank holidays, prices that get adjusted, team members who leave. If your schema contains outdated information, AI gives outdated answers about your business. Build schema updates into your content update process. Set a quarterly reminder to review all your schema markup.
Always validate
No validation. 49% of websites with schema markup have errors. Syntax errors, missing required fields, wrong data types. A schema with errors gets ignored by AI. That is the same as having no schema. Validate every page with the Rich Results Test after implementation and after every change. It takes five minutes and prevents your effort from being wasted. Learn more about what AI reads on your site in what AI reads on your website.
Frequently asked questions
Which schema types give the best results?
FAQPage, Article, Organization/LocalBusiness and Product have the most impact on AI visibility. The combination of multiple types per page (3+) gives 13% higher chance of AI citation. Start with Organization on your homepage and Article + Person on your blog content. Add FAQPage to pages with FAQ sections.
Does schema markup really work for AI visibility?
Yes. 71% of pages cited by ChatGPT use schema markup. Pages with structured data are cited 3.2x more often. 65% of pages in Google AI Mode have schema. It is one of the strongest and most measurable signals you can give AI. The numbers are unambiguous.
Should I use JSON-LD, Microdata or RDFa?
JSON-LD. It is Google's recommended format, sits separately from your HTML structure and is processed most effectively by all AI crawlers. Microdata and RDFa are technically valid but less broadly supported and harder to implement and maintain. Always choose JSON-LD.
How many schema types can I use per page?
There is no limit, but each schema type must be relevant to the page. A blog page with Article + Person + Organization + FAQPage is logical. A product page with Product + Organization + AggregateRating is logical. Do not add irrelevant schema types just to have more schema. Quality and relevance count, not quantity.
How do I test if my schema markup is correct?
Use Google's Rich Results Test to validate your schema. Check for errors, warnings and missing required fields. Test after every implementation and after every change. 49% of websites with schema have errors. Validation is not an optional step. It takes five minutes and ensures your schema is actually being read by AI.
Do I need a developer to implement structured data?
Not necessarily. Most CMS platforms including WordPress have plugins (like Yoast SEO or Rank Math) that can generate JSON-LD automatically. For more complex implementations or custom websites, a developer may be helpful. The key is to validate the output regardless of how you implement it, and to keep it up to date as your business information changes.
Is your structured data working for AI?
VestVale automatically monitors whether ChatGPT, Gemini, Claude and Google AI cite your business. Find out if your schema markup is driving AI visibility.
From £19.95/mo ex. VAT. Cancel monthly.