In short
Schema.org is a shared vocabulary for describing things on the web: organisations, local businesses, products, services, articles, events, people and many more. It was launched in 2011 by Google, Microsoft (Bing) and Yahoo, with Yandex joining soon after. Structured data using that vocabulary is usually added to a page as JSON-LD, a block of code in the page that visitors do not see. Google documents which schema.org types it supports and recommends JSON-LD as the format.
What it looks like
A JSON-LD block is a short list of labelled facts. For a local business it might state the type (for example, a plumber), the business name, address, phone number, opening hours, website and links to official social profiles. For an article it might state the headline, author, publisher and publication date.
Nothing about the visible page changes. The markup sits alongside the content and repeats the important facts in a format software can read without guessing.
Useful types for a small business
| Organization or LocalBusiness | Who you are, where you are, how to contact you, and links to your official profiles. |
|---|---|
| Service | What you offer and the areas you serve. |
| Product and Offer | Items you sell, with price and availability. |
| Article | Blog posts and guides, with author and date. |
| FAQPage | Questions and answers on a page. |
| BreadcrumbList | Where a page sits in your site structure. |
What Google does with it
Google uses structured data to understand pages and to decide whether they are eligible for rich results: enhanced listings such as review stars, product prices, event dates or breadcrumbs. Google’s documentation lists the supported types and the properties each one needs. It also says plainly that valid markup does not guarantee a rich result.
Support changes over time. In 2023, for example, Google restricted FAQ rich results to a small set of well-known government and health websites. FAQ markup is still valid schema.org and still describes your page; it just rarely produces the expanded listing it once did.
Google provides a Rich Results Test, and schema.org runs a separate Schema Markup Validator. Use both after any change. Google Search Console also reports structured data errors it finds while crawling your site.
Why it matters for AI search
AI assistants and AI search features draw on web pages and search indexes to build their answers. Structured data makes your core facts explicit and consistent: your name, what you do, where you work and how to contact you. That is helpful to any system trying to describe your business accurately.
Be realistic about the limits. AI providers do not publish exactly how much weight, if any, their systems give to structured data, and schema markup does not guarantee that an assistant will mention or cite you. Treat it as one part of good foundations, alongside clear page content, consistent details across directories and independent mentions.
Getting it right
The markup must match what is visible on the page. Google’s guidelines treat structured data that describes content users cannot see, or that is misleading, as a policy problem that can cost you rich results. Keep it accurate as prices, hours and services change, which is easier if your website generates it from the same data as the page itself.
Start with Organization or LocalBusiness on your home page, then add types for your service pages and articles. More markup is only better when it is accurate.
When this matters
Relevant to any business that relies on search and AI assistants to be found. RankOne tracks how AI assistants describe your brand over time, and New Era AI’s AI strategy work includes checking your structured data as part of an AI search review.
Questions
Does schema markup improve rankings?
Google has said structured data is not in itself a ranking boost. It helps Google understand the page and can make it eligible for rich results, which may change how your listing looks.
Will schema markup get us cited by ChatGPT or Gemini?
There is no guarantee. It makes your key facts clearer to machines, which helps, but AI providers do not publish how they weigh it. Treat it as a foundation.
Which format should we use?
JSON-LD. Google recommends it, and it keeps the markup separate from the visible page, which makes it easier to maintain.
How do we check our markup?
Use Google’s Rich Results Test for Google features and the Schema Markup Validator at schema.org for general validity.