JSON-LD is a structured data format that lets a webpage explicitly tell search engines and AI systems what it is, an organization, a product, an article, a set of FAQs, instead of leaving them to infer that from unstructured text. It’s written as a small block of code, based on the schema.org vocabulary, and placed in a page’s HTML where it’s invisible to human visitors but fully readable by machines.

Why AI engines rely on structured data

Unstructured text requires an AI model to infer meaning. Is this paragraph describing a company, a product, or a person? Structured data removes the guesswork by labeling that meaning explicitly. This doesn’t guarantee a citation, but it materially reduces the ambiguity an AI system has to resolve on its own, which is one of the concrete, controllable factors in what influences AI citation.

The schema types that matter most for a small business site

Organization

Establishes the basic facts about a business: name, logo, URL, and, where relevant, social profiles and contact details. This is typically the first schema type worth adding site-wide.

WebSite

Describes the site itself, distinct from the organization behind it, and can include a defined search action.

Article / TechArticle

Applied to blog content, this schema labels headline, author, publish date and body content explicitly, directly supporting the E-E-A-T signals covered in How to Improve Your GEO Citability Score.

FAQPage

Applied to a genuine Q&A section, this labels each question and its answer as a discrete pair. It’s one of the most directly citable schema types, since it mirrors the exact question-answer format of an AI chat response. Full detail in How to Write FAQ Content That Gets Cited by AI.

SoftwareApplication / Product

For a product or plugin page, this schema can label the category, operating system or platform, and a featureList of concrete capabilities, much more specific than a general Organization schema.

A minimal example

A simplified Organization schema block looks like this:

{
  "@context": "https://schema.org",
  "@type": "Organization",
  "name": "Example Brand",
  "url": "https://example.com",
  "logo": "https://example.com/logo.png"
}

This same pattern, a @context, a @type, and a small set of relevant properties, applies to every schema type. Only the properties change.

How to add JSON-LD on a WordPress + Divi site

There are three common approaches:

1. An SEO plugin’s built-in schema tool (Yoast, Rank Math, etc.), usually the simplest path for Organization, WebSite and Article schema, since these plugins generate valid JSON-LD from existing page fields
2. A dedicated schema plugin, which offers more granular control for less common types
3. A manual code module, for schema types not covered by an existing plugin (FAQPage generated from a non-block FAQ layout, for example), where JSON-LD can be pasted directly into a Divi Code module

How to verify it’s working

After adding schema, validate it with Google’s Rich Results Test or the Schema.org validator before considering it done. A JSON-LD block with a syntax error is effectively invisible, since most systems will silently ignore malformed structured data rather than partially reading it.

FAQ

Does adding schema markup change how a page looks to visitors? No, JSON-LD is invisible in the rendered page. It exists purely in the page’s code for machines to read.

Is schema markup required for AI citation, or just helpful? It’s not strictly required. AI systems can and do cite pages with no schema at all, but it removes ambiguity and is one of the more directly controllable levers available, unlike, for example, third-party backlinks.

How often does schema need to be updated? Whenever the underlying content changes materially: a new FAQ question, an updated product feature list, a new author. Static schema that no longer matches the visible page content can create inconsistency that undermines trust signals rather than strengthening them.