Glossary/technical

JSON-LD

Last updated March 22, 2026

Definition

Quick answer
JSON-LD (JavaScript Object Notation for Linked Data) is the recommended format for embedding structured data on web pages. In AEO, it provides the machine-readable semantic layer that helps AI engines understand content type, authorship, entity identity, and page relationships.
Full definition

What is JSON-LD?

JSON-LD is a method of encoding structured data using JavaScript Object Notation, designed to be both human-readable and machine-processable. It is the format recommended by Google for implementing Schema.org markup, and it has become the de facto standard for providing machine-readable context to both search engines and AI engines.

In the AEO context, JSON-LD serves as the bridge between your human-facing content and the machine-readable signals that AI engines need to process that content accurately. A JSON-LD block embedded in a page tells AI crawlers: this is a Product page (not a blog post), it was published on this date, it was authored by this person, it belongs to this organisation, and it relates to these topics. Without this explicit context, AI engines must infer all of these attributes from the page text — a process that is inherently less reliable.

JSON-LD is implemented as a <script type="application/ld+json"> block, typically placed in the page's <head> or <body>. This implementation approach is a key advantage: because the structured data is contained in a separate script block rather than interspersed with the visible HTML (as with microdata or RDFa), it is easier to implement, maintain, debug, and update without affecting page layout or content.

For AEO, the most valuable JSON-LD implementations include: Organization (defining brand identity and properties), Product (providing machine-readable product attributes), Article and BlogPosting (establishing content authority and authorship), FAQ (structuring question-answer content for direct AI extraction), DefinedTerm (marking glossary entries as authoritative definitions), HowTo (structuring procedural content), and BreadcrumbList (mapping site hierarchy).

JSON-LD also powers the llm-profile.json file — a standalone JSON-LD document placed at .well-known/llm-profile.json that provides a comprehensive, machine-readable brand profile specifically for AI consumption. This extends JSON-LD beyond individual page markup into site-wide entity definition.

The quality of JSON-LD implementation matters as much as its presence. Incorrect types, missing required properties, or data that contradicts the visible page content can confuse AI engines rather than help them. JSON-LD should be validated using tools like Google's Rich Results Test and the Schema.org validator, and it should accurately reflect the actual content and claims on each page.

Context

Why it matters

JSON-LD is the most efficient and maintainable way to give AI engines the structured context they need to understand your content. It is the format that powers Schema Markup, llm-profile.json, and other machine-readable assets that directly influence how AI engines parse, categorise, and cite your content. Getting JSON-LD right is a high-leverage Technical AEO investment.

Examples

Real-world examples

  • 1

    Implementing Organization JSON-LD on the homepage with name, description, url, logo, and sameAs links to authoritative profiles, strengthening entity recognition across AI engines

  • 2

    Adding FAQ JSON-LD to a product comparison page, enabling Perplexity and AI Overviews to extract question-answer pairs directly into their responses

  • 3

    Creating a comprehensive llm-profile.json using JSON-LD format with Schema.org vocabulary to provide AI crawlers with a structured brand definition

JSON-LD FAQ

Frequently asked questions about JSON-LD

Related terms
Get started

Start with the pages and proof that AI can actually use

Run the free audit to see what blocks AI from citing your site. Use the trial when you need ongoing monitoring, attribution, prompt discovery, and team workflows after the first fixes are live.