What an AI Sitemap Is
An AI sitemap is a machine-readable index of a website's content, designed specifically for AI systems rather than for search engine crawlers. Where a traditional sitemap.xml is a flat list of URLs with optional metadata, an AI sitemap (typically at /ai/ai-sitemap.json or /ai-sitemap.json) provides structured, typed content data organized by content type, topic cluster, or priority ... making it far more useful for retrieval systems that need to understand what a site contains before deciding what to index.
How It Differs from sitemap.xml
The traditional sitemap.xml is designed for search engine crawlers that process URLs one at a time. It tells crawlers which pages exist and when they were last modified. An AI sitemap goes further. It groups content by type (articles, guides, glossary terms, directory listings), provides topic labels, includes entity maps, links to related content, and may include AI-readable summaries. The result is a structured content overview that an AI system can use to build a mental model of your site without crawling every page.
What an AI Sitemap Contains
A well-structured AI sitemap typically includes the site's identity information, a list of content categories with descriptions, a typed index of content items (each with URL, title, description, content type, publish date, and topic tags), links to related machine-readable resources (llm.json, entity-map.json, content-index.json), and priority signals. Some implementations also include AI-readable summaries for each page, further reducing the work an AI system needs to do to understand the content.
Why AI Sitemaps Matter
As AI-powered search and retrieval systems evolve, they increasingly need to understand not just that a page exists, but what type of content it is, what topic it covers, and how it relates to other content on the same site. An AI sitemap provides this context upfront, before the AI system needs to crawl and parse individual pages. Sites that provide this structured overview are more likely to be indexed completely, understood accurately, and surfaced appropriately in AI-powered search results.