Discoverability is the outcome layer.

AI discovery falls into two buckets — the technical discovery and the context discovery. The first decides whether a model can read you. The second decides whether a model recommends you.

A site can be perfectly structured and never get cited. A site can have great content and never get crawled. Discoverability requires both.

Technical discovery is the underlying machinery of a site. The list gets long — meta tags, image alt tags, JSON-LD structured data, clean HTML hierarchy, sitemaps, robots.txt that allows crawlers, schema markup, internal linking, and on. Each one is a signal a model uses to decide whether your site is legible.

Without these, you struggle against sites that have them. Models prefer reading structured content the same way a person would prefer a book over graffiti scattered across a city block.

The second step is the outcome. Structure makes a site eligible. Content is what triggers a citation. The strongest pages don’t just answer the question that landed first — they give the model the context it needs to compose answers across several related questions.

When a person asks an AI a question, the model chooses between two kinds of pages: ones that contain the literal answer, and ones that contain the context to compose an answer. The second kind earns more citations, because it answers more conversations.

Both pages contain content. Example B leverages context — its single page becomes the source AI cites across multiple related questions, not just the one query it was written for.

Discoverability is one of three foundations. Structure, visibility, and discoverability come together in the AI Citation Framework — a step-by-step playbook for earning AI recommendations from zero authority.