Internal Linking

Internal linking: a signal for both SEO and agents

Internal links connect pages within your site. For traditional SEO, they distribute PageRank and help crawlers discover content. For AI agents, they serve a more important role: they define the navigation graph of your site, helping agents understand which pages are most important, how topics relate, and where to go for more detail.

A well-structured internal linking strategy is one of the lowest-effort, highest-impact improvements you can make.

Why it matters for AI agents

When an agent browses your site — either via a crawler indexing your content or in real time during a user session — it follows links to discover related information. Strong internal linking means:

Discovery: All important pages are reachable from a main entry point
Context: Links with descriptive anchor text signal what the linked page is about
Authority: Pages with many inbound internal links are treated as more authoritative
Depth: Agents can follow topic chains to build a complete picture

Anchor text quality

The anchor text (the clickable text of a link) is one of the most important signals in internal linking. It tells agents and crawlers what the destination page is about.

Good (descriptive):

<a href="/docs/json-ld">how to implement JSON-LD structured data</a>

Avoid (generic):

<a href="/docs/json-ld">click here</a>
<a href="/docs/json-ld">read more</a>
<a href="/docs/json-ld">this page</a>

Use natural, keyword-rich anchor text that accurately describes the destination page.

Site architecture and link depth

Link depth is the number of clicks required to reach a page from the homepage. Key rules:

Important pages should be reachable within 3 clicks from the homepage
Every page in your sitemap should be internally linked from at least one other page
Orphan pages (no inbound internal links) are invisible to crawlers and agents

Ideal architecture for a documentation site:

Homepage
├── /docs (hub)
│   ├── /docs/introduction        (depth 2)
│   ├── /docs/json-ld             (depth 2)
│   └── /docs/llms-txt            (depth 2)
├── /blog (hub)
│   ├── /blog/article-1           (depth 2)
│   └── /blog/article-2           (depth 2)
└── /glossary (hub)
    └── /glossary/schema-org      (depth 2)

Hub pages and topic clusters

A proven SEO architecture for AI-ready sites:

Pillar page (hub): a comprehensive overview of a broad topic (e.g., /docs/structured-data)
Cluster pages: detailed sub-pages on specific aspects (e.g., /docs/json-ld, /docs/schema-org)

The pillar page links to all cluster pages, and each cluster page links back to the pillar. This creates a topic cluster that signals subject-matter authority.

/docs/structure (pillar)
  ↔ /docs/schema-org
  ↔ /docs/json-ld
  ↔ /docs/open-graph
  ↔ /docs/llms-txt
  ↔ /docs/sitemap-xml

Cross-linking between blog, docs, and glossary

Cross-contextual links boost site authority and help agents connect concepts:

Blog articles should link to relevant documentation pages
Documentation pages should reference glossary terms
Glossary terms should link back to related docs and blog posts

Example in a blog article:

To configure your site correctly, start by adding 
[JSON-LD structured data](/docs/json-ld) and publishing an 
[llms.txt file](/docs/llms-txt). See the [GEO glossary term](/glossary/geo)
for a full definition.

Breadcrumbs

Breadcrumbs serve dual purposes: they help users navigate and give crawlers a clear hierarchy signal. Always add Schema.org BreadcrumbList JSON-LD alongside visible breadcrumbs:

{
  "@context": "https://schema.org",
  "@type": "BreadcrumbList",
  "itemListElement": [
    {
      "@type": "ListItem",
      "position": 1,
      "name": "Docs",
      "item": "https://example.com/docs"
    },
    {
      "@type": "ListItem",
      "position": 2,
      "name": "JSON-LD",
      "item": "https://example.com/docs/json-ld"
    }
  ]
}

"Related articles" sections at the bottom of pages are an efficient way to add relevant internal links without disrupting the main content. They also increase session depth and expose agents to more content.

What to avoid

Excessive linking: Linking every occurrence of every keyword dilutes the signal
Identical anchor texts to different pages: confuses crawlers
Circular link loops with no additional context: legitimate but should be complemented by navigational links
Links in hidden elements: links inside display:none or CSS-hidden elements may be ignored