← Back to diagnostics

← Back to AI Search Diagnostics

What does indexed but not retrieved mean?

A URL can appear as indexed in Google Search Console while remaining absent from AI-mediated answers (ChatGPT browsing, Perplexity citations, AI Overviews). Indexing confirms crawl storage; retrieval is a separate judgment about whether a fragment is worth grounding for a specific query.

What the System Is Doing

Your pages appear in Google Search Console as indexed, but generative AI systems do not retrieve them when answering relevant queries.

Why the Symptom Appears

Common causes include:

  • Fragment extractability: Content is buried in long narrative or layout-dependent sections
  • Entity ambiguity: The page does not disambiguate which entity it authoritatively describes
  • Trust signals: Missing author, dates, or source attribution for verifiable claims
  • Rendering parity: Critical text exists only after client-side hydration (silent hydration SEO)
  • Intent mismatch: The page ranks for keywords but does not answer the generative query shape

Diagnostic Workflow

  1. Confirm indexing status in Search Console (external fact)
  2. Test AI answer coverage for target queries manually or via monitoring
  3. Audit fragment-level extractability and JSON-LD parity
  4. Check cannibalization—whether another URL on your site owns the same intent cluster
  5. Apply prechunking or entity fixes; re-test after static dateModified updates only when content changed

Remediation Path

To enable retrieval:

  1. Restructure content for atomic retrieval
  2. Implement prechunking framework
  3. Ensure segments are independently scorable
  4. See GEO fundamentals for mechanics

This is a failure state within AI Optimization, where systems fail to retrieve or select a source during AI-driven search and answer generation.

Sources