Direct answer: Pages get skipped when answers are ambiguous, sources are thin, or entity context is inconsistent.
Machine read
Primary entity
Diagnostic pattern library
Extractable answer
Medium
Citation potential
Medium
Main issue
Ambiguous answer targets and low evidence density
Human read
Operators need a repeatable checklist for why a page is parsed but not selected.
Visibility is not a binary indexed-or-not state. Selection happens deeper in the stack.
What to change
- Rewrite the first screen for one explicit query and one explicit answer.
- Add source-backed claims and datestamped updates.
- Strengthen internal links so entity relationships are obvious.
- Remove layout patterns that hide essential definitions.
Hidden failure mode: Teams optimize metadata while leaving answer clarity unresolved in visible content.
Noise check: Chasing model-specific hacks before fixing information architecture wastes cycles.
The playbook
- Owner: Editorial lead + search engineer
- Effort: One week
- Expected outcome: Higher answer selection and clearer model interpretation.
FAQ
Can a page be indexed but still ignored in AI answers?
Yes. Indexing allows retrieval, but selection depends on confidence and utility signals.
Is schema enough to force inclusion?
No. Schema supports interpretation but does not override weak content.
Sources:
When teams say “we are not visible in AI answers,” they often describe a selection problem, not a crawl problem.