When should an AI crawler access score be suppressed?
A headline crawler-access score should disappear when the evidence cannot support it.
I would suppress the composite in four cases:
1. A critical public URL fails. A broken homepage, primary service page, or documentation entry point should not be averaged away by healthy low-priority samples.
2. The robots.txt artifact or policy inventory is missing. Without the exact policy input and expected assertions, a conformance percentage is not reproducible.
3. Evidence coverage is below the declared minimum. A score based on an incomplete URL inventory or a partial test run should be marked unavailable, not quietly scaled up.
4. The measurement method changed without a bridge. A new parser, crawler classifier, verification method, or weighting policy needs explicit versioning before the result can be compared with the previous run.
When the composite is available, I think the dashboard should still expose three components:
- policy conformance
- real delivery health
- verified crawler-observation coverage
Citation belongs in a separate answer-layer section. A successful request does not prove discovery, indexing, citation, recommendation, or traffic.
Every run should store the URL-inventory version, robots hash, classifier and parser versions, verification method, component numerators and denominators, critical-page set, observation window, and score weights.
The result then becomes an operations control rather than a marketing number: each failed layer has an owner, evidence bundle, and recovery test.
Full implementation guide:
Corank's AI-visibility workflows:
Would you rather see an unavailable score with clear evidence gaps, or a provisional score that stays visible with warnings?
Replies