Public methodology
Measured, versioned, and limited on purpose.
The platform separates deterministic site readiness from observed AI visibility and the confidence of those observations.
AEO readiness
A versioned 100-point rubric tests direct answers, modular structure, sourced statistics, citations, authorship, appropriate structured data, entity connections, bot access, static extractability, architecture, and entity consistency. Only tested points enter the score.
Observed AI visibility
Visibility is based only on explicitly labeled official API observations. A provider failure is unavailable, never a zero. Consumer-interface rankings are not claimed.
Confidence
Confidence reflects sample size, repetitions, provider failures, coverage, age, and citation-domain disagreement. It is a transparent band, not false statistical certainty.
Current limits
The anonymous preview uses a shallow, robots-aware sample of up to ten public HTTPS pages. It does not prove future citations, rankings, traffic, or revenue.