Measurement methodology
AI Recommendation Rate Methodology
RankingLocal separates estimated readiness from live measurement so local businesses can see what was inferred, what was actually tested, and what changed after fixes.
AI Recommendation Readiness Estimate
This is a diagnostic estimate based on crawlability, schema, local evidence, reviews, citations, page structure, and REAL Method signals. It is useful for finding likely blockers, but it is not a live AI answer measurement.
Live AI Recommendation Test
A meaningful live test would retain the prompt, named service, answer, mention decision, citation when present, and timestamp. It must say which services and questions were actually tested. The free website checks do not perform this test.
Longitudinal Tracking
To compare observations over time, repeat a defined prompt set and report the sample size, failures, and any changes in the test method. A changed answer alone does not prove which website edit caused it.
Current Offer And Measurement Scope
| Path | Current scope |
|---|---|
| AI Visibility Foundation Check | Free limited website estimate. A Perplexity query may run when available; other named services are Not measured here. One query score is not a recommendation rate. |
| Free Report Preview | Free website foundation checks after business-email verification. Results identify completed, skipped, and unavailable checks. |
| SEO services | Current scope and prices appear on the pricing page. Any live AI answer tracking must be defined in a separate written scope with its actual evidence. |
What The Report Can And Cannot Prove
A website-readiness estimate cannot prove an AI recommendation rate. A measured test can only describe its named questions, services, location, and time window; it cannot guarantee causes, future mentions, leads, or market-wide coverage.