How often brands are cited inside AI answers, by industry, and what fixing it actually moved — measured, not estimated. Data through 2026-09-06.
| Industry | Audits | Avg citation rate | Off-site share |
|---|---|---|---|
| B2B SaaS | 44 | 22% | 96% |
| Other | 14 | 34% | 92% |
Industries below the 5-audit segment threshold remain in collecting and are not shown.
Of 27 prompts re-tested after a fix shipped, 12 cleared the reportable bar — a citation observed under at least medium sampling, beating that prompt's own baseline. The pathway split over those wins: 5 fast retrieval-layer wins, 5 slower training-layer wins, 2 in between. Movement is reported as correlation with controls — engines change for many reasons, and this report never claims otherwise.
Every figure comes from completed RankEcho audits and tracked proof loops: buyer-intent prompt batteries run across multiple AI engines with citation parsing, aggregated anonymously by the industry the audit classified. Free single-sample audits may reuse an engine answer observed within six hours, as disclosed on the methodology page; paid sampling always queries live. Industry rows publish at 5+ audits; outcome figures published once 25+ prompts were measured with 10+ reportable wins.
Aggregate counts, rates, and medians only. This page never publishes domains, prompts, or customer-level results, and nothing here is a guarantee of any individual outcome.