State of AI Citations — methodology
The State of AI Citations
One question keeps coming up in every conversation about AI search: how often do brands actually get cited — and does fixing it move anything? This report will answer both from RankEcho's own measurements: per-industry brand citation rates from completed audits, joined with fix-to-proof outcomes from re-tested prompts. Until both datasets clear their publishing thresholds, the methodology below is the page.
Citation dataset threshold
20
Minimum completed audits before aggregate rates publish; 5 per industry row
Proof dataset thresholds
25 / 10
Minimum measured prompts and reportable wins before outcome figures publish
completed audits21 / 20
measured prompts5 / 25
reportable wins1 / 10
Collecting. No synthetic numbers, no projections, no placeholder figures — this page publishes nothing until the real data clears both thresholds. Progress is shown only as collection context, never as a marketing figure.
What will publish here
- The overall average brand citation rate across completed audits, with the sample size beside it.
- A per-industry table — citation rate and off-site share — for every industry past its own sample threshold; the rest stay marked collecting.
- Fix-to-proof movement: reportable wins, the win rate over measured prompts, median days from shipping a fix to the first observed citation, and the retrieval-versus-training pathway split.
Cite this report: RankEcho, "State of AI Citations: brand citation rates and fix-to-proof outcomes across AI search engines," Nexus Decision Systems LLC, 2026-06-19. https://rankecho.io/state-of-ai-citations
Run a free audit — every audit matures this dataset
Aggregate counts, rates, and medians only. This page never publishes domains, prompts, or customer-level results, and movement is reported as correlation with controls, not a guarantee.