Research
Twelve findings from independent, published research that changed how we built AIEO. Every stat cites its primary source and grades the evidence.
Methodology
The findings below are from independent, published research — not AIEO's own dataset. Sources include Princeton GEO, Ahrefs, Profound, Semrush, Pew, SoCi, Shopify, SE Ranking, Rutgers/Wharton, and others. Each stat links to its primary source and carries an evidence grade per the rubric at aieo.ai/methodology. AIEO's own measurement platform runs n≥30 samples per prompt per engine per week with Wilson 95% CIs; first-party results will publish once the first closed-alpha cohort completes a full sampling window.
12 findings · independent published research
Domain overlap between ChatGPT and Perplexity citations
Qwairy / Ahrefs, 2025
AIO citations also in Google top-10 organic
Ahrefs AIO study
ChatGPT citations from URLs ranked position 21+
Semrush, 5M URLs
Spearman correlation: YouTube mentions × AI citation rate
Ahrefs, 75K brands
Local businesses recommended by ChatGPT
SoCi, 2026
ChatGPT local citations from ThreeBestRated
BrightLocal / LocalFalcon
Statistics-tactic uplift on simulated engine
Princeton GEO (Aggarwal et al., KDD 2024)
Traffic loss from blocking AI crawlers — without reducing citations
Rutgers / Wharton, Dec 2025
Professional citations going to individuals (not companies)
Averi
YMYL legal queries that trigger AI Overviews
SE Ranking
AI shopping inclusion lift from complete Product schema
Shopify, Q4 2025
ChatGPT Reddit share collapse in 5 weeks (Sept 2025)
Semrush, 13-week study
“Eleven percent domain overlap between ChatGPT and Perplexity. That single number is the whole case for multi-engine sampling.”
AIEO methodology · Finding 01
Citation policy
Every stat above labels its source. This report presents third-party published findings, not AIEO's own measurement data. If you reuse these figures, cite the original primary source, not AIEO. Evidence grades follow the rubric at aieo.ai/methodology.
AIEO · AI visibility platform
Wilson 95% CIs. Evidence-graded recommendations. 9 engines tracked weekly.