Original research on how AI engines decide what to recommend. Every number is ours. Every methodology is documented.
We asked AI to recommend Australian real estate brands with web search and without. 5 cities. 21 active brand-city pairs. Ray White entrenched in 5 of 5. LJ Hooker memory-favored in 4 of 5. Two layers, two different levers.
4 AI engines. 20 Australian suburbs. 746 distinct names. Engines agreed on 13. We audited 5 winners against 5 matched-suburb controls on 15 entity signals. The difference was not the website — it was the review pile.
~6,200 AI citations across 12 markets and 5 industries. AI engines run two separate systems for local businesses — one for recommendations, one for explanations.
20 prompts. 3 engines. 19 of ~35 licensed clinics named — 54% of the regulated market. Birth rates didn't predict who AI recommends.
20 prompts. 3 engines. 42 agencies named. Only 11 on all three — and the most-cited agency overall is invisible on ChatGPT.
324 queries. 3 engines. 6 cities. 800+ agencies named. Only 12 on all three — and not one was a franchise.
18 buyer prompts. 3 AI engines. 210 agencies named. Only 4 on all three. 94% boutique dominance — highest of any city tested.
18 buyer prompts. 3 AI engines. 351 citations. Only 5 brands on all three. One engine confused Melbourne AU with Melbourne FL.
18 buyer prompts. 3 AI engines. 254 agencies named. Only 3 appeared on all three. See who AI recommends — and who's invisible.
The full story behind BaliPropertyRules.com — from zero authority to AI-cited source. Every number, every decision.
We audited 200+ AI responses to understand citation patterns. Here's what actually drives AI to cite a source.
SEO agencies optimize for rankings. AI citation requires a fundamentally different approach — different incentives, different metrics, different skills.
Request your free AI Visibility Report. We'll show you exactly where AI mentions your competitors — and where it's silent about you.
Request Your Free Report