How ChatGPT chooses which businesses to recommend
A practitioner's view of the signals ChatGPT appears to weigh when recommending local businesses.
This is our working model of ChatGPT recommendation behavior, drawn from hands-on client work and publicly documented model behavior. It is an explainer, not a published study: ChatGPT tends to favor businesses with strong entity definition, dense third-party citations, recent activity signals, and clear category attribution.
Signal 1 — Entity recognition
ChatGPT must first know the business exists as a discrete entity. Without a Wikipedia article, Wikidata entry, or strong directory presence, the model often refuses to name a specific business at all.
Signal 2 — Citation density
The more authoritative sources reference a business (and the more recent those references are), the higher the recommendation probability — even in models with browsing disabled.
Signal 3 — Category clarity
Ambiguous business descriptions ('marketing services') under-perform precise ones ('AI Search agency for medical clinics in El Paso'). Precision wins recommendations.
Yes when browsing is enabled; otherwise it draws from training. AISO covers both surfaces.
Want this engineered into your business?
A free AI Visibility Audit shows where your business currently stands across ChatGPT, Gemini, and Perplexity.
Request Free Audit