Skip to main content
Measurement · Citation & Visibility Measurement

​
Definition

Unprompted recommendation rate is a measure of how often an AI model names a brand in answer to an open category question, without the buyer naming that brand and without the model searching the live web. It counts recommendations the model volunteers from its own knowledge, not responses to questions where the brand was already supplied. There is a difference between a model that can evaluate a brand and a model that thinks of it. Asked whether a specific brand suits a specific need, a model will assess whatever brand it is handed. Asked an open category question, it names only the brands it holds strongly enough to volunteer, and open category questions are where most buying conversations begin. Unprompted recommendation rate isolates that gap by testing without retrieval, so the result reflects what the model has internalized rather than what it can look up. It is the strictest of the three measures and the slowest to move, because changing it requires shifting what the model believes rather than what it can find.

Parametric presence

Parametric knowledge

LLM brand recall

Brand disambiguation

Share of model

​
Relevant Plate Lunch Collective Services

Entity SEO Context Map AI Search Visibility Assessment