Detailed Definition
Unaided Prompt Mention Rate tests true baseline AI recall. By asking open-ended questions like 'Name the top enterprise cloud providers' rather than 'Tell me about Microsoft Azure', evaluators measure whether the model spontaneously recommends the entity.
Formula / Calculation Standard
Unaided Mention Rate (%) = (Unaided Prompts with Entity Mention / Total Unaided Prompts) × 100Why It Matters for Brands
Aided prompts ('What is Apple?') only measure knowledge retrieval. Unaided prompts measure top-of-mind AI recommendation, which dictates actual customer acquisition.
How Fameproof Measures It
Fameproof uses standardized Version 2 prompt packs containing 24 unaided queries across broad, regional, and industry-specific contexts.
Read Full Evaluation Methodology →Example Entities Measured with this Metric
Related Terms
Generative Engine Optimization (GEO)The methodology of optimizing brand presence, canonical entity representation, and authority to increase mentions in generative AI answers.LLM Share of Voice (SoV)The percentage of AI-generated responses within a category that explicitly cite or recommend a specific brand compared to competitors.Repetition ConsistencyThe variance or stability of AI model responses when identical prompts are submitted across multiple independent API calls.
Related Research Articles
Generative Engine Optimization (GEO): The Complete Guide to AI Model Visibility in 2026Discover how Generative Engine Optimization (GEO) differs from traditional SEO, how ChatGPT, Claude, and Gemini evaluate entities, and how to measure unaided AI visibility.Why ChatGPT, Claude, and Gemini Recommend Different Brands (And How to Audit Them)Why does ChatGPT list OpenAI and Apple first while Claude emphasizes Anthropic partners or open-source leaders? Learn the architectural and pre-training reasons behind model variance.