Skip to main content

Glossary

GEO (Generative Engine Optimization)​

Optimizing content and brand presence for AI models (generative search engines) so AI mentions your brand and products more often and more accurately in its answers.

Share of Voice (SoV)​

How often AI mentions your brand in answers to your topic-map questions — the share of your mentions across all queries (measured on generic questions, not branded ones).

Share of Shelf​

The product equivalent of SoV: your products' share among the brands AI recommends in shopping queries for a given category (the "shelf" in the AI answer).

Topic map​

A set of questions real AI users ask, used to measure visibility. A map is saved explicitly; an active map snapshot is a point-in-time checkpoint.

Attribution — 3 levels of proof​

Order revenue attributed to the AI channel is reported at three levels, never summed:

  • Hard — proven (e.g. a closed path from the pixel / AI UTM).
  • Self-reported — the customer named AI as the source ("how did you hear about us" survey).
  • Probable — signals pointing to AI without hard proof.

Persona​

A defined audience profile (decision criteria, pain points) used to contextualize questions and measure visibility "through the eyes" of a specific segment.

Ground truth (reference facts)​

A set of true facts about your brand/products (prices, features, history, certifications) against which we detect AI hallucinations in the Brand Safety module.

Refresh score​

A content "freshness" indicator. When it drops, the Content Refresh Engine flags the content for a refresh.

Platform-voice model vs internal model​

  • Platform voice — the AI model we query a platform with, impersonating a user (reflects what the consumer actually sees).
  • Internal model — the model used for platform tasks (generation, scoring, analysis). These are two separate layers, chosen independently.

A check option where the AI model uses current results from the web — so the answer reflects the current market, not just the model's knowledge.