Skip to main content

AI Visibility Analytics Options

The tables below list the metrics and dimensions available for AI Visibility (GenAI Lens) analytics requests, where provider is llm. Learn how to use them by reading the Analyzing Prompts guide.

  • Metrics are the values calculated by a metrics analysis (used on its own, or nested inside a date_histogram or top_terms analysis). Use the API name in the metrics array.
  • Dimensions are the fields a top_terms analysis can group by. Use the API name in the dimension field.

Metrics

NameAPI nameDescription
Average FrequencyfrequencyThe average number of mentions per document over the selected period - "how much are people talking about me?".
MentionsmentionsThe total number of mentions of the entity across all results within the selected filters and date range.
ProminenceprominenceThe relative position of an entity within a result, giving higher weight to mentions that appear earlier (for example, in the first paragraph rather than later sections).
SentimentsentimentThe overall tone of coverage on a 0-100 scale, where 0 is negative, 50 is neutral, and 100 is positive.
VisibilityvisibilityA combined score giving a single, balanced measure of an entity's overall presence. It combines how widely the entity is seen (mention rate), how often it appears (frequency), and how central it is to the content (prominence). Can be fractional.

Dimensions

NameAPI nameDescription
ConceptsconceptsKey concepts identified in the AI answers.
JournalistsjournalistsJournalists or authors associated with cited sources.
Key phraseskeyphrasesKey phrases extracted from the AI answers.
LinkslinksLinks cited in the AI answers.
LLMsllmsThe AI assistant or answer engine that produced the answer (for example ChatGPT, Gemini, Perplexity).
CountrycountryThe country associated with the matching answers.
OrganizationsorganizationsOrganizations mentioned in the AI answers.
PeoplepeoplePeople mentioned in the AI answers.
ProductsproductsProducts mentioned in the AI answers.
SourcessourcesSources cited in the AI answers.
Source typessource_typesThe type of cited source (for example news, blog, forum).
SentimentsentimentGroups results by sentiment (for example positive, neutral, negative).