Skip to main content

Entities, Mentions, and Relationships

How Entities, Mentions, and Relationships work inside our data model

Brandon Brown avatar
Written by Brandon Brown
Updated over 3 weeks ago

How It Works

The foundation of the Search Party data model is a simple chain:

Entity → Mention → Relationship

  • Entities are the “nouns” of the system: brands, products, people, publications, or concepts. Example: Nike, Slack Canvas, Sustainability.

  • Mentions are every time an entity appears in an AI response, along with the context: was it a top recommendation, a passing citation, or a direct comparison?

  • Relationships are built by combining mentions and observations into connections between entities. Example: Nike HAS_AFFINITY_WITH Sustainability or Slack Canvas COMPETES_WITH Notion.

This chain takes scattered AI outputs and turns them into a structured, living map of how your brand and ecosystem are perceived inside large language models.

Key Benefits

  • Consistency: Entities are global and standardized across all customers, ensuring accuracy

  • Context: Mentions capture not just that you appeared, but how you appeared

  • Insight: Relationships reveal meaningful connections—affinity, competition, validation, or contradiction—scored over time

Common Use Cases

  • Marketing: See if a campaign is increasing brand affinity with a narrative like “sustainability”

  • Product marketing: Track if a new feature is showing up as a competitor to another tool

  • Communications: Measure whether media outlets are validating or contradicting your brand messages

Did this answer your question?