Data Handling - Securing AI in Sisense Intelligence

Last updated: November 1, 2026

Tier Deployment
Launch     Grow    Enterprise Cloud     On-Prem

Sisense Intelligence brings powerful AI capabilities to every layer of the analytics experience. Sisense believes in responsible AI; as such, we are committed to ensuring that these capabilities meet the highest standards for privacy, control, and enterprise-grade data governance. This documentation outlines how Sisense protects your data, manages AI driven workflows and agents, and empowers your organization to use these tools with confidence.

General Principles

  • No model training on your data - Sisense never uses your data to train any models.

  • Sisense minimizes data exposure and transmits only what is essential to the functioning of its AI features:

  • Local architecture first - many Sisense Intelligence features run entirely inside your Sisense environment.

  • Governed - System level controls exist for all Sisense Intelligence features, with select features allowing data model, dashboard or widget-level configuration.

    • Data designers can keep data models, tables, and columns away from AI features with the hide_from_ai tag. Sisense removes tagged entries before data model information is sent to the LLM. See Creating Data Dictionaries.

Sisense Intelligence Features and Cloud Dependency

Sisense features fall into two categories - Cloud-linked and Local:

  • Cloud-linked: Require access to Sisense cloud services including Large Language Models (LLMs) and a Vector database. These features include all of our Generative AI features and their underlying REST APIs.

    • Assistant

    • Narrative

    • Semantic Enrichment

    • Smart Value Matching

    • REST APIs include:

      • /ai/chats/{chatId}
      • /ai/recommendations/query/{sourceId}/{numOfRecommendations}
      • /ai/describe/query
      • /ai/nlg/queryResult
      • /ai/nlq/query/{sourceId}
      • /ai/recommendations/{datamodelId}/table-relations

  • Local: Run entirely within the customer environment without cloud dependencies.

    • Explanation

    • Forecast

    • Trend

    • Exploration path

    • Simply Ask (NLQ)*

    • Knowledge Graph**

*Simply Ask will be deprecated in 2026. We recommend using the Assistant feature instead, which provides enhanced conversational capabilities with better contextual understanding.

** The Knowledge Graph is a native Sisense service used to power Sisense Intelligence features such as Explanation and Simply Ask. Each deployment maintains its own instance, and tenant data is isolated.