Sharing Data With AI

The Agentic AI Assistant includes a data sharing option that helps organizations balance AI functionality with their data governance requirements. When data sharing is enabled, the assistant can provide the full experience for supported use cases. When it is disabled, the assistant remains available in a more restricted mode. By default, data sharing is turned off.

When data sharing is disabled, GoodData limits or hides AI capabilities that need to send computed analytical results or retrieved document content to the configured LLM. Some skills remain fully available, some can prepare content but cannot execute data analysis, and some are hidden from the assistant.

In general:

• Knowledge Search, Visualization Summary, and Dashboard Summary are hidden.
• Forecasting, Anomaly Detection, and Clustering may still be visible, but the assistant cannot run their data execution tools.
• Key Driver Analysis and What-If Analysis can prepare the analysis or scenario setup, but cannot execute the analysis or interpret results.
• Alert Creation remains available, but alert recipients are limited to the current user.
• Metric Creation, Data Visualization, and Scheduled Export remain available. If a workflow requires natural language interpretation of computed results, that interpretation requires data sharing.

Configure Data Sharing for the Agentic Assistant

Organization admins and solution architects can control whether the agentic assistant is allowed to share data with the configured large language model (LLM).

You can manage this setting at these levels:

• organization
• workspace

This lets you apply a broader default for the whole organization or configure specific workspaces differently.

Default Behavior

Data sharing is opt-in and is turned off by default. This default helps organizations that need tighter control over how data is handled when using AI features, typically:

• enterprises with strict compliance requirements
• regulated industries such as healthcare, finance, and government
• teams that need to balance AI capabilities with privacy and governance requirements

If you use your own LLM setup, this setting can also help you control which models are allowed to receive data.

Auditability

Changes to the data sharing configuration are tracked through platform audit logging so administrators can review configuration changes when needed.