Effective AI Dashboard Creation
The chat history is associated with the current dashboard editing session.
Click to reset the conversation context. Content that is already added to the dashboard remains unchanged.
Clear the chat history to start a new conversation without removing previous dashboard changes.
Remember these things:
- Be specific in each request. Specify chart types, measures, time periods, and filters by name to improve accuracy.
- Use names from the data model. If a measure is named "Net Sales," use "Net Sales" instead of a generic term such as "Revenue." The AI attempts to match both terms.
- Start with a simple request, and then refine the dashboard with follow-up requests.
- Specify one change in each message to improve editing accuracy. Requests that contain many changes can produce less predictable results.
- Verify that data is enabled for AI access. The AI uses measures, attributes, and saved expressions that are enabled in workspace GenAI settings.
- The AI builds content from available and enabled data. The AI does not create data or business logic.
- If a matching report already exists in the catalog, the AI reuses the report instead of creating a duplicate.
- The layout uses a grid system. The AI prevents content overlap and adjusts spacing after processing is complete.
- AI-generated dashboards are standard Birst dashboards. After publication, users can edit, share, and filter the dashboards.
- The AI respects your user permissions. It can only use measures and attributes you have access to.
- Complex requests can require more processing time. If a request is timed out, rephrase the request with more specific details.
An example procedure: