Widget-Level Insights (Narration)

·         Access 'Insight' from a widget's menu

·         View AI-generated summary or narration per chart

·         Get results influenced by applied filters, highlights, DLS, splits, or calculated fields

·         Choose narration mode (paragraph or bullet list)

1.1 Generate Widget-Level Narration

To generate a widget-level GenAI summary, follow the steps below:

1.       Plot a visual/chart widget on the dashboard.

2.       Navigate to view mode of dashboard.

3.       Click the three-dot menu (More Options) on the widget.

4.       Select the 'Insight' option.

5.       The Generate Insight modal will appear, presenting multiple options to generate AI-driven interpretations.

6.       Choose the further required option to generate insights

1.2 Widget-Level Narration Options

Modal Components

1 Buttons

• Generate Narrations
  Produces a detailed, paragraph-style explanation of the data.
  Use when you need a story-like, descriptive interpretation.

• List Key Insights Using Bullet Points
  Displays insights in a structured, bullet-point format.
  Use when you want a concise, scannable list of findings.

• Generate Summary
  Creates a short, high-level overview of the dataset.
  Use when you need an executive summary.

• Show Me Anomaly
  Identifies unusual patterns, spikes, or drops in the dataset.
  Use when validating data quality or monitoring exceptions.

2 Insight Output Box

• Displays the AI-generated content (narration, summary, insights, or anomaly findings).
• Users can scroll within the box to review the full generated text.
• The content can be copied for use in reports, dashboards, or presentations.

·         Only visual widgets support this feature. Text, shape widget, Filter Widget, Parameter Widgets are excluded.

·         The 'Insight' option is disabled when RubiAI is not configured in Admin. A tooltip will indicate this.

·         Summaries are temporary and disappear upon refresh.

·         Drill-down views will generate insights as per the current level.

·         Summary content will respect DLS rules. Data visibility will vary by user access.

Usage Best Practices

• Review Carefully: AI-generated content may contain inaccuracies; always validate with raw data.
• Choose the Right Format: Select narration for detailed reports, summaries for quick overviews, and bullet points for presentations.
• Anomaly Check: Always confirm anomalies against source data before acting on them.
• Iterate if Needed: Regenerate insights if the initial version doesn’t align with expectations.
  

Limitations

• Generated insights are interpretative and may not always be statistically precise.
• Anomalies flagged are suggestions, not confirmed errors.
• Complex datasets may produce overly generalized summaries.