Skip to main content
Auto-Explain is a feature that automatically analyzes data variations and identifies root causes. It examines the underlying data structure to determine which factors contributed to a metric change between two data points.

​
Atoti Intelligence Essentials

This is part of the Atoti Intelligence Essentials offer.

​
What is Auto-Explain

Auto-Explain analyzes data to identify the root causes of variations. When a metric changes between two data points, Auto-Explain examines the underlying data structure to determine which factors contributed to the change. Auto-Explain produces two kinds of output:
  • Root-cause analysis: the root cause members, their contribution percentages, and the contribution tables. This is produced by a deterministic algorithm and does not require an LLM.
  • AI summary: an optional natural-language summary of the results. This requires an LLM to be configured. When no LLM is configured, requesting an AI summary returns a message indicating the AI summary is unavailable instead of a generated summary.

​
Why use Auto-Explain

Auto-Explain reduces the time spent investigating data variations manually. Key benefits include:
  • Identify root causes automatically
  • Get clear explanations of metric changes
  • Work with existing Atoti data models
  • Focus analysis on relevant dimensions through configuration

​
Where to find Auto-Explain

Auto-Explain is available in the context menu of pivot tables in the Atoti UI after successful setup. To access it, right-click on a selected cell in a pivot table and select Auto-Explain from the context menu:
Auto-Explain context menu

​
How to use Auto-Explain

​
Prerequisites

Before using Auto-Explain, ensure the following requirements are met:
  • Auto-Explain is set up in the project
  • The Atoti application is running
  • The Atoti UI is accessible
  • A pivot table with data is available

​
Analyze a variation

Follow these steps to analyze a variation between two cells:
  1. Open a pivot table in the Atoti UI
  2. Select two cells to compare
  3. Right-click one of the selected cells
  4. Select Auto-Explain from the context menu
  5. Wait for the analysis to complete
Auto-Explain analyzes the variation and displays the results:
Auto-Explain results

​
Interpret the results

The Auto-Explain results always include:
  • Root cause members that contribute to the variation
  • Contribution percentages for each factor
  • Hierarchy levels where variations occur
If an LLM is configured and the analysis was requested with AI summary enabled, an additional natural-language summary of the results appears below the contribution data. The AI disclaimer is displayed alongside the summary. If an AI summary is requested but no LLM is configured, a message indicating the AI summary is unavailable appears instead of a generated summary.