Overview
The Root Cause Analysis calculator uses AI-driven decision tree analysis to identify the root causes of selected process behaviors. This calculator automatically analyzes your event log data to discover which factors (such as attributes, values, or conditions) most strongly influence specific outcomes like rework, late payments, or case escalations.
This is an AI-powered calculator that requires minimal configuration and provides intelligent insights into what drives specific process behaviors.
Common Uses
- Identify factors contributing to invoice payment delays
- Understand what causes rework in processes
- Discover patterns that lead to case escalations
- Analyze what drives quality issues or defects
- Determine root causes of process bottlenecks
- Find out why certain cases take longer than others
Settings
Target Behavior: Define the process behavior you want to analyze by creating a filter that selects cases exhibiting that behavior (e.g., cases with rework, late payments, or long durations).
Attributes to Analyze: Select which case and event attributes the AI should consider when building the decision tree. The calculator will automatically determine which attributes have the strongest influence on the target behavior.
Minimum Case Volume: Specify the minimum number of cases required for meaningful analysis. The AI needs sufficient data to identify statistically significant patterns.
Confidence Threshold: Set the confidence level for the decision tree branches. Higher thresholds result in more reliable but potentially fewer insights.
Example
Finding Causes of Late Invoice Payments
Scenario: You want to understand why some invoices are paid late while others are paid on time.
Setup:
- Create a filter selecting cases where Payment Timeliness = "Late"
- Select attributes to analyze: Vendor, Invoice Amount, Department, Payment Terms
- Set minimum cases to 100 for statistical validity
Output:
The calculator generates a decision tree showing:
- Primary Root Cause: Invoices from Vendor Category "International" are 3x more likely to be late
- Secondary Factor: Within international vendors, invoices over $10,000 have 85% late payment rate
- Contributing Factor: Invoices routed to Department "Procurement B" have higher late payment rates regardless of vendor
Insights: The analysis reveals that international vendors, especially for high-value invoices, need different payment processes. The decision tree helps prioritize which process improvements will have the biggest impact.
How to Interpret Results
The decision tree output shows:
- Nodes: Represent decision points based on attribute values
- Branches: Show how cases split based on conditions
- Leaf Values: Indicate the percentage of cases meeting your target behavior at each endpoint
- Importance Scores: Highlight which attributes have the strongest influence
Look for branches with high percentages and large case volumes - these represent the most significant root causes requiring attention.
Output
The calculator displays an interactive decision tree visualization that you can:
- Expand or collapse branches to explore different paths
- Click on nodes to see detailed statistics
- Identify the strongest predictors of your target behavior
- Export findings for presentation or further analysis
This documentation is part of the mindzie Studio process mining platform.