Documentation Index

Fetch the complete documentation index at: https://help.symphonysummitai.com/llms.txt

Use this file to discover all available pages before exploring further.

SDI Routing Health

Updated
  • 3 minute(s) read
Prev Next

SDI Routing Health enables ITSM/AI platform Administrators to track routing prediction accuracy over time through a simple process.

It provides a centralized view of the accuracy and health of ticket routing predictions. It enables Administrators to monitor routing performance across key dimensions such as Impact, Urgency, Priority, Classification, Category, and Workgroup. It identifies accuracy trends and factors contributing to low accuracy, and review detailed mismatches in evaluated tickets. The SDI Health Routing dashboard also provides visibility into the evaluation pipeline status and allows Administrators to trigger a recompute when required.

The following infographic depicts the workflow and connectivity between various components on SDI Routing Health Dashboard.

Benefits

  • Measures the accuracy of routing predictions against evaluated tickets.

  • Provides an overall view of evaluated/routed tickets, overall accuracy, and workgroup accuracy.

  • Tracks accuracy trends based on tickets closed within 1, 7, or 30 days.

  •  Shows the number of tickets evaluated and the current status for each routing dimension.

  • Displays the status of the routing evaluation pipeline from prediction collection through completion.

  • Allows users to recompute routing evaluation data.

Use Case

An organization recently reorganizes its Service Desk workgroups and updates routing rules. Within a few weeks, Analysts notice that more tickets are being reassigned manually, but the Administrator cannot determine whether the issue is caused by the new workgroup structure or declining SDI prediction accuracy.

The Administrator runs the SDI Prediction Accuracy evaluation for the latest period and compares it with previous accuracy windows. The accuracy trend shows a decline in workgroup prediction accuracy. The Detailed Accuracy Table then reveals that tickets from a specific category are frequently being predicted for the wrong workgroup.

Using this insight, the Administrator can identify the affected routing dimension, review the corresponding routing configuration or training data, and take corrective action before inaccurate predictions lead to continued manual reassignment and routing delays.

Viewing SDI Routing Health

To view SDI Health Routing screen, perform the following steps:

  1. Navigate to SDI Routing Health > AI > AIW Configuration.


    Figure: AIW Configuration

  2. Click AIW Configuration. The following screenshot is displayed:


    Figure: Knowledge sync

  3. Click the 3rd icon from the left navigation panel. The SDI Routing Health page is displayed.

    Figure: SDI Routing Health

The following table describes the field description for SDI Routing Health page.

Field

Description

SDI Routing Health

Displays the overall health and accuracy of SDI ticket routing.

Evaluation Period

Allows users to select the evaluation period. This drop-down has following options:

  • Day

  • Week

  • Month

  • Quarter

Filter

Allows to filter the data based on the selected Department. This drop-down displays the options ‘ALL’ and individual Departments.

Refresh

Refreshes the routing health data displayed on the dashboard.

Export

Allows users to export the displayed routing health information.

Evaluated / Routed

Displays the number of tickets evaluated against the total routed tickets for the selected period.

Overall Accuracy

Displays the average routing accuracy across all evaluated routing dimensions.

Workgroup Accuracy

Displays the accuracy of workgroup routing across the evaluated tickets.

Accuracy Trend

Displays routing accuracy over time for tickets closed within 1, 7, or 30 days.

Mousehover displays the detailed information for the day as a tooltip.

Screenshot

1 day / 7 days / 30 days

Indicates the ticket closure periods used to calculate and display accuracy trends.

Accuracy Impact Drivers

Displays factors contributing to low routing accuracy during the selected period.

Low Model Confidence

Indicates the percentage/share of evaluated results affected by low model confidence.

Pipeline Completed

Indicates that the routing evaluation pipeline has completed successfully.

Displays the progress/status of stages including Collect Predictions, Fetch Ground Truth, Evaluate, Aggregate, Mark Expired, Purge Retention, and Completed.

Displays the date and time when the evaluation pipeline was last executed.

Recompute

Initiates a new routing accuracy evaluation and refreshes the latest accuracy metrics.

Note

The Recompute process uses the ticket data from last three months.

Detailed Accuracy Table

Provides a dimension-level breakdown of routing accuracy and recent evaluated predictions.

Dimension

Identifies the routing attribute being evaluated, such as Impact, Urgency, Priority, Classification, Category, or Workgroup.

Accuracy

Displays the routing accuracy for the selected dimension.

Top Mismatch

Displays the most significant mismatch between the predicted and actual value for the dimension.

Evaluated

Displays the number of tickets evaluated for the dimension.

Status

Indicates the current evaluation status of the dimension, such as Good, Bad and Moderate.

Dimension / Predictions

Allows users to switch between the dimension-level accuracy view and prediction-level details.

Pagination

Allows users to navigate through additional records in the Detailed Accuracy Table.