Triage Analytics automates support ticket tagging by tracking the performance of each AI classifier and providing analytics on the number of tickets that are accurately classified. It has four sections: Triage Classifier Performance, Triage Classifier’s reason for no prediction, Ticket Classification Details, and Classifier’s Predicted Value and Accuracy.
Note: All analytics, including Looker tables and insights tables, are processed every 1 to 2 hours.
Triage Classifier Performance
This graph displays all the enabled classifiers within the selected time period. It depicts a funnel sequence, starting with tickets qualified for prediction, followed by those on which predictions were made. Subsequently, it shows predictions written to the helpdesk, tickets among these that were closed or resolved, and finally, the accuracy of the predictions on these closed or resolved tickets.
| Column Name | Description |
| Qualified prediction | Based on the initial setup, this shows the number of incoming tickets that qualify for Forethought Triage to classify for the selected classifier. |
| Made prediction | Number of predictions Forethought classifier made. |
| Was written to helpdesk | Number of predictions for which Forethought successfully wrote the classifier prediction value to the helpdesk. |
| Closed cases | Number of cases that are currently closed from the predictions where Forethought wrote the prediction to the helpdesk. |
| Prediction was correct | Number of predictions for which the prediction was correct and not changed in the helpdesk when it was closed. Accuracy is calculated only for tickets which have status closed or resolved. |
Triage Classifier’s reason for no prediction
Ticket Classification Details
| Column Name | Description |
| Ticket ID | Unique ticket ID of the helpdesk |
| Classifier name | Name of the classifier that is enabled |
| Ticket title | Title of the ticket |
| Ticket body | Body of the ticket |
| Ticket status | Ticket status |
| Did Forethought predict? (Yes/No) |
Yes - When forethought did a prediction No - Whne forethought did not do a prediction |
| Forethought predicted value | The classifier prediction value |
| Most recent value in helpdesk | Shows the last updated filed value in the helpdesk where the prediction was writen to. |
| Prediction was correct (Yes/No/Null) |
Yes - Predicted value was correct No - Prediction value was incorrect Null - Ticket has not been closed/resolved to calculate this. |
| Reason for no prediction | The reason Forethought did not predict |
| Was the prediction written to the helpdesk? (Yes/No) | If the predicted values were written to the helpdesk. |
| Prediction Date | Date when the prediction was performed. |
Classifier’s Predicted Value and Accuracy
| Column Name | Description |
| Classifier Name | Name of the classifier that is enabled |
| Predicted Value | This shows all the values the classifier can predict. (Each row = unique classifier prediction value) |
| Total predictions | Number of predictions that had predicted value |
| Prediction accuracy | Number of tickets that were closed where the prediction value was correct and not changed after the initial prediction. |