Zia can predict anomaly in incoming and outgoing ticket volume. A sudden surge or drop in the number of tickets that are received and number of responses that are sent is analyzed by Zia and flagged as an anomaly for admins to check and fix the problem.
Monitoring anomaly can be particularly useful for a business to understand how customers react towards change in pricing strategy, adoption of new product, sunset of a long running product, and most importantly the issues they face in the product or service.
Increase in ticket count can indicate malfunction or poor customer service, carefully monitoring the customer requests can indicate the underlying reason allowing a business to fix in on time. Overtime if the number of tickets for the same issue significantly drops it can indicate that the customer's are happy with the fix or your product is functioning as expected.
Enabling anomaly prediction
Anomaly prediction is department specific.
Setting up anomaly prediction criteria
Admins can setup an anomaly prediction model for incoming tickets, outgoing tickets or both as per their requirement. For instance, some departments may want to monitor only the outgoing responses, they can set up the anomaly model only for the outgoing responses and get notified about it.
The anomaly criteria can be set in the following ways:
- Number of tickets: Admins can select the predefined number of tickets (5, 25, 50) that should be considered as an anomaly when compared with the trend. Eg., if every Tuesday afternoon you usually receive 20 tickets and the Tuesday you receive 25 tickets (20 + 5 (anomalous number)) Zia will flag it as an anomaly.

- Custom trigger criteria: Admins can choose a custom criteria instead of predefined set of numbers. They criteria includes the volume of tickets and the deviation %.
- The volume or number of tickets can be selected from a scale of:
6 to 25 tickets
26 to 50 tickets
51 to 75 tickets
76 to 100 tickets
More than 100 tickets
- The deviation % can be selected from 10%, 25%, 50%, 100% or dynamic. The deviation % is calculated as Current value - Trend value.
The system will only define an anomaly if both the conditions are met.
For example, if your "Trend Value" is usually 20 outgoing messages, but the "Current Value" jumps to 40, you have a 100% deviation. Only if both the number of tickets and the deviation % match system will recognize it as an anomaly and notify.
Setting up anomaly notification for business hours
Zia can notify the users whenever it detects an anomaly in incoming or outgoing trend allowing admins to take proactive measures and fix the problem. The notification can be triggered at specific business hours to ensure the anomalies don't get overlooked and receive immediate attention. The notifications can be viewed under the Anomaly tab in Zia notifications. The Zia notification icon is displayed at the bottom bar.
To set up anomaly prediction and notification
- Navigate to Setup > Zia > Anomaly.
- Under Prediction for Incoming Tickets tab, select the Number of tickets (5, 25, 50) or click Customize Trigger Criteria.
- Under Customize Trigger Criteria select the desired option for each range and click Apply.
- Toggle Notification Settings for Business Hours, if needed.
- Repeat step 2 and 3 for Prediction for Outgoing Tickets.
- Click Save.

Viewing the anomaly in prediction dashboard
The Prediction dashboard provides a visual representation of the current trends. The dashboard displays the following components:
- Trends Vs Incoming Responses or Outgoing Responses
- Trending Auto Tags
- Sentiment Analysis
- Sentiment Trend Analysis
Trends Vs incoming or outgoing responses
Zia analyzes the ticket traffic for the last 30 days and predicts the trend for current day. When there is a surge or dip in the predicted pattern, it marks it as an anomaly and notifies via Zia Notification Center.
The line graph displays the time versus the number of responses in the x and y axis respectively.