Permission Required
Users with the Support Administrator and Agent permission profile can access this feature.
Check Feature Availability and Limits

1. It's important to note that currently, Zia is limited to the English language, but we are actively working on expanding language support.
   
2. Users with the agent profile can't access the anomaly predictions feature.

Anomaly Predictions 

• Incoming Ticket Responses
  
• Outgoing Ticket Responses
  
  
  

As you can see, the Incoming Responses chart comes with three key components:

• Last 30 days Trend: The trend line shows how the incoming ticket response is supposed to fare going by your past data.
  
• Today's Incoming Responses: The line shows how the day's incoming responses are faring in reality as against how it is supposed to, shown in the trend line. It is this line that shows any spike or dip.
  
• Anomaly: A sudden spike or a dip from what is normal to your help desk is called an anomaly. The red stars in the chart denote anomaly.
  

The example above shows typical incoming traffic patterns for one day. Taking a closer look at the trend, we can see that traffic peaks between 7:00 AM and 3:00 PM daily and increase again after 9:00 PM. However, there is a steep surge in the Actual Line at 2:00 PM today, which interprets that your incoming ticket responses have increased radically for the day against the trend for the last 30 days.

Viewing Top Tags

Click on a data point on the dashboard to view the top trending tags during that particular hour. Each bar on the chart represents a tag, while its size indicates the number of tag occurrences in incoming or outgoing responses. This information is beneficial to know the context in which an anomaly occurred.

Setting up Anomaly Predictions

To get notified about anomalies, you must turn on anomaly predictions, choose what you want to predict, and define what is considered an anomaly.

To do these, follow the steps below:

1. Click the Setup icon (  ) in the top bar.
   
2. Click Zia under the General menu.
   
3. Click Anomaly under the Zia sub-menu.
   
4. On the Anomaly Services page, select the department from the top of the page.
   Anomaly Predictions will be enabled for the department chosen here.
   
   
   
   
5. Toggle on the Anomaly Predictions setting and do the following:
   
1. Select whether you want Zia to predict anomalies in incoming tickets, outgoing tickets, or both.
   
2. Select the minimum number of incoming ticket responses for Zia to consider as an anomaly.
   You can choose between 5, 25, or 50 responses.
3. Click Customize Trigger Criteria to specify the permissible deviation percentage for different ticket response volumes.
   For example, consider you set 50 as the deviation percentage for incoming ticket responses, and the trend line predicts 100 responses at a given hour. In this case, Zia will only trigger an anomaly when the actual incoming ticket responses are 150 or higher at that hour.
   
   
   
   
4. Toggle on the During Business Hours setting to receive anomaly notifications during specific business hours.
   
5. Select the desired business hour from the drop-down menu.
   
6. Click Save to reflect the changes for incoming responses. (or)
   Click Save Apply Same Settings for Outgoing Tickets.
   Note: You can also customize the settings for outgoing responses individually.

Zia Notifications 

1. Click the Setup > Zia 
2. On the Predictions page, do the following:
   
1. Select the department for which you want to enable notifications from the top of the page.
   
2. Toggle on the Zia Notifications setting.
   
3. Click Notify To to add agents who will receive notifications about anomalies.
   This will open the Users to Notify panel.
4. Add agents, teams, and agents under a role using the drop-down menu.
   
5. Click Save.
   

1. Click the Zia icon (  ) in the bottom right corner of the screen in your Zoho Desk account.
   The Zia Notifications panel opens on top of your Zoho Desk page in the same tab.
2. View the anomalies in the incoming or the outgoing ticket responses for the current department.
   
3. Click on the notification to open the anomaly graph to review the surge or dip against the trend.
   

• The Zia notifications stay for 15 days, after which they get removed.

Sentiment Analysis 

• Ticket List - An aggregation of the overall sentiment is displayed alongside individual tickets.
  
• Ticket Details - The sentiment of every incoming ticket response is displayed along with the keyword attributes.
  

The Zia Dashboard displays two graphs to monitor and track customer sentiments. The first graph shows the percentage of positive, neutral, and negative sentiments and the number of responses that showed these sentiments over the last 24 hours.

The second graph shows a break up of positive, neutral, and negative sentiments at each hour over the last 24 hours. As an option, you can also view the trend for the previous 7 days. By visualizing the graph, you can quickly tell whether there were significantly more negative emotions at any hour of the day.

Ticket Auto Tags 

Note:

• A department must have a minimum of 3000 individual tickets via email, help center, web forum, community, or social for Zia to train and generate auto tags.
• Learn more about this feature in detail.
  

 Zia Answer Bot 

Here are the actions that you can perform on the Answer bot panel from inside a ticket:

• Chat with Zia: You can chat with Zia about customers' queries and receive answers from the knowledge base articles.
  
• Use this in your reply: It allows you to paste the paragraphs suggested by Zia into the ticket response. The reply editor opens with the content pasted into it automatically.
  
• Link to article: It opens the article details window on top of your ticket. You can perform the following actions on this window:
  
• Share link to ticket: A link to the article will be added to the ticket response.
  
• Paste to ticket: The article content will be pasted into the ticket response. Please note that the entire article will be pasted once this button is selected.
  
• Helpful?: You can provide feedback about the suggestions by clicking on the "thumbs up" or "thumbs down" icons. Zia learns from your feedback, so the more accurate it is, the more precise suggestions you will get.