Smart Suggestions

Smart Suggestions are recommendations provided by the system when assigning a category or technician group to a request. This helps technicians quickly select the required option. The system provides suggestions by analyzing the request subject and description and by considering the configured accuracy level and the number of requests to display precise results.

note

• Smart Suggestions are supported only in the Request Management module.
• Currently, SmartSuggestions apply only to the Category and Technician Group fields.
• For each category, 50 to 300 requests are required for Smart Suggestions to function.
• Only requests created after the model has been trained are considered for evaluation.

How It Works

The system displays suggestions based on the number of requests. You can understand this by considering the following example.

If the system has requests in the following four categories:

• Hardware: 100 requests
• Software: 40 requests
• HR: 110 requests
• Non-IT: 130 requests

The trained model will consider only Hardware, HR, and Non-IT categories for displaying suggestions. It will not consider Software because it has fewer than 50 requests.

From the eligible categories, the model will first consider the one with the fewest requests (i.e., Hardware with 100 requests). Based on this, it will consider 100 requests from each of the remaining categories to achieve a total of 300 requests. i.e., Hardware: 100, HR: 100, and Non-IT: 100 = 300 requests.

Therefore, 100 requests from each of the three categories will be considered for displaying suggestions. The trained model considers an equal number of requests from the eligible categories.

If the number of requests across eligible categories is unequal, the trained model will not consider those categories for evaluation.

Configuration

To configure Smart Suggestions:

1. Navigate to the Admin > Automation > Smart Suggestions page.

2. Configure the following parameters:

   • Enable Field Suggestion: Enable this option if you want suggestions to appear in the fields. By default, it is disabled. Once enabled, the following parameters will become available.
   • Accuracy Level: Set the accuracy threshold for displaying suggestions. The configured level must be lower than the model's current accuracy to ensure suggestions are visible.
   • Field Suggestion: Select the fields for which you want suggestions to appear. Currently, the Category and Technician Group fields are available.
   • Resolution Suggestion: Select an option to view suggestions for resolving the request. The available options are: Knowledge Base and Response Template.
   note

   Suggestions are displayed based on the percentage of the subject and description content that matches the existing Knowledge Base articles or response templates.

   • Notification Text: Enter the text to include in the email notification for the Knowledge Base or response template.

3. When finished, click Train Model to train the model and display the results. Select the options you want to train the model for. The options selected in the Field Suggestion and Resolution Suggestion sections appear in the list. Changes to the trained model apply only to new requests.

4. Click Update to save the changes. You can also click Cancel if you want to discard the changes.

After you train the model and update the settings, logs appear on the Configuration Audit page. Additionally, suggestions appear for the selected field and resolution options on the Request List and Request Details pages, respectively, as shown in the following examples.

Example 1: Suggestions for the Technician Group field are shown below.

Example 2: Suggestions appear when you insert a Knowledge Base article into a request, as shown here:

Example 3: Suggestions appear when you insert a response template into a request, as shown here: