Getting Started with AI Ticket Suggested Fields

This AI Suggested Fields article will help administrators to configure the feature from the ground up. 

Overview

AI Ticket Suggested Fields are a feature of TeamDynamix which use ticket history to train generative AI models, allowing the models to suggest ticket field values based on previous assignment. This feature is configured in Work Management TDAdmin for use by TeamDynamix Technicians.

This article goes over the bare-bones steps required in order to get AI Ticket Suggested Fields configured in TeamDynamix Work Management environment.

Environment

  • TDAdmin
  • TDWorkManagement Technician access to a ticketing application

Getting Started with AI Suggested Fields

Step 1: Create an AI Model

The first step with AI Suggested Field configuration is to train an AI model with a list of tickets to serve as good examples for your previous field assignments. We recommend choosing a large set of closed/resolved tickets which have been properly assigned values for the fields you wish to suggest.

  1. In Work Management TDAdmin, navigate to Applications > {{ your ticketing application }} > AI & Automation > AI Models
  2. Click the + New Model button to train your AI model
  3. Select the ticket fields to be learned by this model. Options include: Type, Priority, Form, Services, Responsible Group
    • Note: the model will use more fields than just the ones selected here for its context, but these are the ones that it will learn to suggest
  4. In the section Select Ticket Filters for Model Training, build a set of filters for your model. TeamDynamix suggests the following;
    • Include tickets that have accurate assignment and responses (closed and resolved)
    • Include as many of these tickets as possible (1k is the lower limit, and the model will use a random subset if more than 250k records are selected 
    • Do not include defaults (ex. if you have a default service for "General Support", you may want to exclude that from the training resources)
  5. Choose the users to Notify when AI model training is completed
  6. Click Save and Train to initiate the learning process for your AI model
    • Note: this may take a decent amount of time, depending on your set of tickets to be used for training. It should take less than an hour, and a decent expectation is somewhere around 15 minutes.
  7. After the AI model is trained, open the configuration page and click "Enable" to activate it

Step 2: Configure AI Suggested Fields

After your model has completed training, you can set up the actual field suggestions that will be suggested in the application.

  1. In Work Management TDAdmin, navigate to Applications > {{ your ticketing application }} > AI & Automation > AI Suggested Fields
  2. Use the checkboxes under Field Application to select which ticket classifications should use AI Suggested Fields
  3. Under each checked ticket classification, select the checkboxes for which fields should be suggested on those classifications
  4. Click Enable at the top of the page to activate AI Field Suggestions for the application

 

Step 3: Add AI Permissions to Application Security Roles

Once AI Field Suggestions is configured, you will need to allow your technicians to see field suggestions.

  1. In Work Management TDAdmin, navigate to Applications > {{ your ticketing application }} > Users & Roles > Security Roles and click on the security role to update
  2. Under Ticket, check the boxes for the following:
    • The user will be able to view AI Suggestions
      • The user will be able to view and accept/reject AI suggestions when creating and editing tickets
    • [optional] The user will be able to view raw AI Suggestion data
      • The user will be able to view all suggestion values for a given field on a ticket, as opposed to a filtered set of suggestions
  3. Click Save to apply changes

 

Step 4: Test AI Suggested Fields

  1. Navigate to Work Management /TDWorkManagement and open a ticketing application
  2. Select a ticket, and click Edit to open the editing form
  3. Observe and test the options below the fields you selected for your AI model to suggest
Print Article