Create a Virtual Agent topic

Use Virtual Agent Designer to create and modify Virtual Agent topics (conversations). Build your topic to meet a specific objective or goal, such as fulfilling a user's request or helping a user complete a task.

시작하기 전에

Do the following:

• Review Getting started with Virtual Agent Designer.
• Configure Virtual Agent settings. For example, if you're creating large language model (LLM) topics, turn on Now Assist in Virtual Agent. For more information, see Configuring assistants overview. If you're creating Natural Language Understanding (NLU) topics, configure NLU settings. For more information, see Configure Natural Language Understanding in Virtual Agent.
• Plan your conversation before you begin. For more information, see Designing a Virtual Agent topic.
• Verify that you're in the appropriate application scope before you create or update a topic. For example, if you're creating ITSM topics, verify that you're in the ITSM Virtual Agent Conversations scope (and not the scope for the ITSM NLU Model for Virtual Agent Conversations). For details, see Change the application scope for Virtual Agent Designer topics.

Role required: virtual_agent_admin or admin

이 태스크 정보

The topic is in the Inactive state until you publish it. Publishing a topic changes the topic state to Active and automatically deploys the topic to your Virtual Agent clients.

The Large Language Model (LLM) topic type is faster and easier to set up compared to using Natural Language Understanding. You don't have to select NLU intents or train NLU models for your virtual agent. For more information, see LLM topic discovery in Virtual Agent.

When NLU is enabled, the Set up Natural Language Understanding (NLU) section is displayed in the Topic Properties page. You can use this section to specify the associated NLU model and intent for the topic. You can also optionally specify keywords that Virtual Agent can use as a fallback if NLU doesn't return a matching intent and topic. For more information, see Natural Language Understanding (NLU) topic discovery in Virtual Agent.

프로시저

1. Navigate to All > Conversational Interfaces > Assistant Designer > Virtual Agent > Designer.
2. Select LLM or NLU/keyword depending on the type of topic you're creating.
3. Select the type of topic.
   • For LLM topics:
     1. Select Create asset.

        

     2. From the Create asset window, select Topic.
        중요사항:

        Selecting any of the other topics in the window will open the create flow in the respective application.
   • For NLU/Keyword topics, select Create topic.

     

4. On the form, fill in the fields.
   For a description of the field values, see Create a topic form.
5. On the topic creation form, select Create.
6. On the Flow tab, build the conversation flow.

   Flow steps	Description
   Add controls	To create the topic flow: Select the controls for user inputs, bot responses, or utilities from the palette and drop them onto the canvas. When you add a control to the canvas, it becomes a node in the conversation flow. For each node, complete the corresponding property sheet. In the topic header bar, select Save. The topic header bar displays a Saving in progress... message. The toast message Successfully saved your topic appears in the lower-right corner afterward for 4 seconds, and can be closed by selecting the X in the message. For details about each control and the corresponding properties that you define, see User inputs, Bot responses, and Utilities. As you add controls, watch for a yellow warning badge or a red error badge in the corner of the control. The number on the badge indicates the number of issues in that control, such as missing mandatory information in the property sheet. (Highlighting a badge or the "Incomplete" banner at the top of the property sheet details the items in need of correction.)
   Move a node	Select the node, drag it to the new location in the flow, and Select Save.
   Delete a node	Select the X in the upper right corner of the node. In the topic header bar, select Save. The topic header bar displays a Saving in progress... message. The toast message Successfully saved your topic appears in the lower-right corner afterward for 4 seconds, and can be closed by selecting the X in the message. 주: You can't delete a node that has a script variable used by subsequent JavaScript expressions in the flow. Also, you can delete a decision node only when it has a single branch.

7. Select Test in the header bar to test your topic in a chat window.
   You can review test results in the adjoining tabs. For detailed information, see Testing LLM topics and Testing NLU/Keyword topics.

   If the Now Assist panel, Microsoft Teams application, or Slack application is configured for your environment, preview options for those channels are displayed in the Test button list. Select Preview in Now Assist panel or Preview in Microsoft Teams in the list to test your topic in those environments. For more information, see Integrating Virtual Agent with messaging apps.

8. When you're finished with the topic, select Save in the topic header bar.
   The topic is in the Inactive state until you publish it.
   The topic header bar displays a Saving in progress... message. The toast message Successfully saved your topic appears in the lower-right corner afterward for four seconds, and can be closed by selecting the X in the message.

다음에 수행할 작업

Publish the topic to deploy the topic to your Virtual Agent clients. The topic state changes to Active.

If you created a topic that might be frequently used in certain contexts (environments), consider promoting the topic as part of the default or custom chat experience.

• NLU model mapping in Virtual Agent Designer
  ServiceNow model groups help you to manage the localization of both topics and Natural Language Understanding (NLU) models more easily. You can map Virtual Agent topics to a new or existing model group directly from Virtual Agent.
• Topic mapping to a multilingual NLU model group
  When you map a topic to a multilingual model group in Virtual Agent Designer, the topic mappings for enabled languages are set up automatically.
• Modify NLU utterances and entities for a Virtual Agent topic
  View, test, and modify NLU utterances for a Virtual Agent topic on the NLU Intent tab in Virtual Agent Designer.
• Add NLU to an existing keyword topic
  Create and map an NLU model group and intent for the topic from Assistant Designer.
• Add nodeless NLU entities as input variables to a topic
  You can add nodeless NLU entities as input variables to a topic if NLU is enabled for Virtual Agent. These variables can be slot-filled from NLU service provider predictions or provided outside of the scope of the topic.
• Branch a Virtual Agent conversation with a Decision utility
  Your conversations may need to branch depending on end user input or choices. Use the Decision utility to branch the conversation when certain conditions are met.
• Insert a header card in a Static Choice or Dynamic Choice control
  When you create a Virtual Agent topic, you can include images and YouTube videos on Static Choice and Dynamic Choice user input controls.
• Change the application scope for Virtual Agent Designer topics
  Set the application scope before you create or update a topic. For example, if you're creating ITSM topics, verify that you're in the ITSM Virtual Agent Conversations scope (and not the scope for the ITSM NLU (Natural Language Understanding) Model for Virtual Agent Conversations).
• Define script variables for a topic
  Use the Variables sidebar to define script variables for a topic. Script variables can store information that is not found in ServiceNow tables. This information can then be shared elsewhere in a topic.
• Create a small talk topic
  Build small talk topics that let Virtual Agent engage in casual conversation with users. A small talk topic provides a response to a casual question that users might ask during a conversation, such as the time or date. A small talk topic can occur anytime within a conversation session and can be unrelated to the original conversation intent.
• Control topic discovery and visibility
  Admins and topic authors can enable Virtual Agent to discover topics for use in chats, include a topic in the list users see, or both. The topics can be setup topics and small talk topics.
• Use system-derived entities in an NLU topic
  Associate system-derived entities such as such as date, time, duration, or location with a node in a topic.