Use system-derived entities in an NLU topic

  • Rversion finale: Australia
  • Mis à jour 12 mars 2026
  • 2 minutes de lecture

    • Associate system-derived entities such as such as date, time, duration, or location with a node in a topic.

    Avant de commencer

    Create the Booking Reservations NLU model and system entities (FromLocation and ToLocation) as described in Create a system-derived entity. In this example procedure, you will create a topic that can slot fill locations that are used in an utterance.

    Role required: virtual_agent_admin or admin

    Procédure

    1. Create a new topic called Booking Reservations.
      When completing the Properties tab, make sure that you choose the Booking Reservations NLU model that you created in an earlier exercise.
    2. Drag a Text bot response component onto the canvas, and then select it.
      In the properties sheet, fill in the fields.
      Tableau 1. Response Properties sheet
      Field Description
      Node name Type Intro.
      Response Message Type Let's get you booked on a flight.
      Figure 1. Intro Text response properties
      In the Text response properties, the Node name is "Intro," and the Response Message is, "Let's get you booked on a flight."
    3. Drag a Text user input component onto the canvas.
      In the properties sheet, fill in the fields.
      Tableau 2. User Input Properties sheet
      Field Description
      Node name Type Fromwhere.
      Prompt Type Where are you departing from?
      Input format Select Text. This doesn't restrict the input format.
      NLU entity Select FromLocation.

      This list only shows existing entities, such as system entities or entities that you created for the model.
      Do not ask users to confirm recognized entity

      Select this option to avoid having Virtual Agent ask the user to confirm the location.
      Advanced (optional)
      Topic switching
      Enable NLU to switch topics Slide this toggle switch to enable topic switching.
      Figure 2. Fromwhere Text user input response properties
      In the Fromwhere Text user input node, the prompt reads, "Where are you departing from?" The NLU entity used is FromLocation. Topic switching is enabled.
    4. Drag another Text user input component onto the canvas.
      In the properties sheet, fill in the fields.
      Field Description
      Node name Type Towhere.
      Prompt Type Where are you heading?
      Input format Select Text. This doesn't restrict the input format.
      NLU Entity Select ToLocation.

      This list only shows existing entities, such as system entities or entities that you created for the model.
      Do not ask users to confirm recognized entity Select this option to avoid having Virtual Agent ask the user to confirm the location.
      Advanced (optional)
      Topic switching
      Enable NLU to switch topics Slide this toggle switch to enable topic switching.
      Figure 3. Towhere Text user input response properties
      In the Towhere Text user input node, the prompt reads, "Where are you heading?" The NLU entity used is ToLocation. Topic switching is enabled.
    5. Drag another Text bot response component onto the canvas, and then select it.
      In the properties sheet, fill in the fields.
      Field Description
      Node name Type Confirmation.
      Response Message Type Great! I have you booked on a flight from [Fromwhere] to [Towhere]. Bon voyage! Use dot-walking to replace [Fromwhere] and [Towhere] with the input variables you created in the previous nodes.
      Figure 4. Confirmation text response properties
      The Confirmation node displays the message, "Great! I have you booked on a flight from blank to blank," which will be slot-filled from the Fromwhere and Towhere input variables.
    6. Click Save.

    Que faire ensuite

    Use additional system-derived entities for your Booking a Reservation topic. For example, you can use system-derived entities to capture the dates that the user wants to travel.

    For more information about default system entities, see NLU system entities.