Virtual Agent Designer user input controls

Use the collection of input controls provided by Virtual Agent Designer to prompt and capture information from the user in a conversation.

Common input control properties

In addition to data pills and condition controls, each input control has the following common properties:

Input variable for the user response
Each control stores the user response in an input variable that is accessible from other controls within the same topic using the getValue and getDisplayValue methods. The variable name is based on the name that you assign to the input control. For example, if you assign the name First prompt to an input control, the variable name is first_prompt. The variable name is lowercase, with an underscore character instead of a space. For more information on these methods, see Virtual Agent scripts.
Figure 1. The Variable name field is populated automatically
The Variable name field is populated automatically under Text input.

Large Language Model (LLM) user input controls

The following user input controls are supported in LLM topics when you create an LLM Virtual Agent topic. LLM user inputs include the following.
Note: Avoid placing more than five user input controls in a single LLM topic, or more than five in a single Input Collector. Excessive user input controls can overwhelm the LLM and may lead to inaccurate or unexpected responses.

Natural Language Understanding (NLU) entity properties

If NLU is enabled, the following properties are displayed in the property sheet for the Text, Static Choice, Dynamic Choice, Boolean, Date Time, and Carousel input controls. To view or enable these properties, go to the Advanced options section.
  • NLU entity: The entity that corresponds to the input variable for the control. You select the entity from the list of entities associated with the topic intent, as defined in the NLU model that you’re using for the topic. When Virtual Agent recognizes the entity value in user utterances and if the value meets the entity prediction confidence threshold, then Virtual Agent extracts the value. The value is stored in the input variable for the node. This storing action is called slot filling. Slot filling recognizes when multiple values for the same NLU entity occur in a conversation. Users are prompted with a list to confirm which node value to select when multiple nodes are tied to the same NLU entity. This property is available only for NLU topic discovery.
    Figure 2. NLU entity property
    Slide the NLU entity toggle to select an NLU topic when topic discovery is enabled.
  • Enable NLU to switch topics: Enables NLU prediction for the node. If enabled, users can enter text to answer questions, regardless of the type of input control being used. Virtual Agent uses this utterance to match another existing intent, letting the user switch topics. To enable this property on a node, navigate to Advanced options > Topic switching.
    Figure 3. Enable NLU to switch topics property
    Under Advanced options - Topic switching, slide the Enable NLU to switch topics toggle to switch topics property.
Default value
An input prompt can have a preset value that you define. This value can be either a string or a script that returns a string. This property is in the Advanced options section.
Figure 4. Default value property
Default value property under Advanced options to predefine a value for user input.
Confirmation messages
These bot response messages signal the end of the node interaction and ask the users to confirm their response. This property is in the Advanced options section.
Figure 5. Confirmation messages properties
Confirmation messages properties under Advanced options for input completion and default value confirmations.