Configure Dialouge Tree as a Task
Updated
Dialogue Tree as a Task enables you to design deterministic, branching conversational flows that are often required in compliance-mandated environments. Unlike free-flowing AI conversations, dialogue trees ensure predictable and controlled interactions.
This task type provides direct access to key capabilities such as Zero-Shot Model configuration for intent detection without large training datasets. It supports deterministic branching to ensure compliance and predictable user journeys. Intents are integrated into the flow and drive the branching logic.
Note: It is mandatory to configure Intents inside the User Replies node of the Dialogue Tree to proceed with the flow.
To learn more about creating intents, refer to the guide: Intents in Dialogue Tree as a Task.
Steps to Create a Dialogue Tree as a Task
Follow the steps below to create a Dialogue Tree as a Task:
Step 1: Add a Dialogue Tree Task
Navigate to the Tasks manager.
Click the caret icon beside + Add Task in the top-right corner.
Select Dialogue Tree as a Task from the dropdown.
Step 2: Define Task Details
On the Create New Task window, enter the following details:
Name: Enter a unique name for your task (e.g., Customer Verification Flow).
Description: Provide a short description of the task’s purpose.
Click Save in the bottom-right corner.
Step 3: Configure Dialogue Tree Canvas
On the Dialogue Tree canvas, click the + icon to add a User Replies node.
In the Customer Says configuration screen:
Name: Enter a descriptive name for the node (e.g., User Response for Verification).
Intents: Select the intent(s) you want to associate with this node. Note: Intents must be preconfigured in your ML Model. Refer to Intents in Dialogue Tree as a Task for details.
Continue adding nodes as required to build your branching flow.
Step 4: End the Flow
Every dialogue tree must end with an End Dialogue Tree node to ensure proper termination of the conversation.
Step 5: Save the Task
Click Save to finalize your Dialogue Tree Task.
Important Notes
Fallback Handling: If a fallback occurs, the flow automatically redirects to the default Router task.
Handover: You can hand over the Dialogue Tree flow to another task using the Handover to Another Task node.
Routing Restrictions: Routing to another bot is not supported.
Model Support: Only Zero-Shot models are supported for intent detection in Dialogue Tree tasks.