Adding a Dialogue Tree Tool to a task automatically initiates the associated dialogue tree. The tree runs in the background to ensure the interaction follows the predefined conversational flow. Dialogue tree nodes can be used to execute hardcoded workflows.

Procedure

• In the Tools window, select Add Tool (top-right corner) and choose Dialogue Tree from the dropdown list.

  

• In the Dialogue Tree Tool window, enter a Name and Description.

• Under Setup Function, select + Create Dialogue Tree to begin building the dialogue tree.

• Under Select inputs to capture, define the Input Variables that will be passed into the tree.

• Define an Output Variable that returns context to the LLM about the dialogue tree execution.

  

• When the dialogue tree is complete, select Save (bottom-right corner).

Configure a Follow-up Tool for a Dialogue Tree

When a Dialogue Tree is configured as a tool, you can optionally define a Follow-up Tool that runs automatically once the Dialogue Tree completes. This enables deterministic workflows by executing the next predefined action without requiring the AI Agent to invoke an LLM to determine the next step.

The Follow-up Tool setting is available within the End Dialogue Tree node. Once configured, the selected tool is triggered immediately after the Dialogue Tree finishes execution.

To configure a Follow-up Tool

​

1. Open the required Dialogue Tree that is configured as a tool.

2. Hover over the End Dialogue Tree node and click the Edit Element icon.

3. Configure the Tool Output to be returned when the Dialogue Tree completes.

4. In the Follow-up Tool (Optional) field, select the tool that should run after the Dialogue Tree ends.

5. Click Save to apply the configuration.

How Follow-up Tools Work

When a Follow-up Tool is configured:

• The selected tool is automatically executed when the End Dialogue Tree node is reached.
• No intermediate LLM call is required to determine the next action.
• Workflow latency is reduced, resulting in faster execution and improved responsiveness, particularly for Voice AI Agents.
• The capability is best suited for deterministic workflows where the next action is known in advance, such as calling an API, updating records, routing conversations, or triggering downstream processes.
• If no Follow-up Tool is configured, the standard execution flow continues and control is returned to the AI Agent after the Dialogue Tree completes.

​