Skip to main content
Version: Current

Invoke Agent

Description​

The Invoke Agent activity starts another Agent from the current workflow and returns the Automation ID for that execution.

Use this activity to delegate work to another Agent, pass input values to its main workflow, and choose whether the current workflow waits for completion.

Common Capabilities​

Process Data Support​

info

This activity supports dynamic configuration using variables from the Process Data drawer.

You can bind values from Model Data, Form Data, System Data, Enterprise Variables, and Activity Outputs.

Learn more → Using Process Data

Design Time Configuration​

Display Text​

Specify the name displayed for the activity on the Agent Builder canvas.

Agent Name​

Choose the Agent that should be invoked.

Automation Name​

Specify the name of the Automation instance created when the Agent is invoked. Use a meaningful name to help identify the execution when monitoring Automations.

Execution Strategy​

Choose the appropriate execution strategy. The following options are available.

OptionDescription
SynchronousInvokes the Agent and waits for its execution to complete before continuing the current workflow.
AsynchronousInvokes the Agent and immediately continues the current workflow without waiting for the Agent to complete.

Input Parameters​

Displays the input parameters configured for the selected Agent's Main Workflow in the Workflow Parameters dialog.

Map each input parameter to a value from the current workflow. Use values that match the data type of the corresponding parameter.

Timeout​

Specify the maximum duration the current workflow waits for the invoked Agent when Synchronous execution is selected.

Wait All Incoming​

Execution Control

Controls how this activity behaves when multiple incoming execution paths converge.

Determines whether the activity executes when:

  • Any incoming path completes.
  • All relevant incoming paths have completed before execution.

Learn more → Execution Control

On Error​

Error Handling Strategy

Defines how the workflow behaves if this activity encounters a runtime error.

Supported strategies include Fault, Continue, Suspend, and Retry. Select Inherit to use the workflow-level error-handling policy.

Learn more → Error Handling

Outputs​

Automation ID​

Returns the unique identifier of the Automation instance created for the invoked Agent.

Bind this value from Process Data Explorer → Activity Outputs → [Invoke Agent activity name] → Automation ID to the Automation ID field of Get Agent Output or Get Automation Info.

Remarks​

  • Each Automation ID identifies a specific execution of the selected Agent.
  • For Asynchronous execution, continuing the current workflow does not indicate that the invoked Agent has completed. Arrange output retrieval after the invoked execution completes when downstream steps require its final outputs.
  • Use Get Agent Output to retrieve the execution status and main workflow outputs, or Get Automation Info to retrieve execution metadata.
  • If the invoked Agent cannot be started, the activity does not return an Automation ID and reports the corresponding error.

Example​

A parent workflow delegates order processing to an Agent named Process Order.

  1. In the Process Order Agent's main workflow, configure an input parameter named OrderNumber and an output parameter named Message in the Workflow Parameters dialog.
  2. Populate Message using Set Workflow Output before the main workflow completes.
  3. In the parent workflow, add Invoke Agent, select Process Order, and map OrderNumber to the current workflow's order number.
  4. Specify an Automation Name, select Synchronous, and configure Timeout for the expected execution duration.
  5. After the invoked Agent completes successfully, use Get Agent Output with the same Agent and the returned Automation ID to retrieve Message.
  6. Use Get Automation Info with that Automation ID when execution metadata is also required.
Help improve this page