Skip to main content
The Start Integration Workflow action is available on Send Message Task, Task, and User Task nodes within BPMN processes.

Overview

The Start Integration Workflow action allows a BPMN process to trigger an integration workflow, send input data, and receive the workflow output at a designated result key. This enables processes to delegate data fetching, transformations, and external system integrations to lightweight workflows without creating subprocesses.
For triggering workflows from UI Flows (without a BPMN process), see the Start Workflow UI Action.

How it works

When a Start Integration Workflow action is triggered:
1

Input data is sent

The input data mapped in the action configuration is sent as start variables to the workflow.
2

Workflow executes

The integration workflow runs with the provided inputs, executing its configured nodes (REST calls, data transformations, AI operations, etc.).
3

Output is captured

Workflow output data is captured on the specified result key in the process instance upon completion.
4

Process continues

The BPMN process continues to the next node after the workflow completes.
The process waits for the workflow to complete before continuing. To receive output data, you must add a Receive Message Task node after the Send Message Task that triggers the workflow.

Configuration

Selecting a workflow

The Select Workflows dropdown displays:
  • All workflows within the current application version
  • Any workflows referenced in the application (e.g., via the Library)

Input mapping

Map process variables to workflow input parameters:
{
  "customerId": "${processInstanceData.customer.id}",
  "requestType": "credit_check"
}

Result key

Configure the result key where workflow output data will be stored in the process instance variables after completion.

Receiving output

To capture workflow output in your process:
  1. Add a Receive Message Task node after the node with the Start Integration Workflow action
  2. The workflow output is automatically mapped to the configured result key
  3. Subsequent nodes can access the output data via the result key path

When to use

ScenarioRecommended Action
Call external APIs and return dataStart Integration Workflow
Run data transformationsStart Integration Workflow
Execute AI operations (extraction, generation)Start Integration Workflow
Multi-step user journeysUse subprocesses instead
Processes requiring state managementUse subprocesses instead

Last modified on February 27, 2026