The Abort node is a Flow Action node in AI+ Studio deployment pipelines that stops workflow execution and returns a custom message to the user. The Abort node gives you control over what the user sees when a workflow is interrupted. Without this node, users may not receive clear feedback when execution stops.

It is especially useful in scenarios where:

• Guardrails block input or output
• A validation condition fails
• You want to enforce strict workflow rules
• Errors occur, but you still want to return a user-friendly message

Guardrails can block requests and return fallback messages when policy violations are detected. The Abort node lets you define and control these messages directly within the pipeline. 

How it Works

The Abort node works by combining two components:

• Update Properties node: Creates and assigns the message to a variable
• Abort node: Reads the variable and returns the message while stopping the pipeline

This design separates message creation from execution control, which allows you to reuse logic and maintain consistent messaging across workflows.

When triggered, the Abort node immediately stops further execution of the pipeline and returns the configured response. Any nodes placed downstream are not executed. The behavior is deterministic and ensures that once a workflow is aborted, no additional processing occurs.

Key characteristics include:

• Stops pipeline execution immediately
• Returns a message stored in a variable
• Does not execute downstream nodes
• Logs the abort event in Audit Trails

Abort Node Configuration

Follow the steps below to add Abort Node in your AI Deployment Pipeline:

• Navigate to your AI Use Case Deployment Pipeline. 

• Either open an existing pipeline or create a new one. 

• Click the ‘+’ button on the pipeline canvas. 

• Select Abort under Flow Actions. 

• The Abort node configuration screen will appear. 

• Select the Abort Variable form the dropdown that contains the message to display. This variable must be populated before the Abort node (using Update Properties node).

• Click ‘Save’ to save your Abort node configuration.

Example Flow

A common use case involves guardrail handling:

1. A request is processed in a Prompt node
2. A guardrail detects restricted content
3. The pipeline continues if Continue Pipeline post Error toggle is enabled and you have configured the Error Varibale in Prompt Node.
4. An Update Properties node sets a custom blocked message
5. The Abort node is triggered
6. The workflow stops and displays the message

Example output: “Your request cannot be processed due to policy restrictions.”

​

​

Behavior in Audit Trails

The Abort node is fully logged in Audit Trails, which helps you monitor and debug workflow execution. When a workflow is aborted, the system records the event and marks the request as stopped.

Audit Trail details include:

• Request status as Stopped
• Abort node entry in debug logs
• The selected Abort Variable
• The final message returned to the user

This visibility helps you identify why a workflow stopped and what response was delivered.

​

Use the Abort node when you need to stop execution and return a controlled response.  You can configure the Abort node in the following scenarios:

• Handling guardrail-blocked input or output

• Enforcing validation or business rules

• Stopping execution on critical errors

• Returning predefined fallback messages

Best Practices

• Always define the message in an Update Properties node before the Abort node
• Use clear and user-friendly messaging for blocked scenarios
• Avoid hardcoding messages across multiple nodes
• Use consistent variable naming (for example, abort_message)
• Test fallback and abort paths using pipeline testing