Skip to content

Workflow Automation

Automation enables organizations to reduce repetitive tasks, increase operational speed and accuracy, and achieve more with fewer resources. With the Workflow Engine, teams can create simple or complex automations using a structured, configuration-driven approach — then deploy them wherever they are needed across the platform.

About Workflows and the Workflow Engine

The Workflow Engine provides a framework for defining and executing multi-step business processes. Rather than writing custom scripts for every routine operation, teams define workflows as a series of connected nodes — each representing a trigger or action — that execute in a determined order.

Using the Workflow Engine to automate routine operations such as creating follow-up tasks, sending notifications, or updating records makes work more efficient and productive, allowing employees to focus on higher-value strategic work.

Key Capabilities

  • Triggers: Define how a workflow starts. Manual triggers allow users to initiate a workflow on demand, with support for additional trigger types as your requirements evolve.
  • Actions: Perform work within a workflow. Code actions execute Python logic with access to workflow context, previous node outputs, and a curated set of built-in functions.
  • Context Management: Share data across workflow steps. Workflow-level context provides global state accessible to all nodes, while node-level context isolates data for individual steps.
  • Execution Ordering: The engine automatically resolves node dependencies using topological sorting, ensuring each step executes only after its prerequisites are complete.
  • API Access: List available workflows, retrieve configuration details, and trigger execution programmatically through dedicated API endpoints.

How It Works

A workflow is defined by three components:

  1. Nodes — Individual steps in the workflow, each with a type (trigger or action), a name, and optional configuration such as inline Python code.
  2. Edges — Connections between nodes that define the execution sequence and data flow.
  3. Metadata — Descriptive information about the workflow, including its name and identifier.

When a workflow is executed, the engine resolves the dependency graph, determines the correct execution order, and processes each node sequentially. Results — including success status, context data, and any errors — are returned as a structured execution report.

Example: Inventory Reorder Alerts

Consider an inventory management scenario where stock levels need to be monitored automatically. A workflow could be configured to:

  1. A manual trigger initiates the stock check process.
  2. A code action queries current inventory levels and compares them against predefined thresholds.
  3. A subsequent action generates a reorder notification and creates a purchase order record when stock falls below the minimum level.

Each step has access to the output of previous steps through the workflow context, enabling decisions based on accumulated data.

Testing and Measuring Success

Testing workflows before activation is essential to ensure they perform as expected and handle all possible scenarios. Thorough testing helps catch errors early, improve reliability, and prevent data issues in production.

Defining clear success criteria provides the foundation for measuring whether your workflows meet business requirements:

  • Specific: Define precise objectives. For example, "correctly flag reorder items when inventory drops below the configured threshold" rather than "handle inventory well."
  • Measurable: Establish quantitative metrics. Track execution success rates, processing accuracy, and error frequency across test runs.
  • Achievable: Base targets on realistic expectations informed by your workflow complexity and data volume.
  • Relevant: Align criteria with the business purpose of the workflow. An order processing workflow may prioritize accuracy, while a notification workflow may prioritize delivery speed.

Use sample data and varied scenarios to validate workflow behavior before deploying to production environments.