Workflow automation

Workflow automation uses AI or AI-generated code to remove predictable, repetitive work when triggers, inputs, outputs, exceptions, and ownership are clear.

• Identify a high-value workflow to automate, for example by following our AI Adoption Guide
• Make sure this is a predictable, repetitive workflow
• Identify on trigger, inputs, outputs, exceptions, failure visibility, and ownership.

• Do not use when nobody agrees how the manual process works.
• Do not use if there are existing tools that can already do it.
• Do not use if the data source or destination has no API.

• Capture trigger, inputs, steps, outputs, exceptions, and failure impact.
• Use a tool like Zapier or n8n with existing connectors and AI nodes.
• Start with internal workflows before tackling external ones.

• Do not automate broken work.
• Use deterministic rules where rules are enough.
• Assign a real owner after launch.
• Related: Agentic Engineering