What to prepare for an AI workflow automation discovery
Bring a concrete process, constraints and acceptance criteria to the first scoping conversation.
Describe one real workflow
Choose a process with a clear starting event and a recognizable result. Map the people, systems and handoffs involved. Record where delays or rework occur using examples you are authorized to share. A specific workflow gives the scoping discussion a better foundation than a broad request to automate everything.
- Name the input, owner and expected output.
- Describe how exceptions are handled today.
- Remove private customer information from sample material.
Set operating boundaries
Explain which actions may run automatically and which need human review. Include access constraints, data retention needs and existing approval rules. Distinguish read-only reporting from changes to external systems. These boundaries influence the integration design, testing plan and recovery procedure.
- List the connected systems and required permissions.
- Identify irreversible actions and their approval points.
- Define what the automation should do when a dependency is unavailable.
Agree measurable acceptance criteria
Define what a successful pilot must demonstrate before discussing expansion. Include correct outputs, delivery, exception handling and rollback. Use the current Zion services and plans pages for the present offering; the discovery process establishes the specific scope and commercial terms. This guide does not describe a prepackaged industry product.
- Choose representative success and failure examples.
- Agree who reviews the pilot evidence.
- Document the accepted scope before connecting production actions.