How to choose the first workflow to automate

What this solves

Teams often choose the first workflow emotionally, not commercially, which makes AI feel disappointing or risky.

• Nobody agrees what first success should look like.
• The first use case is chosen because it seems advanced.
• The workflow touches too many edge cases too soon.

What AI is good at here

• Helping compare candidate workflows objectively.
• Highlighting which tasks are repetitive and which are exception-heavy.

What AI should not own

• Leaders still define success criteria and risk tolerance.
• AI should not choose priorities without commercial context.

Best first implementation

• Score candidate workflows on repetition, value, clarity, and error cost.
• Choose the cleanest workflow with visible upside, not the broadest one.

What good looks like

• The first automation proves something commercially real.
• The team gains trust from a contained win.

Common mistakes

• Starting with a workflow that crosses too many teams.
• Picking a task with unclear ownership or poor source data.