AI Workflow Readiness Checklist

Most AI budgets are wasted because the work starts with a model rather than a business problem. This checklist takes one workflow and tests whether it is ready: is the problem measurable, is the real process understood, can the data lawfully be used, and can accountable people control the outcome.

Twelve items across four steps, each asking for evidence rather than an opinion. It ends with three decisions: ready for a pilot, prepare the workflow first, or do not automate this yet.

What it covers

• Define the business problem and baseline before any model is chosen
• Map the real process, including the workarounds nobody documented
• Confirm lawful data use, ownership, and who can pause the system
• Set acceptance criteria, a stop threshold, and a decision date upfront
• Three outcomes with a clear rule for which one applies to you