The Six Questions

The six questions are the decision discipline to run — in order — before committing real money or real weeks to any AI build: (1) Name the problem in your own words. (2) Is the workflow real — counted, not felt? (3) Is this AI or is this automation? (4) Is the technology actually ready? (5) Where does the human go? (6) Prototype by hand before you productionize. They won’t make every project succeed; they catch most of the ones that were going to fail, before they cost anything.

Why it matters: These are decision questions, not technical ones — a business owner who can’t evaluate a machine-learning model can absolutely evaluate whether spending sixty grand to order less lettuce makes sense, and that’s the only question that was ever theirs to answer. The order matters too: each question assumes the ones before it were answered honestly, which is why vendor-shaped problem statements (question one) quietly break everything downstream.