AI Start Here — What Should I Automate First?

List your AI ideas. Get each one scored, a do-not-start list, and a 30-day pilot brief for the best first project.

Everyone on your team has an AI idea. Picking the wrong first one wastes a quarter and makes people give up on AI. What Should I Automate First? scores every idea the same way and tells you which one to start, which to park, and which should never be automated at all.

What you get in one run:

• A Pilot Fit Scorecard: every idea scored 1–5 on value, frequency, data readiness, risk and effort, with the math shown
• A do-not-start list, including ideas that need a simple template or report instead of AI
• One recommended first pilot, and why it beats the runner-up
• A 30-day pilot brief: scope, human check, baseline, success threshold, stop rule and week-by-week plan
• What to verify before you commit

Works best with 3–12 ideas and rough numbers (how often, how long, where the information lives). Without numbers you still get a ranking, clearly labeled where scores are estimated.

Not for: choosing software vendors, building the automation, or guaranteed savings. Uses that make decisions about hiring, credit, health or legal matters are flagged for professional review.

How to start

List 3–12 tasks you're thinking of handing to AI or automation, and the one goal that matters most (e.g., "reply to customers within 8 hours"). I'll score each task on value, frequency, data readiness, risk and effort, tell you which ones not to start (some won't need AI at all), and turn the best one into a 30-day pilot plan. Add how often each task happens and how long it takes for a sharper ranking.