Your First 90 Days With an AI Management System

Starting an AI management system feels enormous. Shrink it to ninety days.

Days 1–30: Understand. Inventory every AI system you use, build, or buy — including the tools employees adopted on their own. Note what each one does, whose data it touches, and who owns it. Write your one-page AI policy and get leadership to sign it.

Days 31–60: Assess. Run a plain-English risk assessment on your highest-stakes systems: what could go wrong, how bad, how likely, what now? Set three to five AI objectives with owners and deadlines. Identify the competence gaps — who needs to learn what?

Days 61–90: Operate. Put basic controls in place for daily use: approved tool lists, human review where stakes are high, logging that actually happens. Start documenting decisions as you make them. Hold your first management review: what did we learn, what changes?

That is a working management system. Not a perfect one — a real one, sized to your organization, improving on purpose.

After day 90, the job becomes rhythm: monitor, audit yourself, review, improve. The loop takes over.

You do not need to boil the ocean. You need ninety days, an honest inventory, and the discipline to start. The best system is the one that exists.

I explain frameworks; I don't give legal counsel.