The 100-Day AI Leadership Plan

Most "first 100 days" memos read like onboarding plans for a job that does not exist yet. AI leadership is different — the job description gets rewritten every quarter, and the calendar is the only thing that holds steady. So I structure the first 100 days as four overlapping arcs: Learn, Align, Deliver, Review — each thirty days long, each starting before the previous one ends.

Learn (day 1–30) is unfashionably slow. You are not there to ship; you are there to find the floor. Map the data flows, sit in on three operations reviews, talk to the four people whose lives will change most when AI lands. Align (day 20–50) is where you trade preferences for principles — build-vs-buy, foundation-vs-custom, RAG-vs-fine-tune — and write them down in language a board member can repeat. Deliver (day 40–80) is where the principles meet a roadmap, but only with use cases you can describe in one sentence and measure in one number.

Review (day 70–100) is where most plans rot. The review is not the steering committee — it is the public moment where you say which assumptions were wrong. Make that moment loud, make it specific, and put the revised plan on the same page as the original. Everything you do in year two will depend on whether the org believes you can do that without flinching.