Most AI strategies are either a slide of possibilities or a pile of tool subscriptions. A useful 90-day plan picks a few workflows, sets the rules that make AI safe to use, and ships one measured result — so the next plan is based on evidence rather than enthusiasm. These three phases fit the work into one quarter.

Days 1–30 — Assess and choose

Goal: Know where AI is used today and pick where to focus.

  1. Inventory current AI tools, accounts and informal use across teams.
  2. List candidate workflows and score them on value, readiness and risk.
  3. Choose one priority workflow to ship and two to prepare.
  4. Name an executive owner for the plan and an owner for each workflow.

You finish with: An inventory, a scored shortlist and named owners.

Days 31–60 — Set the foundation and build

Goal: Make AI safe to use and start the first workflow.

  1. Publish a short AI use policy: approved tools, data rules and review points.
  2. Configure the approved workspace for the priority workflow’s team.
  3. Record the baseline for the priority workflow before changing it.
  4. Build the smallest version with approval points and a stop rule.

You finish with: A published policy, a configured workspace and a first workflow in parallel run.

Days 61–90 — Measure, report and plan the next quarter

Goal: Prove the result and decide what comes next.

  1. Compare the priority workflow against its baseline.
  2. Report results, costs and lessons to leadership in one page.
  3. Decide whether to expand, adjust or stop, with reasons.
  4. Set the next quarter’s workflows and owners from what the first one showed.

You finish with: A measured result, a leadership report and the next 90-day plan.

Common pitfalls

  • Starting with a tool purchase instead of a workflow.
  • A plan with no executive owner.
  • Skipping the policy, so people keep using personal accounts.
  • Reporting activity instead of a measured result.

Read as Markdown Fractional Chief AI Officer