Start from roles, not tools
AI changes a relationship manager's day differently from an engineer's or a recruiter's. A program built around the tool teaches everyone the same buttons. A program built around roles shows each person what changes in their own work.
Start by listing the roles in scope and, for each, the two or three tasks where AI makes the biggest difference.
Foundation: the rules and the concepts
Everyone needs to know what AI can and can't do, and the company's rules for using it. Most people also need to understand why a prompt works and why client data stays out.
Agents teach this to each person's gaps. Someone who already knows the rules skips them. Someone who doesn't, learns them before they practice.
Practitioner: safe practice
The sandbox is where people try AI on their own tasks without risk. They brief an assistant, check its answer, and keep sensitive data out. The coach shows what changed between attempts.
Nothing passes while client data gets through. That single rule prevents most of the incidents organizations worry about.
Proficient: real tasks
The Field agent finds the moments in each person's week where AI helps: the meeting to prepare, the report built by hand, the draft that needs checking. Each becomes a mission, done for real and debriefed.
Measuring readiness
Licenses and logins tell you who has access. Levels tell you who can use AI well in their work. Report readiness by team and role, so leaders can see where the organization is ready and where it needs support before the next rollout.