Goal
You map one normal working day and find the five moments where Copilot earns its keep — in your actual work, not in theory.
Why this matters
“Minutes saved per employee” is an average that hides everything important. Value lives in specific recurring tasks in specific roles. This map replaces the vague estimate with a concrete, role-specific picture — and it scouts the task that becomes your agent in week 4.
The drill
Take one normal working day. Not your best, not your worst. A Tuesday.
- List it out. Block by block, write what you actually do. Email, the 9:00 status call, the vendor proposal, the board update, the 1:1s. Be honest and specific — aim for 12 to 20 activities.
- Tag each one. Next to every activity, one of three tags: Does it (Copilot could handle most of it — summarizing, drafting, first-pass analysis), Helps it (faster or better with Copilot, you stay in the driver’s seat), Stays out (human-only: the decision, the relationship, the judgment call).
- Rank what’s left. For everything tagged “does it” or “helps it”, score frequency (daily, weekly, monthly) and friction (quick, annoying, painful). Frequency times friction is your leverage score.
- Pull your top 5. The five highest-scoring moments are where Copilot earns its keep in your day.
Extra round
Read your own map. The top is usually high-frequency, medium-friction work — that’s normal; boring recurring tasks are where AI quietly compounds. The “stays out” column matters just as much: it shows you where not to use Copilot. And the scouting note: find the one task in your top 5 that’s most repetitive and most rule-based. Circle it. In week 4, that’s your first agent.
Micro-win
You have a one-page map that’s worth more than any minutes-saved slide — because it’s yours and it’s specific.
Remember
Don’t hand your map to your team later. Help them build theirs. That’s the difference between adoption that sticks and adoption that’s mandated and resented.