Most AI rollouts don't fail because the technology fell short. They fail because someone scaled before they were ready — before the judgment, the workflow, or the guardrails could carry the weight. This course is the reckoning you do first: an honest look at you, your workflow, and your fences, with a score you can retake as you grow.
Readiness lives in three places. All three have to hold — the weakest one sets your speed limit.
Can you tell a good output from a plausible one in this domain? Do you know what great looks like before the tool shows you anything? AI amplifies judgment. It does not supply it. If you can't grade the work without the tool, you can't safely delegate it to the tool.
Is the task defined enough to hand to anyone — a clear start, a clear finish, a definition of done, a worked example to compare against? A task you can't describe is a task you can't delegate, to a person or a machine.
Do you know what never leaves your hands? Does everything that does leave have a named check? Do you know, in advance, what pulls a task back to you? If the guardrails live in your head, they don't exist under pressure.
| Step | What it looks like |
|---|---|
| 1. Fence first | Draw your Human-Owned Zone map before you scale anything. Know your center, your never-delegate list, and the named check on everything that leaves. |
| 2. Pilot small | One task, real work, low stakes, full attention on the checking. Run it for weeks, not days. Keep score. |
| 3. Scale what survived | Only what held up under real checking earns expansion — one ring outward, one task at a time. Every expansion is earned by track record, never granted by enthusiasm. |
Twelve statements, four per dimension. Score each honestly: 0 = no, 1 = somewhat, 2 = yes.
| I can spot a wrong-but-confident output in my domain. | 0 · 1 · 2 |
| I know what a great result looks like before I start. | 0 · 1 · 2 |
| I can explain why an output is good or bad, not just feel it. | 0 · 1 · 2 |
| I've done this task well myself, without AI. | 0 · 1 · 2 |
| The task has a clear start, finish, and definition of done. | 0 · 1 · 2 |
| I have a worked example to compare outputs against. | 0 · 1 · 2 |
| I know exactly where AI's part starts and stops. | 0 · 1 · 2 |
| Mistakes get caught before they reach anyone who matters. | 0 · 1 · 2 |
| I have a written never-delegate list. | 0 · 1 · 2 |
| Every delegated task has a named, written check. | 0 · 1 · 2 |
| I know in advance what pulls a task back into my hands. | 0 · 1 · 2 |
| I could comfortably explain my setup to the person most affected by it. | 0 · 1 · 2 |
| Total | Where you are |
|---|---|
| 0–9 | Not yet — and knowing that puts you ahead of most. Build the fundamentals: judgment reps, one defined task, a written fence. Retake in a month. |
| 10–17 | Ready to pilot. One task, low stakes, full checking. Your lowest-scoring dimension is your homework. |
| 18–24 | Ready to scale what's proven. Expand one ring at a time, and keep retaking — readiness is a moving target because your work keeps moving. |
Date:
You: / 8 Workflow: / 8 Guardrails: / 8
Total: / 24 Band:
My weakest dimension and this month's homework:
Acute Incite · powered by theCommons Academy · The Readiness Reckoning · Find what helps. Fence what's risky. Forge your own way of working.