Copied to clipboard!
Module 12 Foundations intermediate 50 min

Your AI playbook: write it now

Capstone

What you'll be able to do

  • Produce a one-page playbook covering what to delegate, which models to use, and which checkpoints stay human
  • Convert Modules 1 through 11 into a single durable artifact
  • Leave the track with something you would actually pin above your desk and reuse

Write Your One-Page AI Playbook

This is the point of the whole track. Not a certificate, but a one-page playbook you would pin above your desk: what you hand to AI, what you keep fully human, which tool you reach for, where your context lives, and the checkpoints that never move. In this capstone you turn everything from Modules 1 through 11 into a durable artifact in your own voice. The AI facilitates; you decide every line, because this is your playbook.

Recommended tool

claude

Claude is good at facilitating one question at a time and then formatting a clean one-page document in your voice, which is exactly what turning eleven modules into a single keepable artifact requires.

Bring to the exercise

  • Your AI Context Document
  • Five to eight tasks you do most often
  • Your model chooser from Module 3 and your human-stays-loud rules from Module 6
  • The reusable prompts and checkpoints you built along the way
Step 1

Sort Your Tasks Into Three Buckets

Start where Module 6 left off: decide what to delegate, what to collaborate on, and what stays fully human.

What to substitute before pasting

  • [PASTE YOUR CONTEXT DOCUMENT] Your one-page AI Context Document.
  • [LIST 5-8 TASKS] The tasks you genuinely do most, not the ones that sound impressive.
Be my facilitator for writing a one-page AI playbook I will actually reuse. Ask one question at a time and wait for my answer.

Here is my AI Context Document:
[PASTE YOUR CONTEXT DOCUMENT]

Tasks I do most:
[LIST 5-8 TASKS]

For each task, help me place it in one of three buckets and say why:
- DELEGATE: hand to AI, I review
- COLLABORATE: AI and I work it together
- KEEP HUMAN: stays fully mine, AI does not decide

If a task could hurt a person, is binding, or is something I must sign, push me to justify before it leaves the KEEP HUMAN bucket.
What good output looks like
  • Every task lands in exactly one bucket with a reason.
  • Anything binding, signable, or harmful is pushed toward KEEP HUMAN.
  • It asks one question at a time instead of dumping the whole thing.

Verify before using AI's output

Step 2

Write Your Model Chooser

Pull Module 3 into a personal chooser: which tool you reach for, for which kind of task, by default.

Now build my personal model chooser. Based on my tasks and context, give me a short, plain-language rule for which tool I reach for by default:

- Grounded in one of my own documents -> NotebookLM
- A long, careful draft in my voice -> Claude
- Quick help inside my Google tools -> Gemini
- A general brainstorm -> whichever is handy

Adjust these rules to MY actual tasks and name the one tool I should stop defaulting to out of habit.
What good output looks like
  • A short chooser mapped to your real task types, not a generic chart.
  • It names a default-by-habit you should break.
  • Each rule is one line you could remember.

Verify before using AI's output

Step 3

Pin Down Your Context and Checkpoints

Capture where your context and reusable prompts live, and the human checkpoints that never move, from Modules 4, 6, and 10.

Help me write two short sections of the playbook.

1. WHERE MY STUFF LIVES: where my AI Context Document and my reusable workhorse prompts are saved, so future-me can find them in seconds.

2. CHECKPOINTS THAT NEVER MOVE: the points where I verify before AI work goes out. At minimum cover: before money, before sending, before signing, before filing. For each, one line on exactly what I confirm.

Keep both sections tight enough to fit on one page.
What good output looks like
  • Names a concrete home for your context doc and prompt drawer.
  • Lists the non-negotiable checkpoints with a one-line check each.
  • Stays short enough to live on a single page.

Verify before using AI's output

Step 4

Add the Working-Check and a Revisit Date

Close the loop with Modules 9 and 11: how you will know it is working, and when you will revisit it.

Add a short final section: HOW I WILL KNOW THIS IS WORKING.

Include:
- The cheap evaluation habit I will keep (from my eval harness) and roughly how often I will run it
- One or two signs that a model or tool choice has stopped being the right call
- A specific date to revisit and update this whole playbook

Keep it to a few lines. This is the part that stops the playbook from going stale.
What good output looks like
  • Names a real, repeatable evaluation habit and a rough cadence.
  • Lists signs that a tool choice has gone stale.
  • Sets a concrete revisit date.

Verify before using AI's output

Step 5

Format the One-Pager

Have the AI assemble everything into one clean page in your voice, the artifact you walk out with.

Pull everything together into ONE clean, one-page AI playbook in my voice. Use these sections:

1. DELEGATE / COLLABORATE / KEEP HUMAN (my tasks in buckets)
2. MY MODEL CHOOSER
3. WHERE MY STUFF LIVES
4. CHECKPOINTS THAT NEVER MOVE
5. HOW I WILL KNOW THIS IS WORKING (plus my revisit date)

Format it as clean Markdown I can save and pin. Plain and direct, no hype, no filler. If any section is thin, tell me what to add rather than padding it.
What good output looks like
  • One page, all five sections, in your voice.
  • Clean Markdown you could save and reread in a month.
  • It flags thin sections instead of padding them.

Verify before using AI's output