Capstone: Personal AI Playbook

This is the universal capstone. Everyone builds a Personal AI Playbook regardless of which other capstone they choose. It’s your portable toolkit — the collection of what actually works for you, refined through real use.

The playbook isn’t built in one sitting. It’s assembled as you progress through the levels and capstones, capturing what you learn along the way.

What Goes in the Playbook

1. Prompt Library

Reusable prompts that you’ve tested and refined for your specific work:

Prompt: "Help me create a template for organizing my prompt library. 
I want to store reusable prompts with:
- Name (what it does in 3-5 words)
- Category (communication, analysis, research, planning, etc.)
- The prompt itself (with [placeholders] for variable inputs)
- When to use it (trigger conditions)
- Tips (what I learned about making it work well)
- Last tested date

Create the template with 3 example entries based on common 
workplace tasks."

2. Workflow Recipes

Multi-step processes you’ve built (from capstones or daily work):

Prompt: "Create a template for documenting an AI workflow recipe. 
Each recipe should capture:
- Workflow name
- Purpose (what problem it solves)
- Steps (numbered, with the prompt for each step)
- Input needed (what you start with)
- Output produced (what you end up with)
- Time estimate
- Common pitfalls and how to avoid them

Create the template with one example: a workflow for turning raw 
notes into a polished document."

3. Lessons Learned

What you’ve discovered about working with AI effectively:

Prompt: "Help me create a 'lessons learned' template for tracking 
what I discover about using AI effectively. Each entry should have:
- Date
- What I tried
- What worked / what didn't
- The insight (one sentence I'd tell a colleague)
- Category: prompting technique, tool discovery, workflow 
  improvement, mistake to avoid

Create the template with space for entries and include 3 starter 
entries based on common beginner discoveries."

Exercise 1: Seed the Playbook

Start by capturing what you already know:

Prompt: "I'm creating a personal AI playbook — a reference doc of 
prompts, workflows, and lessons learned. Help me seed it with 
content from what I've learned so far.

My role: [your role]
AI tools I use: [list them]
Tasks I use AI for most: [list 3-5]
Biggest win so far: [describe]
Biggest frustration: [describe]

Based on this, generate:
1. My top 5 reusable prompts (based on my common tasks)
2. A starter workflow for my most frequent AI-assisted task
3. 3 lessons learned I've probably already discovered
4. 3 areas I should explore next"

Industry variant (healthcare): Seed with healthcare-specific prompts — patient communication templates, compliance-safe research workflows, policy summarization for staff training.

Exercise 2: Capture from Capstone Work

After completing another capstone project, extract the reusable parts:

Prompt: "I just completed a capstone project: [name — e.g., Expense 
Report Builder]. Here's what I built:

[paste your key outputs — templates, workflows, prompts]

Extract the reusable components for my playbook:
1. Any prompt that could be used for other purposes (generalize it)
2. The workflow pattern (abstract from the specific project)
3. Techniques I used that apply broadly
4. Mistakes I made and how to avoid them next time"

Repeat this after each capstone or significant AI project.

Industry variant (healthcare): Extract compliance-aware patterns — which prompts needed de-identification, which workflows required human review gates.

Exercise 3: Teach It to Someone Else

The best test of understanding is teaching. Pick your strongest playbook entry:

Prompt: "I want to teach a colleague how to use this workflow:

[paste your best workflow recipe]

Create a 5-minute walkthrough I could present or share:
1. The problem it solves (why they'd care)
2. Step-by-step demo with a concrete example
3. Common mistakes and how to avoid them
4. How to customize it for their own work

Keep it conversational — like I'm showing them over coffee, not 
presenting at a conference."

Actually share it. Post in your team channel, show a colleague, bring it to your next 1:1. Teaching is the highest-leverage fluency activity.

Industry variant (healthcare): Create the walkthrough for a clinical colleague — emphasize the safety guardrails alongside the productivity gains.

Exercise 4: Audit and Improve

Periodically review your playbook for quality:

Prompt: "Here's my current AI playbook:

[paste your full playbook or key sections]

Audit it:
1. Which prompts are probably outdated? (AI tools evolve fast)
2. Which workflows have steps that could be simplified?
3. What's missing? Based on my role as [role], what common tasks 
   don't have a prompt or workflow yet?
4. Which lessons learned should I promote to 'rules I always follow'?
5. Rate each prompt: still useful / needs update / retire"

Do this monthly. A stale playbook is worse than no playbook — it teaches bad habits.

Industry variant (healthcare): Include a compliance audit of your playbook — are any prompts at risk of inadvertently processing sensitive data?

Final Deliverable

A living document containing:

1. Prompt library — Tested, categorized, with usage notes
2. Workflow recipes — Multi-step processes for recurring tasks
3. Lessons learned — What works, what doesn’t, and why
4. Teaching materials — At least one walkthrough you’ve shared with others

Reflection:

• Which playbook entry saves you the most time per week?
• What’s the one prompt you’d recommend if someone could only learn one?
• How has your playbook changed since you started? (That change IS your fluency slope.)

Keeping It Alive

The playbook isn’t done when the program ends. It grows with you:

• Weekly: Add any new prompt or technique that worked well
• Monthly: Audit for stale entries, update what’s changed
• When you learn something new: Capture it immediately — the insight fades fast
• When you teach someone: Add their questions and your answers — they reveal gaps

The best playbooks are messy, personal, and constantly evolving. Don’t optimize for polish — optimize for usefulness.

Industry variant (healthcare): Seed your playbook with prompts specific to your clinical or administrative context — patient communication templates, compliance-safe research workflows, de-identified case study generators. Include a “compliance checklist” section in your playbook that you reference before every AI interaction involving sensitive domains.