🏆 LEVEL 2 CAPSTONE

Build Your AI-Powered Work System

Apply everything from Level 2. In ~90 minutes you'll have a complete AI workflow that transforms how you work every day.

1 SystemBuilt & Tested
90 minYour Time
Real WorkApplied
RepeatableEvery Day

What You've Mastered

Seven high-impact skills from Level 2. Now you'll weave them into a single system.

🔍 Deep Research

Synthesize hours of research in minutes. Find signal in noise.

✓ Complete

✍️ AI Writing

Draft faster, edit sharper, match any tone or format.

✓ Complete

📊 Data Analysis

Understand what your data says without being a statistician.

✓ Complete

🧭 Strategy

Think through complex decisions and spot blind spots.

✓ Complete

🧠 Claude Projects

Persistent AI workspaces with your context and files.

✓ Complete

⚙️ Automation

Offload repetitive work to AI-powered workflows.

✓ Complete

🗂️ AI OS

Daily habits and a prompt library that compounds over time.

✓ Complete

Choose Your Capstone Track

Pick ONE track that fits your work. Each takes ~90 minutes and produces something real you'll use every week.

Track A: The Knowledge Worker

Build a complete research-to-report pipeline for a real topic in your field.

End result: A research brief + written report + one automation to keep it updated weekly.

Track B: The Team Leader

Build a manager's AI system: a Claude Project for team comms + 3 automations + a team guide.

End result: A documented AI system your team can actually use.

Track C: The Specialist

Go deep on your profession. Build a Claude Project with full professional context, 10 saved prompts, and one automation that saves 2+ hours/week.

End result: A personalized AI expert who knows your exact work.

Track D: The Builder

Create a "mini AI product" — a documented prompt chain that solves a specific problem, written as an SOP you could hand to a colleague.

End result: A shareable, reusable AI workflow with a live demo.

Track Instructions

Track A: The Knowledge Worker Pipeline

🖥️HANDS-ON EXERCISE90 min

Research → Analyze → Report → Automate

  1. Choose a real topic relevant to your work that you'd normally spend hours researching.
  2. Run the Research Sprint prompt (below). Get 8–10 synthesized sources in 15 minutes.
  3. Paste the research output back into Claude and ask it to identify the 3 key themes.
  4. Ask Claude to write a 500-word executive brief using those themes.
  5. Run the Data Analysis prompt if you have any relevant numbers or trends.
  6. Ask Claude to write one slide's worth of key takeaways in bullet form.
  7. Set up a Zapier automation: weekly trigger → Claude generates an update on your topic → emailed to you.
  8. Review the full output. Edit what needs editing. Save the best prompts to your library.
Research Sprint Prompt
You are a research analyst. I need a synthesis on: [YOUR TOPIC] Find and summarize 8–10 key sources or findings. For each: - Source name / author / year - One-sentence main finding - Relevance to [YOUR CONTEXT/ROLE] Then identify 3 themes that appear across multiple sources. Flag any conflicting findings or gaps in the evidence.
Executive Brief Prompt
Using these research themes: [PASTE THEMES] Write a 500-word executive brief on [TOPIC] for [AUDIENCE]. Structure: Key Finding → Why It Matters → What to Do About It Tone: Direct, evidence-based, no jargon. End with: 3 recommended actions.

Track B: The Team Leader System

🖥️HANDS-ON EXERCISE90 min

Manager AI System in 90 Minutes

  1. Identify the #1 task that eats your team's time every week (e.g., status reports, meeting prep, emails).
  2. Create a Claude Project. Write a system prompt for your role as a manager (use the template from Lesson 16).
  3. Upload your team's most-used template, style guide, or SOP document.
  4. Build Automation 1: Email summarizer — new emails from key stakeholders → Claude brief → Slack.
  5. Build Automation 2: Meeting prep — 30 min before a meeting → Claude drafts agenda + talking points → sent to you.
  6. Build Automation 3: Status report — every Friday 4pm → Claude generates weekly summary → Google Doc.
  7. Create a 1-page "AI Starter Guide" for your team: 5 prompts, how to access Claude, what it's useful for.
  8. Demo one workflow live to your team. Measure: how long does it take now vs. before?
Manager System Prompt Template
You are an executive assistant supporting a [YOUR ROLE] managing a team of [SIZE] in [INDUSTRY]. CONTEXT: [Describe your team's main work, challenges, and communication style.] ALWAYS: - Summarize key points before explaining details - Flag action items and owners clearly - Match the directness level of a senior leader NEVER: - Add unnecessary caveats or filler phrases - Make decisions on behalf of the team - Skip flagging things that need human judgment OUTPUT FORMAT: Summary | Details | Action Items | Timeline

Track C: The Specialist Build

🖥️HANDS-ON EXERCISE90 min

Your Professional AI Expert

  1. List the top 5 tasks in your profession that take the most time and follow a repeatable pattern.
  2. Create a Claude Project. Write a detailed system prompt with your professional context, constraints, and standards.
  3. Upload 2–3 reference documents: your professional guidelines, templates, or past high-quality work.
  4. Write and test 10 prompts — 2 for each of your top 5 tasks. Save them to your prompt library.
  5. Identify the one task that is most repetitive and most automatable.
  6. Build one automation in Zapier or Make that handles that task with minimal input from you.
  7. Run a full "day in the life" test: use your Project for everything you do today.
  8. Measure: how many hours did your AI system save you? Document this.
Specialist System Prompt Framework
You are a professional assistant supporting a [YOUR PROFESSION] with [X years] of experience in [SPECIALTY]. CONTEXT: [Describe your practice/organization, typical clients/patients/cases, key tools you use.] PROFESSIONAL STANDARDS: Always adhere to [your field's] best practices. Flag anything requiring [professional judgment / verification / client review]. ALWAYS: - Use correct professional terminology for [your field] - Structure output in [your field's standard format] - Include appropriate disclaimers when needed NEVER: - Make final recommendations without flagging need for professional review - Include client-specific identifying information - Deviate from documented [standards/protocols/guidelines]

Track D: The Builder

🖥️HANDS-ON EXERCISE90 min

Build a Shareable AI Workflow

  1. Identify one specific problem in your organization that AI could solve — something others face too.
  2. Design a 3–5 step prompt chain: each step takes the output of the previous one as input.
  3. Test every step with real data. Iterate until each step produces reliable output.
  4. Document it as a 1-page SOP: Problem → Steps → Prompts → Expected Output → Common Issues.
  5. Record a short demo (Loom or screen recording) showing the full workflow in action.
  6. Share it with one colleague. Have them try it without your help. Note where they get stuck.
  7. Refine based on their feedback. Fix the gaps.
  8. Submit: your SOP doc + demo link + a 2-paragraph explanation of the time saved.
Prompt Chain Documentation Template
WORKFLOW NAME: [Name] PROBLEM IT SOLVES: [Specific pain point] TIME SAVED: [Estimated hours per week] REQUIRES: Claude [free/Pro] + [any other tools] STEP 1 — [Name]: Input: [What do you paste in?] Prompt: [FULL PROMPT] Output: [What do you get?] STEP 2 — [Name]: Input: [Output from Step 1 + anything new] Prompt: [FULL PROMPT] Output: [What do you get?] [Continue for each step] COMMON ISSUES: [What goes wrong and how to fix it]

Reflection Questions

Answer these in writing before marking the capstone complete. Honest answers = better outcomes.

Q1.

What's the one workflow you built that now saves you the most time? Be specific.

Q2.

What was harder than expected? What was surprisingly easy?

Q3.

What AI workflow would you build next — if you had one more week?

Q4.

How would you explain AI's role in your work to a skeptical colleague in 2 sentences?

Q5.

What's your biggest remaining question about AI after completing Level 2?

💡
When you're done, mark this lesson complete in the sidebar. That's your signal that Level 2 is finished and the skills are yours to keep.
COMING SOON

Level 3: Advanced AI Engineering

You've mastered using AI. Level 3 is about building with it. Go from power user to AI architect.

🏗️
RAG SystemsBuild apps that search your own documents
🤖
Multi-Agent AIOrchestrate teams of AI agents
🔌
Claude APIBuild products powered by Claude
🧬
Fine-TuningTrain models on your data
🛡️
AI Safety Eng.Build responsibly at scale
🚀
Deploy & ScaleShip AI features to real users

Level 3 launches soon. Complete the capstone to get early access.

← Back to Lesson 18Review from the beginning →