Lab: Build Your AI-Powered Week

This is the capstone. Everything from the last 13 days comes together right here.

You're going to design, build, and test your AI-powered work week — a system where every recurring task has an AI workflow, at least one runs automatically, and you walk away with something you'll use starting Monday.

Set aside 25-30 minutes. This one's worth the time.

Exercise 1: Map your recurring tasks (5 minutes)

Open a blank doc or spreadsheet. Think through a typical work week, Monday through Friday, and list every recurring task. Not one-offs. The stuff you do every week (or every day) that follows a pattern.

Don't filter. List everything. Here's a starter to jog your memory:

• Email triage and responses
• Meeting prep and follow-ups
• Status reports or updates
• Data entry or processing
• Research and summarization
• Presentations or decks
• Client communications
• Scheduling and calendar management
• Expense tracking
• File organization
• Content creation (social posts, newsletters, blog drafts)
• Review and editing

Now, for each task, write three things:

1. How long it takes you now (rough estimate)
2. AI mode: Delegate / Collaborate / Do myself (from yesterday's lesson)
3. Best tool for this: Cowork, ChatGPT, Claude, scheduled automation, Zapier, built-in AI feature

You should end up with a table that maps your entire work week to AI workflows. This is your operating manual.

Exercise 2: Set up one scheduled automation (10 minutes)

Look at your table. Find one task that meets all three criteria:

• You do it every week (or every day)
• It's in the "Delegate" column
• It has predictable inputs and outputs

This is your automation candidate.

Open Claude Desktop and Cowork. Set up a /schedule command for this task. Be specific:

What it should do (step by step):

When it should run (daily at 7am? Every Friday at 3pm?):

Where inputs come from (which folder, which app):

Where outputs go (which folder, what filename format):

What format the output should be (spreadsheet, report, presentation, summary):

Example:

/schedule every weekday at 7:00am — "Check my Gmail for any emails received since yesterday evening. Summarize each email in 1-2 sentences, flag any that need a response today, and save the summary as morning-brief-[today's date].md on my Desktop."

Set it up. Then trigger it manually once to test. Review the output. Adjust the prompt if needed. Run it again.

When you're satisfied, leave it scheduled. You'll have your first fully automated AI task running by tomorrow morning.

Exercise 3: Build one multi-step workflow (10 minutes)

Now find a task from your table that's in the "Collaborate" column — something where AI does the heavy lifting but you add judgment and refinement.

Build a complete multi-step workflow for it in Cowork. Remember the chaining principles from Day 9: describe the end state, let the steps flow into each other, be specific about outputs.

Here are some templates to adapt:

Meeting prep workflow:

"I have a meeting with [Name] from [Company] tomorrow. Research their company's recent news and quarterly results. Check my email history in Gmail for any past conversations with them. Create a one-page briefing doc with: background on the person and company, summary of our relationship history, 5 recommended talking points, and 3 questions to ask. Save it to my Meeting Prep folder."

Weekly report workflow:

"Read all files added to my Project Updates folder this week. Synthesize them into a weekly status report with: key accomplishments, blockers, next week's priorities, and any decisions needed from leadership. Format it as both a one-page summary and a 5-slide PowerPoint deck. Save both to my Weekly Reports folder."

Content creation workflow:

"Research the topic [X]. Find 5 recent articles or reports with interesting data points. Summarize the key insights. Draft a 500-word article in my voice (reference my writing style from the samples in my Writing Samples folder). Create 3 social media post variations — one for LinkedIn, one for Twitter, one for an email newsletter. Save everything to my Content Drafts folder."

Pick one. Adapt it to your actual work. Run it. Review the output.

Then ask yourself: how long would this have taken me manually? How long did it take with AI? What's the quality difference?

Exercise 4: Design your AI Monday (5 minutes)

Now put it all together. Write out what your ideal Monday morning looks like with your AI system running.

Here's an example:

6:30am — Cowork's scheduled task runs. By the time I open my laptop, my morning brief is sitting on my Desktop with overnight email summaries and today's calendar.

7:00am — I review the brief over coffee. Flag two emails that need responses. Ask Cowork to draft replies based on my context.

7:15am — Review and send the draft responses. Check the auto-generated meeting prep doc for my 9am meeting.

7:30am — Start actual deep work, already caught up on everything without opening my inbox once.

Write your version. Be specific about which automated tasks fire, which workflows you trigger manually, and where your human judgment gets applied.

This isn't a fantasy exercise. This is your playbook for next week. Make it real.

Reflection questions

Before you close this out, answer these honestly:

1. How many hours per week do you estimate your AI system will save you? Be conservative. Even 3-5 hours is transformational.

1. What's the one workflow you're most excited about? The one that made you think "why wasn't I doing this already?"

1. What's your biggest remaining uncertainty? Where are you not sure AI can handle it well enough? (That's fine. Those are your "collaborate" tasks where you stay closely involved.)

1. What would you add to your system in 30 days? Not everything has to be built today. What's the natural next workflow to add once you've proven the first few work?

1. Looking back at your Day 1 audit — how has your "AI opportunity surface" changed? You saw possibilities two weeks ago. Now you have systems. What feels different?

Write your answers down. Come back to them in a month and see how reality compared to the plan.

You just built something real

This isn't a hypothetical. You have a scheduled automation that will run without you. You have a multi-step workflow you can trigger whenever you need it. You have a map of your entire work week with AI integrated into every recurring task.

That's a system. That's what we set out to build on Day 1.

Tomorrow is the final day. We'll talk about where all of this is headed, how to keep getting better, and what your continued learning path looks like. It's a short one, but it matters.

You didn't just learn about AI. You built an AI system. That's the difference between knowing and doing.