# Module 11: Claude Code for Non-Engineers
## Demo Scenarios

These scenarios demonstrate what Claude Code can build. The audience is non-technical. The goal is comprehension, not coding ability. Show each demo, explain what happened, and discuss when this capability matters.

---

## Demo 1: The Data Pipeline (5-minute build)

**Scenario:** Your finance team exports a CSV from their accounting system every Monday. They manually reformat it, calculate three derived columns, and paste it into a Google Sheet. Takes 45 minutes.

**What to show:**
Open a terminal. Show Claude Code. Say:

```
Write a Python script that:
1. Reads a CSV file called 'weekly-export.csv'
2. Adds a column called 'margin_pct' calculated as (revenue - cost) / revenue * 100
3. Adds a column called 'yoy_change' that compares each row's revenue to the same customer's revenue from the prior year file called 'prior-year.csv'
4. Adds a column called 'flag' that marks any row where margin is below 20% or yoy_change is negative
5. Sorts by margin_pct ascending (worst first)
6. Saves the output as 'weekly-report.csv'
```

**Teaching point:** A non-engineer described the logic in plain English. Claude Code wrote, tested, and delivered a working script. The finance team runs it every Monday by double-clicking a file. What took 45 minutes now takes 10 seconds.

---

## Demo 2: The Internal Tool (15-minute build)

**Scenario:** Your sales team tracks customer visit schedules on a shared whiteboard. It's always out of date. Nobody wants to build a tool because IT has a six-month backlog.

**What to show:**

```
Build a simple web app for tracking customer visit schedules. It should have:
- A table showing: Customer Name, Last Visit Date, Next Visit Date, Assigned Rep, Status (Scheduled / Overdue / Completed)
- The ability to add new entries
- Color coding: green for scheduled, red for overdue (past the next visit date), gray for completed
- A filter dropdown to view by sales rep
- Store data in localStorage so it persists between sessions
- Make it look clean and professional

Save it as a single HTML file I can share with the team.
```

**Teaching point:** This is a functional internal tool built in 15 minutes. No IT ticket. No six-month wait. A non-engineer defined the requirements; Claude Code built it. Show the running app in a browser.

---

## Demo 3: The Automation Script (10-minute build)

**Scenario:** Your marketing team downloads five analytics reports from different platforms every morning. They open each one, pull two numbers from each, and paste them into a daily dashboard. Takes 30 minutes.

**What to show:**

```
Write a Python script that:
1. Reads five CSV files from a folder called 'daily-reports/':
   - google-analytics.csv (pull: sessions, bounce_rate)
   - hubspot.csv (pull: new_leads, mql_count)
   - mailchimp.csv (pull: open_rate, click_rate)
   - linkedin.csv (pull: impressions, engagement_rate)
   - adwords.csv (pull: spend, conversions)
2. Extracts the most recent day's data from each
3. Compiles them into a single summary table
4. Calculates day-over-day change for each metric
5. Formats the output as a clean HTML email body
6. Saves it as 'daily-dashboard.html'

Include error handling for missing files (some days a report might not be available yet).
```

**Teaching point:** The marketing team still pulls the reports (that requires login credentials). But the 20 minutes of manual data assembly is now automated.

---

## Decision Matrix: When to Use What

Show this on screen during the module.

| Situation | Use This |
|---|---|
| I need to write an email, analyze a document, or think through a problem | **Claude (chat)** |
| I need to work with files on my computer, create presentations, or browse the web | **Claude Cowork** |
| I need to build a script, tool, or automation that runs repeatedly | **Claude Code** |
| I need to embed Claude into our company's product or systems | **Claude API** |
| I need a human engineer | **When the stakes are high, the system is complex, or the code will run in production without human review** |

---

## Discussion Questions

1. Which of these three demos would save the most time in your organization?
2. Who on your team would be the right person to learn Claude Code? (Hint: it doesn't have to be an engineer.)
3. What's on your IT backlog that could be built in an afternoon with Claude Code?
4. What are the risks of building tools this way? (Data security, maintenance, quality)