Claude Code for Non-Engineers
70 minutes
Module 11 · Tier III
Claude Code
Non-engineers build tools.
Charter Oak Strategic Partners
The tool gap.
You
“I need a tool that cleans this CSV and produces a report.”
vs.
Engineering
“We will get to it in Q3.”
Claude Code fills that gap.
A terminal that
writes its own code.
You describe what you want in plain English.
Claude Code writes the script, runs it, checks the output.
If something breaks, it fixes it and tries again.
Five tools under the hood.
Read
Reads files and directories
Write
Creates new files from scratch
Edit
Modifies specific lines in a file
Bash
Runs terminal commands
Search
Finds patterns across files
Data Cleaner
Messy CSV in. Clean, validated dataset out.
> Clean this CSV: standardize dates to ISO format, remove duplicates, flag rows missing email addresses, export as cleaned-data.csv with a summary.
Reading raw-contacts.csv — 2,847 rows
Found 4 date formats: MM/DD/YY, YYYY-MM-DD, DD-Mon-YYYY, Unix timestamp
Standardized all dates to ISO 8601
Removed 143 duplicate rows
Flagged 67 rows missing email
Exported cleaned-data.csv — 2,704 rows
Internal Tool
> Build an HTML page with a form where a sales rep enters: customer name, visit date, visit objectives, and key contacts. When they click Generate, create three things: a formatted visit brief, a draft calendar invite text, and a pre-visit checklist. Clean professional styling. No external dependencies. Everything in one HTML file.
Writing visit-scheduler.html
Added form with 4 fields + Generate button
Added brief, calendar invite, and checklist generators
Testing layout in browser — looks clean
This sat on an IT backlog for six months. Built in fifteen minutes.
Three CSVs. One dashboard.
> Build a dashboard that reads sales.csv, expenses.csv, and headcount.csv. Show: total revenue, total expenses, revenue per employee, expense ratio, month-over-month changes. HTML with Chart.js. Save as dashboard.html.
Reading 3 data sources
Writing dashboard.html — 6 charts, responsive layout
Testing — all calculations verified
One command. Updated dashboard. Replaces the Friday afternoon ritual of copying numbers between spreadsheets.
Five interfaces. One decision.
Quick questions, brainstorming, one-off text
Ongoing work with persistent context and documents
Multi-step file work, scheduled documents, plugins
Building tools, scripts, automations, dashboards
High-volume, system-integrated, real-time processing
Cowork vs. Claude Code
Output gets read
Documents, reports, presentations
Scheduled recurring work
Visual interface, folder-based
Output gets run
Scripts, tools, automations
Built once, executed repeatedly
Terminal interface, command-based
Does the output get read or run? That is the decision.
Your first build.
Choose a task from your work. Describe it to Claude Code in plain English. Watch it build the solution.
Fifteen minutes. Present what you built.
Eyes-open adoption.
Data Security
Code runs locally. Understand what data leaves your machine via API calls.
Code Quality
Claude Code output works. It may not be production-grade. Review before deploying.
Maintenance
Scripts need updating when data formats change. Document what you build.
Build it. Or engineer it.
Internal tools your team uses
Personal automations and scripts
Prototypes and proofs of concept
Data pipelines for internal reports
One-off scripts and file processors
Customer-facing features and products
Production systems with real users
Code touching sensitive data at scale
Security-critical authentication flows
Anything that breaks a business if it fails
The highest-value move: build the prototype. Hand it to engineering and say "make this production-ready."
You have built a tool.
Now encode a procedure.
Module 12: Skills turn your best practices into reusable instructions that Claude follows automatically.