Lesson 2: First project — say what you want

A beginner's most common mistake isn't technical. It's a brief like this:

"make me a calculator"

The AI will build something — but probably not what you had in mind. Not because it's stupid, but because you kept half of the brief in your head. A good brief has four parts:

WHAT should exist — the goal in one sentence

WHY it exists — who will use it and for what

HOW it should look — form, device, language

WHEN it's done — behaviour you can verify

Exercise: write a brief

Imagine you want a margin calculator for a small business owner: they enter the purchase and selling price and want to see how much they earn. Write the brief for Claude Code the way you'd really write it — then click Evaluate.

The checker is a simple aid — it checks the structure of your brief, not the quality of your idea. But if it flags something, Claude would almost certainly miss it too.

How the app grew out of it

This is a real session — the model brief above, Claude Code below. Including the moment when the first version wasn't perfect and feedback was needed.

Try the result

The app from the replay is real — this is exactly what one short session produces.

Summary

A bad brief isn't short — it's incomplete. Half of it stayed in your head.
Four parts: WHAT · WHY · HOW · WHEN it's done.
Give specific, numbered feedback — one problem, one item.

That was the foundation. In the next lesson you'll learn to iterate so the AI gets it right on the first try. And if you'd rather have AI deployed in your company without the learning curve — let's start with an initial consultation; the fee is credited toward the project.

First project:say what you want

Exercise: write a brief

How the app grew out of it

Try the result

Summary

First project:
say what you want