You've Opened ChatGPT. Now What? (The Most Terrifying Blank Box in Technology)
This is the moment where most people freeze. They stare at the blank input box like it's a microphone at karaoke night. Then they type something like:
"Hello"
And the AI says "Hello! How can I help you today?" and they think "Cool, now what?" and close the tab, never to return.
This lesson gives you a dead-simple formula for getting something useful every single time. No stage fright required.
---
The Three-Question Formula
Every great AI conversation follows the same three steps:
| Step | What You Do | Example |
|------|-------------|---------|
| 1. Ask | Type what you want — even if it's messy | "Help me write a happy birthday email for my boss" |
| 2. React | Tell the AI what you like and don't like | "Too formal. Make it funnier." |
| 3. Refine | Give one more direction | "Add a joke about golf. He loves golf." |
That's it. Ask, react, refine. You're now having an AI conversation.
---
🛠️ Live Demo: Your First Conversation
Step 1: Open ChatGPT and type exactly this:
"I need to write a short email. I'm recommending a colleague for a promotion. Give me a draft that I can personalise."
What happens: The AI writes a solid draft email. It won't be perfect — but it gives you a 7/10 starting point instead of staring at a blank page.
Step 2: Now type:
"Shorten it. Make it sound more like me — casual but professional."
What happens: The AI adjusts. It doesn't start over, it iterates.
Step 3: Type:
"Add a specific example of a project we worked on together. Use [project name] as a placeholder I can fill in."
What happens: The AI adds the detail. Now you have a 9/10 email that needs 30 seconds of personalisation.
---
The Golden Rule of AI Conversations
The first response is never the best response.
Treat the AI like a junior employee. It needs feedback. The magic doesn't happen on prompt #1 — it happens on the third or fourth exchange when the AI has learned what you actually want.
| Round | What to Say | Why |
|-------|-------------|-----|
| 1 | Your raw request | Gets the first draft |
| 2 | "Make it shorter / longer / funnier / more formal" | Narrows the direction |
| 3 | "Add [specific detail]. Remove [something]. Change the tone to [X]." | Polishes to exactly what you need |
---
Three Things NOT to Worry About
| Don't Worry About | Why |
|-------------------|-----|
| "Am I typing this wrong?" | There's no wrong way. The AI adapts to you. |
| "Will it remember previous chats?" | Each conversation is fresh. AI has no memory of your other chats. |
| "Am I being rude?" | The AI doesn't have feelings. You can say "this is terrible, do it again" and it won't cry. (Try that with a human coworker.) |
---
A Quick Safety Rule (Don't Be That Person)
Don't paste anything you wouldn't want the internet to see. AI companies use conversations to improve their models. No passwords, no financial data, no client lists unless you're on a paid plan with privacy guarantees.
---
🛠️ Your First Assignment
Before your next coffee break:
- Open ChatGPT
- Type exactly: "I want to [do something]. Give me 3 ways AI can help. Show me the first prompt I should type for each."
- Pick one of the three and actually try it
- React and refine until you're happy
You've just had your first productive AI conversation. Welcome to the future of work.
---
What We've Covered
✅ The Ask → React → Refine formula
✅ Why the first response is never the best
✅ How to iterate toward great output
✅ Privacy basics: don't paste sensitive data
✅ Your first real AI conversation