Hey friend,

Don’t lie to me now…

Have you ever typed something into ChatGPT, read the response, and thought "that is the most generic, useless, slightly-too-cheerful thing I have ever read in my life?"

Yeah. Same.

And here's the thing… it wasn't the AI's fault.

It was the prompt.

I know, I know. Nobody wants to hear that. But stay with me because once you understand this, you’ll start getting responses 10x better for the same amount of effort. The difference between a terrible AI output and one that actually saves you an hour of work is almost entirely in how you asked for it.

So today I'm going to show you exactly how to ask. Step by step.

(FULL PROMPT AT THE BOTTOM)

By default, AI is trying to be helpful to literally everyone on earth simultaneously, which means it ends up being spectacularly useful to nobody in particular.

Fix this by giving it a role before you ask for anything.

Instead of: "Write me an email to a client"

Try: "You are a direct, warm account manager who has worked with this client for two years and knows they hate waffle. Write me an email to a client."

That one sentence changes everything about the output. Try it right now. You'll see.

AI cannot read your mind. I know it feels like it should be able to… it's AI, it knows everything, it once explained quantum physics to me in the style of a disappointed dad, but when it comes to YOUR business, YOUR customer, YOUR situation, it knows nothing.

You have to tell it.

Instead of: "Write me a headline for my landing page"

Try: "Write me a headline for my landing page. I sell an online course to overwhelmed freelance designers who want to raise their prices without losing clients. They're scared of seeming greedy. The main outcome is that they leave with a script they can use in their next client conversation."

See how much more specific that is? That's the version that gives you something you can actually use.

This is the one nobody does, and it’s a tragedy.

AI has a default mode. It's a bit corporate, a bit enthusiastic, very fond of the word "leverage," and absolutely obsessed with starting sentences with "Certainly!" You don't want any of that.

So tell it.

Add this to the end of almost any prompt: "Do not use jargon. Do not be overly formal. Do not start with a compliment or an affirmation. Sound like a knowledgeable human wrote it, not a committee."

You're welcome.

One output is never the best output.

Ask for ten headlines. Ask for five different email openers. Ask for three ways to structure this section. Then pick the best one, or combine two of them, or use one as a jumping off point.

Professional copywriters write thirty options and throw away twenty seven. AI lets you do the same thing in eleven seconds.

Stop settling for the first thing it gives you.

If the output isn't right, tell it why. Out loud. Like you're talking to a person.

"This is too formal, make it more casual." "This is too long, cut it in half." "The third option is closest but make it more urgent." "This sounds like it was written by someone who has never spoken to an actual human before, try again."

You're not stuck with the first version. The whole point is the back and forth. Most people quit after one attempt and conclude that AI doesn't work for them. The people getting the good stuff are on attempt four or five and it took them eight minutes total.

"You are a [ROLE]. I need you to [TASK]. Here is the context: [YOUR SPECIFIC SITUATION]. The audience is [DESCRIBE THEM]. The tone should be [DESCRIBE TONE]. Do not [LIST THINGS YOU HATE]. Give me [NUMBER] options."

That's it. That's the framework. Save it somewhere. Use it every time. Watch your outputs become actually useful.

It's what I read every morning before I open anything else because I like knowing what's going on in the world without spending forty five minutes arguing with an algorithm about it.

Free, five minutes, genuinely good. Worth your time.

Be the smartest person in the room. 1440 navigates 100+ sources to deliver a comprehensive, unbiased news roundup — politics, business, culture, and more — in a quick, 5-minute read. Completely free, completely factual.

Sign up now!