April 16, 2026
Here's a scene that plays out thousands of times a day.
Someone opens ChatGPT for the first time, types "help me write an email" — and gets back a stiff, generic draft that sounds nothing like them. They shrug, close the tab, and tell someone later: "I tried it. It's okay, I guess."
The tool didn't fail them. The prompt did.
Nobody tells you this upfront. But once you see it, you can't unsee it.
When you give AI a vague prompt, you get a vague answer. Every time.
"Help me write an email" could mean a hundred different things. So AI makes a guess — usually a safe, bland, corporate sounding one — because it has no idea who you're writing to, what you need them to do, or what tone you want.
It's like walking up to a stranger and saying "make me something to eat." They might hand you a granola bar. Technically food. Not what you wanted.
Here's the same request, done right:
"Write a short, friendly email to a client named Sandra. I'm following up on a proposal I sent two weeks ago. I want to check in without sounding pushy. Keep it under five sentences."
That gives AI something to work with. You'll get a draft you might change a word or two in and actually send.
The difference isn't the tool. It's the instruction.
You don't need all four every time, but the more you include, the better the output:
Who you're talking to. A message to your boss reads differently than one to a new client or a longtime friend. Tell AI who the audience is.
What you want them to do or feel. After reading this, should they reply? Take action? Feel reassured? Give AI the outcome, not just the task.
The tone. Friendly? Professional? Direct? Casual? AI will match whatever you name.
Any constraints. Length, format, things to include or avoid. "Under 100 words" and "don't mention price yet" are both fair game.
Before:
"Write me a social media post about my business."
After:
"Write a LinkedIn post for my small accounting firm. I want to reassure small business owners that tax season doesn't have to be stressful. Friendly and calm tone. Under 150 words."
The second one takes 20 extra seconds to write. The output is five times more useful.
If the result isn't quite what you wanted, just say so:
AI remembers the conversation. You're not starting over — you're refining.
Most people who think AI isn't useful for them have only ever given it vague prompts. Fix that one habit and the whole tool changes.
Specific in, specific out. That's really all there is to it.
If you want a simple framework for writing prompts that work across emails, documents, research, and more — that's exactly what I teach in the Clearly, AI course. Practical, no jargon, takes about 20 minutes to get through the first module.