Lesson 1 of 8
Say what you actually want
6 min read
You ask an AI to "write an email" and get back something bland and generic, full of blanks. It's tempting to blame the AI. But the problem is usually the ask: you knew what you wanted — it didn't.
The AI can't read your mind
When you leave something out of a prompt, the AI doesn't leave it blank. It fills the gap with its best guess — the most average, most likely version. Ask for "an email" and you get a generic email. Every detail you don't say is a detail it decides for you, and it decides for the average case, not yours.
An unstated detail isn't left empty — the AI fills it with its most likely guess. The more you specify, the less it has to guess, and the closer the answer lands to what you meant.
Say the parts that matter
You don't need to specify everything — just the parts you'd care about if a person did the task. Usually that's a few things: who it's for, how long it should be, the tone, and any must-include detail. Naming even one or two of these sharpens the answer dramatically. It's the difference between "write an email" and "write a short, friendly email to my landlord about the broken heating."
The details worth stating are the ones a human would ask you about: audience, length, tone, and anything that must be included. Name those, and the AI stops guessing about them.
Asking with just a description and no example — the way most people prompt — has a name: zero-shot. It's perfectly fine for simple things; the next lesson shows when adding an example is worth it.
The gist
- —Whatever you leave out, the AI fills in with its most likely guess — not yours
- —Name the parts a person would ask about: audience, length, tone, must-includes
- —One or two specifics turn a generic answer into a usable one
You ask, "Give me some ideas for my sister's birthday." The list comes back generic and useless. Which single addition would sharpen it most?
Continue in the app
Take the whole Talking to AI course — tracked
Get your personalized path, progress and streaks in the app — this lesson and every next one, in order.