- A prompt is whatever you give an AI to respond to.
- It can be a question, an instruction, or a document with “summarize this”.
- A short prompt only works when the AI already knows the background.
Same four words, two answers
Prompt alone
“can i text back ?”
Nothing else to go on.
It answered: relationship advice. It asked who had texted me.
Prompt + background
“can i text back ?”
Plus: what it last sent me, and how the texting works.
It answered: yes, reply here and I’ll answer.
I’d set my assistant up to text me, and I replied with those four words.
The first answer was a fair reading. It had no idea I was replying to its own message. Nobody had told it.
I didn’t write a better prompt. I gave it what it was missing.
A good prompt has three parts
Weak prompt, better prompt
| Instead of | Try |
|---|---|
| “Write a bio.” | “Write a 60-word bio for my coaching site. Warm, first person, no buzzwords.” |
| “Summarize this.” | “Summarize this for a client with two minutes. Three bullets, then the decision I need.” |
| “Give me ideas.” | “Give me five post ideas for busy salon owners. One line each.” |
Three mistakes I see people make
- Writing a longer prompt when the AI is missing background.Give it the background instead.
- Asking for everything at once.One job per prompt.
- Not saying what done looks like.Then being annoyed by the length.
Talk it through with me. Tell me what your work looks like and we’ll find where AI helps first.
Quick answers
What is a prompt in AI?
The input you give an AI model to respond to. It can be a question, an instruction, a document, an image, or a mix.
What makes a good prompt?
Say what you want, who it’s for, and what done looks like.
Then give it something real to work from: the proposal, their last message, an example you like.
Is a prompt the same as a question?
A question is one kind of prompt. So is an instruction, or a photo with “what’s wrong here?”
Apps also add their own instructions before yours. That’s called a system prompt.
Do I need to learn prompt engineering?
No. Most of the old tricks don’t matter now, because the models got better at plain requests.
What still matters: being clear about the job, and giving the model what it needs to see.
What is a system prompt?
Instructions the app or its builder gives the model before your message: its role, its rules, its tone. You usually can’t see it.
How long should a prompt be?
As long as the job needs. One line is fine when the AI already knows the background. A new task with no background needs a few sentences.
Next: prompt vs context, and choosing what the model sees, called context engineering.
AI Foundations starts here: give AI a clear picture of your business in a form you can reuse, so you stop explaining from zero.
Background: Wikipedia: Prompt engineering