AI prompts: how to ask so the answer is useful
How to write prompts that work, with ready examples for everyday tasks.

A prompt is what you type to an AI. The better you state the task, the better the answer. No magic words are needed, only clear instructions.
Four parts of a good prompt
- Task. Say what you want done, in one sentence.
- Context. Give the facts the AI cannot know: who the reader is, what you already have, what must be avoided.
- Format. Name the shape: a list, a table, three sentences, a draft email.
- Example. If you can, show one sample of the result you like.
Habits that help
- Ask for one thing at a time. Long mixed requests give muddy answers.
- Say what to do, not only what to avoid.
- Check the facts in the answer. A model can state wrong things with confidence. This is called a hallucination.
- If the answer misses, say what was wrong and ask again. A second round is normal.
System prompts and settings
A system prompt sets standing rules for a whole chat, such as tone or role. The setting temperature controls how varied the answers are in tools that expose it.
Ready prompts
See everyday prompts for tasks you can copy and adapt.
A risk to know
Text you paste can contain hidden instructions. This is prompt injection. Do not let an AI act on content you do not trust.
All pages in this section
- 12 everyday prompts you can copy
Twelve copy-and-paste prompts for daily tasks, with the reason each one works, and a short method to write your own: role, context, goal, format, examples.
Read next
- Prompt
A prompt is the input you give an AI model: your question or task, plus any instructions, examples and text it should use.
- System prompt
A system prompt is a block of standing instructions, set by an app or developer, that shapes how a model behaves across a whole conversation.
- 12 everyday prompts you can copy
Twelve copy-and-paste prompts for daily tasks, with the reason each one works, and a short method to write your own: role, context, goal, format, examples.
- Prompt injection
Prompt injection is an attack where text a model reads, such as a web page or file, contains instructions that make it act against its user.