AI Glossary
Prompt
The instruction you give an AI tool describing what you want it to do.
In plain English
A prompt is simply what you type or say to an AI tool, and a clearer prompt gets a better result. A dependable way to structure one is Role, Task, Context, Constraints, and Output: who the AI should act as, what to do, the background it needs, the limits to respect, and the form the answer should take.
Example: 'You are my scheduling assistant. Draft a polite reply proposing three times next week, under 80 words, in a warm but professional tone.' That's a prompt.
What a prompt is not
A prompt is not a search query. You are not typing keywords to match; you are giving instructions to a tool that will try to do what you describe. That means full sentences and useful detail help, where in a search box they might get in the way.
It is also not a magic spell. There is no secret phrase that unlocks better answers. The quality comes from being clear about what you want. And a prompt is not a one-shot command you have to get perfect on the first try. Refining it, or editing the directly, is part of the normal process.
A simple formula for writing one
A dependable way to structure a prompt (the formula we use and teach) has five parts: Role, Task, , , and Output.
- Role: who the AI should act as: 'you are my scheduling assistant.'
- Task: what you want done: 'draft a reply proposing three times.'
- Context: the background it needs: who it is for, your business, the tone.
- Constraints: the limits: 'under 80 words, no jargon, do not invent prices.'
- Output: the form of the answer: 'give me three options as a bulleted list.'
Why it matters
The prompt is the single biggest lever you control. The same tool produces a throwaway result from a vague request and a genuinely useful one from a specific request: no change to the technology, only to the instruction. Learning to write clearer prompts, a skill called , transfers between every AI tool you will ever use.
One caution: because the tool follows your framing, a confident wrong answer to a confidently worded prompt is still wrong. Clear prompts reduce mistakes; they do not remove the need to check facts, numbers, and anything customer-facing.
Related terms
Keep exploring
How to Write AI Prompts
The Prompt Formula (Role, Task, Context, Constraints, Output) with a live Prompt Builder.
Read the guide →AI Prompt Library
Reusable prompts by task, each with required inputs and review-before-use notes.
Read the guide →AI assistant demonstration
Try an AI assistant built around an approved knowledge base.
Try the demonstration →Prompt: Quick Questions
What makes a good prompt?
Specificity. A good prompt says who the AI should act as, what you want done, the background it needs, any limits to respect, and the form the answer should take. Vague requests get generic results; specific requests get usable ones.
- Prompts
- Do I need special words or secret phrases?
- No. There are no magic words. Clear, plain instructions with enough context beat clever phrasing. If a result is not right, the fix is almost always to add detail or a constraint and try again, not to find a special keyword.
- Why did I get a bad answer from a simple prompt?
- Usually because the prompt left too much to assumption. The tool filled the gaps with generic choices. Add who it is for, the tone you want, and the facts it should not guess, and the same request produces a much better result.
- Is the first answer the final one?
- Treat it as a draft. Refining through a short back-and-forth ('shorter,' 'more formal,' 'keep only the second option') is normal and is where most of the value comes from. A person still reviews anything important before it is used.