# Meta-Prompting

> Asking a model to write or improve a prompt for you.

**Field:** Artificial Intelligence  
**Category:** Working with AI  
**Also known as:** Meta Prompt, Prompt Generation, AI-Written Prompts

## Think of it like

It is like asking a cook to write the recipe. Then a new cook makes it.

## Definition

Meta-prompting means you ask a model to write or fix a prompt for you. You describe the job, the reader and what a good answer looks like. The model writes a full prompt with steps, rules and an example. You then use that prompt with the same model or another one. You can also paste in a weak prompt and ask what is wrong with it.

## Example

Jack needs a system prompt for a help desk bot. He describes the shop and its rules in five lines. The model writes a full prompt, and Jack trims it.

## Example code

```text
Write a system prompt for a help desk bot for a bike shop.
It should answer in under 80 words, never promise refunds,
and pass hard questions to a person.
Then list three ways the prompt could go wrong.
```

A prompt whose answer is another prompt, plus the weak spots to test.

## Why it matters

Models are good at writing clear, full prompts. It saves time and catches gaps you would miss.

## Common mistake

Using the written prompt without testing it. A prompt that reads well can still fail on real inputs.

## Three principles

- **Describe the job.** The task, the reader, the goal.
- **Let it draft.** The model writes the prompt.
- **Still test it.** Run it on real inputs.

## Related terms

- [Prompt Engineering](https://learn.benibix.com/term/ai/prompt-engineering)
- [System Prompt](https://learn.benibix.com/term/ai/system-prompt)
- [Iterative Prompting](https://learn.benibix.com/term/ai/iterative-prompting)
- [Prompt Template](https://learn.benibix.com/term/ai/prompt-template)
- [Context Engineering](https://learn.benibix.com/term/ai/context-engineering)

## Taught in

- [Iterative and Meta-Prompting](https://learn.benibix.com/learn/ai-working-with-ai/iterative-and-meta-prompting): Prompt Engineering, 6/16

![Two ways to get a better prompt. Left, iterative prompting: a loop of ask, read the draft, ask for one change. Right, meta-prompting: you describe the job to a model, it writes a full prompt, and that prompt is run on a model to get the answer.](https://learn.benibix.com/diagrams/ai3-prompt-help.svg)

---

Source: https://learn.benibix.com/term/ai/meta-prompting
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-18
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
