# Prompt Engineering

> The practice of writing and testing instructions so a model produces useful, consistent outputs.

- Category: Context & Memory
- Stage: Established
- Also called: Prompt design, Prompting
- Canonical: https://www.joelcsaji.com/ai-dictionary/prompt-engineering

## Plain English

It is learning how to ask the AI in a way that reliably gets the kind of answer you need.

## Example

A support bot prompt includes role guidance, examples, output format rules, and escalation instructions.

## Why it matters

Prompting remains the first layer of control, even as serious systems add retrieval, tools, and evaluations.

## Related

- [Prompt](https://www.joelcsaji.com/ai-dictionary/prompt)
- [Context Engineering](https://www.joelcsaji.com/ai-dictionary/context-engineering)
- [Evals](https://www.joelcsaji.com/ai-dictionary/evals)

## Mentioned in

- [Fable 5, Export Controls, and GLM 5.2](https://www.joelcsaji.com/blog/fable-5-export-controls-and-glm-5-2)

## Sources

- [OpenAI: Text Generation](https://platform.openai.com/docs/guides/text-generation)
- [OpenAI: Prompting](https://platform.openai.com/docs/guides/prompting)