---
title: "What Is Prompt Engineering?"
description: "Prompt engineering is the craft of writing clear instructions and context so an AI model reliably produces the output you want. Meaning and core techniques."
type: "glossary"
locale: "fr"
term: "Prompt Engineering"
canonical: "https://agenticschool.dev/fr/glossary/prompt-engineering"
dateModified: "2026-07-07"
---

# What Is Prompt Engineering?

- Definition: Prompt Engineering
- Updated: 2026-07-07
- Keywords: prompt engineering, what is prompt engineering, prompt engineering meaning, prompt engineering vs context engineering, ai prompt engineering
- Canonical URL: https://agenticschool.dev/fr/glossary/prompt-engineering
- Locale: fr

> Prompt engineering is the craft of writing clear instructions and context so an AI model reliably produces the output you want. Meaning and core techniques.

Prompt engineering is the practice of writing the instructions, context and examples you give an AI model so that it reliably produces the output you actually want. Because a language model responds to whatever you put in front of it, small changes in wording, structure and the context you supply can be the difference between a vague answer and a precise, useful one. Prompt engineering is not magic phrasing; it is the disciplined habit of being specific about the goal, the format, the constraints and the examples, and of giving the model the right information to work from. It is the entry point to everything else - agentic coding, RAG and multi-step workflows all rest on prompts that are clear enough to act on.

## What a good prompt contains

A strong prompt is rarely a single clever sentence. It states the role and goal, gives the relevant context, spells out the output format and constraints, and often shows one or two examples of what "good" looks like. The more the task matters, the more it pays to be explicit rather than hoping the model guesses your intent.

- Goal: say exactly what you want and why, not just the topic.
- Context: paste the facts, data or files the model needs to answer well.
- Format and constraints: define the shape of the output and what to avoid.
- Examples: show one or two samples of the result you expect (few-shot).

## Prompt engineering vs context engineering

Prompt engineering is about how you phrase the request; context engineering is about what information the model has in its window when it answers. As tasks grow into agents and RAG, the leverage shifts from clever wording to supplying the right context - retrieved documents, tool results, prior steps. The two work together: a clear prompt tells the model what to do, good context gives it what it needs to do it.

## Why it still matters with agents

Even the most capable coding agent follows the instructions it is given. Clear goals, explicit acceptance criteria and the right context are what turn an agent from unpredictable into dependable. Prompt engineering is therefore not a beginner phase you outgrow; it is the interface through which you direct every AI system, and it is the first skill that separates people who get consistent results from those who do not.

## FAQ

### What is prompt engineering in simple terms?

Prompt engineering is writing your instructions to an AI model clearly enough that it reliably gives you the result you want. It means being specific about the goal, the format and the context, instead of hoping a vague request produces a good answer.

### Is prompt engineering still relevant?

Yes. Models get better, but they still do exactly what the prompt and context tell them. Clear goals, constraints and examples remain the difference between consistent, useful output and hit-or-miss results, especially once prompts drive agents and automated workflows.

### What is the difference between prompt engineering and context engineering?

Prompt engineering is how you word the request; context engineering is what information sits in the model window when it answers. Simple tasks lean on phrasing; agents and RAG lean on supplying the right retrieved context. In practice you use both together.

### What are common prompt engineering techniques?

The reliable staples are being specific about the goal and format, giving relevant context, showing a few examples (few-shot), asking the model to reason step by step for hard problems, and iterating on the wording based on the outputs you get back.