Glossary

Prompt engineering

Designing and refining the instructions given to a language model to reliably get the output you want.

Also called: prompt design

Prompt engineering is the practice of crafting the instructions, examples, and context given to a large language model so that it reliably produces the intended output, without changing the model's underlying weights. It covers wording, structure, such as asking for output in a specific format, and technique, including providing a few worked examples, "few-shot" prompting, or asking the model to reason step by step before answering.

It differs from fine-tuning, which changes model behavior by retraining on new examples: prompt engineering works entirely at inference time and is faster and cheaper to iterate on, but is also more fragile, since a small wording change or a different model version can shift results. Effective prompts also manage the context window carefully, since irrelevant or excessive context can dilute a model's attention on what matters.

Prompt engineering matters because most production LLM use, including AI agent systems, is built on prompts rather than custom-trained models, and prompt quality is often the biggest lever available for output quality. Common pitfalls include writing prompts that work on a handful of manual tests but fail on realistic input variety, and skipping systematic LLM evaluation of prompt changes, which makes regressions easy to miss.

Last reviewed September 22, 2026

In the index now

Related terms

Large language model (LLM) A neural network trained on vast amounts of text to predict and generate language, the technology behind most modern AI assistants. Context window The maximum amount of text, measured in tokens, a language model can consider at once when generating a response. LLM evaluation Systematically testing a language model's outputs for quality, accuracy, safety, and consistency before and after deployment. Hallucination When a language model generates confident, fluent output that is factually wrong or unsupported by its source material. Fine-tuning Further training a pretrained model on a smaller, targeted dataset to adapt its behavior to a specific task or domain. AI agent A system that uses a language model to plan and take multi-step actions, such as calling tools or APIs, toward a goal.

Related guides

How to choose an AI search visibility tool AI answer engines publish no rankings, so these tools sample prompts and report citations — know what "visibility" means before you buy one. How to evaluate LLM applications Offline evals, LLM-as-judge, human review and production monitoring, aimed at the whole pipeline rather than the model alone. How to use AI for data analysis safely Text-to-SQL and AI data assistants can speed up analysis or produce fluent wrong answers — the difference is what sits between the model and your data.