Prompt Engineering

Prompt engineering is the craft of the instruction itself: how you phrase the task, what examples you include, what format you ask for, and how you structure the request. It’s the first lever everyone reaches for, and it still matters — a clearer prompt produces better output.

What’s changed is its weight. As models got better at understanding intent, clever wording stopped being a durable advantage; it’s now table stakes. The leverage moved outward to what you put around the prompt — the documents, history, and tools you supply (context engineering) and what the system remembers across tasks (memory engineering). Prompt engineering is still necessary, just no longer where the differentiation lives. Provider guides now put it after measurement: Anthropic’s prompt engineering guide assumes you already have success criteria and a way to test against them before you start rewording, which is evaluation work.

See AI vocabulary in plain English for this term alongside the other language of models, data, agents, and infrastructure.