What is "prompt engineering"?

Prompt engineering is the practice of carefully wording instructions, context, and examples so an AI model produces the most useful, accurate response. Because large language models generate answers based on patterns in the text they've seen, small changes in phrasing, structure, or the order of information can steer the model toward better reasoning, more relevant details, or a clearer format, even though the underlying model itself hasn't changed at all.

It has nothing to do with manufacturing computer chips, which is a physical hardware process handled by semiconductor engineers, an entirely different discipline. It's also unrelated to repairing electrical wiring, a hands-on trade skill with no connection to software instructions.

Techniques like asking a model to "think step by step" or providing a couple of worked examples before the real question can noticeably improve answer quality, showing that how you ask can matter almost as much as what you ask.