Prompt engineering
The craft of writing the instructions you give an AI model to get a better result. A prompt is simply what you type in; prompt engineering is the practice of phrasing, structuring, and giving enough context that the model does what you actually want.
A prompt is the piece of text you hand to an AI model: a question, an instruction, a paragraph of context, or all three at once. Because a large language model responds to whatever it is given, the wording of that request has an outsized effect on the answer. Ask vaguely and you get a vague reply; ask precisely, with examples and a clear goal, and the same model can produce something genuinely useful.
Prompt engineering is the name the field has given to doing this well. In practice it means being specific about the task, spelling out the format you want, giving the model relevant background, and sometimes showing it a worked example or two so it can follow the pattern. None of this changes the model itself, the way fine-tuning does; it simply gets more out of the model that already exists.
It is worth knowing the term because so many AI products are, under the surface, a carefully written prompt wrapped around a general model. When a company describes a clever assistant built in weeks rather than years, skilled prompting is often a large part of how they did it.