Where the term comes from In this channel's material, prompt context appears in Steph Smith's prompting lesson: the way to change what an AI produces is to give it context, a vision, and examples — and the output changes in response. What each part does - Context tells the model who or what it is working with — e.g., 'people who love the internet'.- Vision describes the shape you want — e.g., 'split into smaller groups'.- Examples show the model the kind of result you mean — e.g., 'like their favorite social media app'.Each piece narrows the space of possible answers so the model aims where you want it. How the practitioners use it In Steph Smith's workflow, a single meetup icebreaker prompt improves when she layers in the audience ('people who love the internet'), a structure ('split into smaller groups'), and a reference point ('like their favorite social media app'). The broader habit in the channel's material is the same: shape the answer you want before asking again. Use the redo button before adding more, tell it explicitly 'this is the kind of answer I want', and treat wrong-but-close answers as valuable — they help you refine what you actually want. Prompt context is one part of that refinement loop, alongside knowing the mistakes the model is likely to make and how to check and edit. - Prompt context = context + vision + examples given to the AI before or inside a prompt.- Supplying even one of these ('like their favorite social media app') measurably changes the output.- Pair it with the redo button and explicit feedback ('this is the kind of answer I want') to refine results fast.- Wrong-but-close answers still help: they clarify what you actually want.