Promptframe

A promptframe is a wireframe annotated for generative AI. For each region of the screen whose content a model will produce, it records what that content is supposed to achieve and the requirements the prompt must meet. It is a fairly new term, and practitioners do not yet use it in exactly one way, so treat this as one reasonable reading rather than a settled standard.

What it is for

In a conventional wireframe, a text block is filled with placeholder copy and the designer’s intent is clear from the layout. When a model will write that text, the layout alone says almost nothing about what should appear. A promptframe fills the gap. Beside a “product summary” box it might state the purpose (help someone decide), the constraints (under 40 words, no claims the data does not support), the tone, and what to do when the input is missing.

Why bother

Without it, prompts get written by whoever builds the feature, and the content decisions a designer would normally make are made by default. The promptframe puts those decisions back in the design process, where they can be reviewed before anything is built. It also gives you something to test the output against.

In practice

Keep it close to the layout, one note per generated region, and write requirements that can be checked. “Friendly” is a mood; “never longer than two sentences” is a requirement. Pair it with real sample outputs, including bad ones, because the failures are where most of the design work lives. Where wording is fixed, ordinary microcopy is still the better tool.