A hybrid design artifact that combines wireframe-level structure with natural language annotations describing the intended behavior, content, and interaction of each element — particularly useful when working with generative AI tools that require structured, intent-rich prompts to produce accurate outputs. Promptframes bridge the gap between a designer's intent and what an AI generation tool needs to understand that intent.
Common contexts
- Annotating a wireframe with intent descriptions before feeding it to an AI code generation tool to produce a working component
- Creating a structured layout brief for a content team working with a generative image tool to produce on-brand illustrations
- Building a promptframe for a landing page so a non-designer stakeholder can iterate on AI-generated versions without losing structural intent
Use when
Use a promptframe when handing off design intent to a generative AI tool, a developer unfamiliar with the domain, or a non-designer who will iterate on a layout — any context where visual structure alone is insufficient to convey behaviour.
Avoid when
Don't invest in creating promptframes for static, non-interactive content or for experienced teams with established shared vocabulary — the annotation overhead adds no value when the recipient already understands the intent from the wireframe alone.
A promptframe forces you to articulate design decisions you usually leave implicit — the process of writing the annotations often surfaces more ambiguity than the AI output itself ever would.
Real-world examples
- Figma's AI layout feature accepts promptframes — rough sketches with annotations describing intent — translating hand-drawn layout intent into production-ready component arrangements.
- Design teams at IBM use promptframes during early concept exploration to rapidly generate 10–15 layout variations from a single annotated sketch, compressing what previously took a week of wireframing into hours.
- Promptframing emerged as a distinct practice in 2023 when teams discovered that annotating rough sketches with natural language instructions produced more relevant AI-generated layouts than text prompts alone.