As of 13 August 2026, AI can make a prototype in Figma.

What to actually do

1. Use a tool built for this

   The route this page recommends

2. Do it yourself

   Second choice

   A chat interface, chat-fluent skill, and roughly 1 hour until you can act on the result.

   How to actually do it

   1. Open the product brief, user research, current copy and any brand guidance, then gather the primary user goal, main journey, required states and device requirements.
   2. Paste those materials into the prompt, replacing every bracketed slot, and ask the chatbot to produce the screen structure and interaction map.
   3. Open Relume and enter the same product description, target user and main journey to generate the initial sitemap and wireframes.
   4. Compare Relume's generated screens with the chatbot's screen list, remove screens that do not support the main journey, and add any missing error, empty or success states.
   5. Export the selected Relume wireframes to Figma and rename the pages and frames to match the agreed screen list.
   6. In Figma, add the supplied copy, connect the primary buttons and links in Prototype mode, and create the essential validation and completion paths from the interaction map.
   7. Run the prototype from the first screen using the main journey, then compare every label, destination and state against the brief before sharing the Figma link with a colleague.

   Prompt

   Act as a UX prototyping assistant. Turn the following product brief into a low-fidelity clickable prototype plan for Figma.
   
   Product or feature: [describe it]
   Target users: [describe the users]
   Primary user goal: [describe the goal]
   Main journey: [list the steps from entry to completion]
   Required content and data: [paste the content or describe what must appear]
   Brand or visual guidance: [paste any guidance or write none]
   Device and accessibility requirements: [describe them or write none]
   
   Produce:
   1. A screen-by-screen structure with a unique name for every screen.
   2. The purpose of each screen and the user action it supports.
   3. The exact headings, labels, button text, error messages and empty states needed, using invented placeholder content only where I have not supplied copy and labelling it clearly as placeholder text.
   4. The interaction map, including every button, link, back action, validation state and success or failure path.
   5. A list of components and reusable patterns to create in Figma.
   6. A short list of assumptions and unresolved product decisions.
   7. A concise checklist for testing the prototype with a colleague.
   
   Keep the flow to the minimum number of screens needed to demonstrate the main journey. Do not invent business rules, prices, legal claims or user research findings. Separate confirmed requirements from assumptions. Format the answer so I can build it screen by screen in Figma.

   Open it prefilled in ChatGPT or Claude, or copy it into Gemini, which takes no prefill link.

3. Hand it to a person

   The distant third

   A person who owns the outcome does this end to end, worth it when the failure is dear.