Figma AI (Make / First Draft)

Figma's AI accelerates designers inside the tool that defines the profession. First Draft generates starting layouts on your own design system rather than a generic reinterpretation, prompt-driven edits absorb the mechanical revisions, Make turns designs into living prototypes without a handoff, and the smaller assists handle content replacement, image cleanup and search across files. The tedious stretches compress; the taste stays yours. Designers working in Figma daily, crowded by mechanical work, are the audience. AI drafts are scaffolding rather than signature: distinctive visual identity still comes from the designer, heavy reliance on generation flattens work toward the average of what the model has seen, and a plausible first draft should never skip the critique a human draft would face.

Example tasks

Limits

Compares

Figma's AI lets founders produce credible product design before the design hire. First Draft lays out screens from a description, Make turns them into clickable prototypes investors and users can actually try, and the work accumulates in real Figma files your first designer will inherit rather than replace, which is the quiet difference from generating throwaway mockups elsewhere. It fits founders whose product needs to look designed while the design budget is zero. AI drafts start conversations rather than finish products, shipping-quality polish still wants human design sense, and a clickable demo reads as nearly-built to non-technical audiences, so caption the gap before it becomes a promised ship date. Non-visual founders may move faster in the app builders that skip design entirely.

Example tasks

Limits

Compares

Figma Make gets a PM to a working prototype without booking design or engineering time. Describe the flow, or start from the team's existing designs, and it produces an interactive front-end prototype inside the tool the design team already lives in, ready for user feedback or a stakeholder walkthrough the same day and looking like your product because it reads your files. For PMs who need ideas clickable, inside the team's Figma, it is the direct route. It builds front-end prototypes rather than products: no real backend, database or auth, so anything that must persist data graduates to a full-stack tool, and a convincing prototype reads as near-finished software to stakeholders, which makes labelling it clearly part of the job.

Example tasks

Limits

Compares

Figma Make is the prototyping route with the least friction for design teams. It reads your actual files, tokens and components, so prototypes come out looking like your product rather than a generic reinterpretation, and everything happens inside the tool designers already work in, with generated code available as an engineering starting point. Teams for whom fidelity to existing designs is the requirement get the category's best answer. The limits are structural: no backend, data persistence or auth, so prototypes that must behave like products graduate to the full-stack builders, generated code needs engineering review before production use, and a polished prototype invites stakeholders to assume the product is nearly done, so set expectations explicitly.

Example tasks

Limits

Compares

Figma's AI image tools cover the edits design work constantly needs. Background removal, upscaling of low-resolution assets and generated fill imagery run without leaving the canvas, which spares designers the round-trip to an editor for routine asset preparation and keeps the design file as the single working surface. Designers preparing assets inside Figma designs are the whole audience, and that boundary is the trade: these are canvas conveniences, so photographic retouching, compositing and darkroom work belong to the dedicated image software.

Example tasks

Limits

Compares