Sign InOpen Brain
arXivPaperNeeds Review

Does AI Save Time on Product Design? A Randomized Controlled Experiment of AI Prompt-to-Design Workflows

In a randomized trial, Figma Make cut completion time by about 20% among finishers, with larger gains for product managers and less certain benefits for professional designers.

arXiv · Sep 22, 2026
Open Source Open MarkdownOpen JSON
Source Summary

A randomized trial assigned **50 product designers** and **50 product managers** to three standardized tasks with or without Figma Make. Among participants who finished, tool access was associated with about **20% shorter completion times**.

Practical Implication

Solo builders can use prompt-to-design workflows to accelerate prototype production and let non-design specialists contribute more directly. The larger reported gains for **product managers** suggest the tool may help most when design execution is not already a core skill.

Agent-Ready Context
A randomized trial assigned **50 product designers** and **50 product managers** to three standardized tasks with or without Figma Make. Among participants who finished, tool access was associated with about **20% shorter completion times**.

Solo builders can use prompt-to-design workflows to accelerate prototype production and let non-design specialists contribute more directly. The larger reported gains for **product managers** suggest the tool may help most when design execution is not already a core skill.

The result covers completers rather than every enrolled participant, and the supplied material gives no quality outcome. Benefits for professional designers were **task dependent**, so speed should not be treated as a general productivity or design-quality gain.
Connected Context · Feed7 Judgment

This adds randomized evidence that prompt-to-design tools can shorten standardized prototype work for completers, especially non-design specialists. It does not validate the candidates’ stronger claims about taste, coherence, or scalable quality because no quality outcome is supplied and professional-designer gains varied by task. Speed and design quality therefore remain separate evaluation axes.

The Missing Layer: Design Taste in AI Agents — Hassan El Mghari, Together AIThe trial measures production speed, while this candidate treats generated UI as a draft requiring references and iteration; together they separate faster first output from finished design quality.One Designer + AI. Hundreds of Deliverables. — Vincent Wendy, AI EngineerThe solo-designer case describes how primitives, structured data, manual exceptions, and visual QA can govern scaled output, providing operational controls absent from the trial’s speed result.The Signal Layer: What to Build When Anything Can Be Built — Lena Hall, AkamaiThe randomized speed gain reinforces the premise that implementation is becoming easier, while this candidate argues that problem selection, evidence, and product judgment remain limiting work.The End of the Static Screen: Architecting Intent-Driven UX — Gus Iwanaga, commercetoolsThe trial supports faster prompt-driven prototyping but does not test repeated-interface stability; this candidate identifies schemas, layout rules, and design-system constraints as prerequisites for that broader use.
Context Map
craft#design-engineering#dev-ux
Uncertainty
The result covers completers rather than every enrolled participant, and the supplied material gives no quality outcome. Benefits for professional designers were **task dependent**, so speed should not be treated as a general productivity or design-quality gain.