Does Figma Make speed up design task completion?

The paper reports a randomized controlled trial in which 50 product designers and 50 product managers attempted three standardized design tasks with or without access to Figma Make, a prompt-to-design tool that lets people build prototypes through conversational workflows with LLMs. "Among participants who completed the study tasks," access was "associated with approximately 20% shorter completion times," and the estimates account for demographics, professional experience, and AI tool familiarity. The gains were "larger among product managers." Participants with access also reported higher task ease and perceived usability.

The paper's own reading is narrow on the time claim and wider on the interpretation. The time savings are tied to "structured, reference-based design tasks," and the authors say plainly that "whether these benefits extend to broader design work remains uncertain." For professional designers, the abstract says only that benefits "may be task dependent." For product managers, the authors read the pattern as "productivity benefits beyond time savings": ease of use plus self-assessed design quality, since managers in the treatment condition "rated their designs more favorably relative to what they thought they could produce outside the study." Their proposed mechanism is that prompt-to-design tools "may help product managers translate their design intentions and ideas into interface changes." The paper presents this as a suggestion, not a tested pathway.

Against the neighbors, this is a measured speed gain, which sets it apart from Does AI really save time, or just change how we spend it?. That coding study found no significant difference in total time because saved work time moved into prompting and reading generations. Here total completion time did fall, though the excerpt says nothing about how time was spent inside the shorter total. It also qualifies When does AI actually boost worker productivity?: the larger gain went to the group with less professional design practice. But the trial times structured tasks and does not test skill acquisition, so the two findings do not directly conflict. For how non-specialists work with prompt-to-prototype tools, Where do vibe coding students actually spend their debugging time? describes the interaction pattern this trial's time measure would sit on top of.

The excerpt leaves several things open. The time estimate is conditional on completing the tasks, and it does not report how many participants finished in each condition. It gives no per-task breakdown, no size for the gap between managers and designers, and no independent rating of design quality; the quality evidence is self-assessed. It reports nothing on learning or retention, and nothing on what the tool did to the work itself. What the evidence supports is that prompt-to-design access can shorten structured design tasks, especially for product managers. It does not show that design work in general gets faster or that the output gets better, and time saved is a weak proxy for value, as the time-reallocation note argues.