Figma researchers found that access to its Make AI tool shortened the time needed to finish fixed interface-design tasks, with the largest gains among product managers. The randomized trial involved 100 product designers and product managers; its time comparisons include only tasks that participants completed.
Figma Make turns text instructions into editable interface designs. In the study, participants were randomly assigned within each job role to use Make or edit manually in Figma Design. Both groups started with the same social-media screen and reference images. They attempted to create a dark-mode version, add a settings-menu item and build a comment flyout, always in that order. Make users could refine the generated result by hand.
Product managers saw the largest time savings
Across completed task attempts, with the three tasks weighted equally, the researchers estimated about 20% less time with Make access. The dark-mode task alone showed no clear time difference. The more involved comment-flyout task accounted for much of the gain.
For product managers, the model estimated 22 minutes 46 seconds less cumulative time across the three tasks, a 35% reduction based on completed-task data. Designers' gains were smaller and clearest on the comment flyout, where the estimated saving was 7 minutes 41 seconds. Make access also raised the estimated chance of completing that task by 19.9 percentage points across both roles.
The trial measured fixed tasks, not full projects
Four Figma employees wrote the study, while an external research firm recruited and moderated the participants. Each participant had at least two years of experience in their role and used Figma at least monthly. Moderators timed tasks until submitted designs met the study team's reference-based completion criteria.
The time analysis excludes unfinished attempts even though Make changed completion rates on the hardest task. The researchers also did not have independent experts judge final design quality. Because all participants followed three prescribed tasks, the results do not establish the same time savings for open-ended projects with changing requirements, review and revision.