After Adobe decided not to pressure the $20 billion acquisition of Figma due to regulatory issues stemming from a lack of understanding of how software is created, people eagerly awaited Figma’s Config 2024, particularly regarding its AI plan.
Figma’s strategic approach in response to industry demands
This is my understanding after having conversations with Figma’s executives during the conference; CEO Dylan Field, CTO Kris Rasmussen, VP Of Product Sho Kuwamoto, VP Of Design Noah Levin, Product Manager Katie Szeto, etc.
Numerous AI-related announcements, demonstrations, and discussions confirmed Figma’s commitment to the development of AI while proving more than capable of surviving without Adobe’s AI experience.
To be more precise, as Dylan Field showed in the opening keynote, the new “Make Design” feature accepts an ordinary language prompt and generates, along with the text, the first sketch of a potential mobile app or website design, images included, using a diffusion model.
It also allows users to continue with the design using additional prompts. The new “Make Prototype” feature can then build clickable UI controls and interconnections that transition from one screen to the next.
Some examples of repetitive work that can be done away with are text creation and layer renaming. The same can be done with other types of text in Figma, for example, a brief description of a recipe, and translating it. It also can replace general layer names with meaningful ones.
Allow Figma to utilize customers’ designs for AI training through registration. To turn it on or off, Figma will limit the ability to change this setting only to a customer’s admin users, not designers, via specific settings.
Dylan pointed out the fact that sometimes the results could be quite distorted. And while the demos appeared close to magical most of the time, there were, of course, duds on hand. The takeaway is the same as for every company weaving AI capabilities into its software: AI does not entirely remove the involvement of people in the process.
A detailed conversation on Figma’s response to their feature
Figma highlighted some that are not in beta but are being worked on including a way for the enterprises to use their design systems to power the “Make Design” feature — a capacity that Figma is piloting with Google designers.
Expect some minor hitches and be ready to make some regressions. Last week, days after the event, a user who tried using “Make Design” to make a weather app saw that the result was nearly the same as the Weather app in iOS and questions of piracy arose.
During the demonstration of these readily available new features, Dylan pointed out that the results are not always accurate. But even if the demos appeared almost like witchcraft for a large part of the time, there were occasional flops. The takeaway is the same as for every company weaving AI capabilities into its software: AI does not do away with the need to engage people in the process.