Users desire Figma to provide more efficient and integrated tools for 'designing with code' or 'prototyping in code' that do not require designers to learn traditional coding. The current tooling is perceived as less efficient than external AI-assisted code-based design tools. This would allow designers to work directly with semantic frontend code concepts within Figma, similar to how they currently use Figma's visual tools.
There are 2 groups of people who say Figma is worthless now, post-AI: Group 1: The non-UI designers. They look at the UI's produced by AI and think they're great, when they're really not. They might be acceptable, but they are not top-tier. AI growth in the UI design space has lagged behind AI growth in development/programming. At best, you will get cookie cutter. But for those who want to outshine their competition in terms of UI/UX, the human eye for design is still a necessity. Group 2: Designers who design in code. "Designing in code" is not a new thing, it was an idea around long before AI became so prolific. Designing in code back in the day meant painstakingly having to hand write HTML/CSS. Iterating (a very necessary step of UI design) would cause too much friction, due to the slow nature of code. Fast forward to today, and 'designing in code' now simply means instructing an AI to make design/style changes. It's faster than writing code, for sure, but it's not as fast as I am in Figma, and it's not as creative as I am in Figma. Having to verbally speak (or type) instructions to AI in order to get it to hopefully produce design that I want is far inferior than me just making a quick design iteration/experimentation in Figma. BUT, I am not entirely against using AI to generate designs. Recently, I've found myself using a hybrid approach. I first start off by designing the UI in Figma. Not every screen, maybe just a page or two. I establish the colors, typography, embellishments, styles, alignment practices, white space, etc.. Then, I use AI to build that UI in the frontend, and I start developing a style guide that the AI uses. As long as that style guide is robust enough, I can then instruct AI to generate unique pages and components on its own. It gets it right most of the time, but you still need an eye for design to ensure that what it does produce is truly good and consistent. I augment this process sometimes by using Figma to design areas that require more unique thought, and I also use tools like unicorn studio, rive, and greensock (or framer motion), to elevate the UI/UX in ways that non-UI designers who rely on AI cannot. This is a competitive advantage. So, Figma is still a very necessary tool in my toolbelt. The 30 Days UI challenge releasing this week is tailored to help you develop that eye for design that is still very much needed.