Designers are excited about AI, but most don’t know what to do with it

Artificial intelligence. It’s pretty cool, I guess? Look at those neat videos. And the thousands of product design iterations just to get those creative balls rolling. Sure. Awesome. Or is it? Maybe. Who knows. All that seems to be the summary of Figma’s 2025 AI Report, based on a survey of 2,500 designers and developers.

While tools like ChatGPT and Figma’s AI features are embedded in daily workflows, the report reveals a stark disconnect. Enthusiasm for AI’s potential is high, but its practical impact remains uneven, the numbers show, constrained by vague goals, quality concerns, and cooling expectations. The report underscores a paradox: professionals see AI as essential to their future, but struggle to meaningfully harness it today. It kind of fits my own experience. It’s there, but not there yet.

Figma’s study shows that a staggering 76% of AI projects prioritize vague objectives like “experimenting with AI” over concrete goals such as revenue growth, with an eye-popping 9%. It makes me weep for all the gigawatts evaporating in the name of a revolution that’s not actually happening, at least for designers and developers. 

The ambiguity reflects the technology’s nascent state, Figma’s Head of Insights Andrew Hogan tells me in a phone interview. “There’s a lot of play and experimentation happening—it’s natural,” he explains, comparing the current moment to early mobile app development, where rapid iteration preceded clear use cases. One survey respondent likened building AI products to “running a restaurant where the menu changes daily,” a metaphor Hogan calls “the quote of the survey.”

So much contradiction

I’m not so sure about that parallelism with mobile app development, which struck me as a much faster, much more impactful revolution than AI, in practical, tangible economical terms, not just paper gains. Past technological shifts, like desktop publishing or the iPhone, delivered seismic industry changes within months. By comparison, AI’s impact feels incremental and anecdotal. Sure, there are brilliant examples of big AI impacts in some industries—mostly audiovisual—but having a synthetic research minion, a repetitive-task assistant, or an artificial creative buddy don’t seem quite as revolutionary as a billion smartphones taking over our lives. 

Hogan acknowledges the tension and, at the same time, has a warning: Companies risk dismissing AI too early if experiments fail to yield quick wins, potentially missing strategic advantages. He also says that, while the research highlights these contradictory data points between expectations/desires and reality, the data shows real progress: 34% of Figma users shipped AI products this year, up from 22% in 2024. The question is whether the vague goals—again, back to the figure of 76% of companies saying let’s play, throw some mud against the wall and see if it sticks—will harden into measurable ROI before disillusionment sets in.

The research shows that there are efficiency gains thanks to AI. But there’s a dichotomy here, too. Seventy-eight percent of professionals say it speeds up their work (up from 71% last year), but only 58% believe it improves quality, while 47% feel it makes them better at their jobs. What about the ones who think the quality is just the same or worse, and the 53% who don’t think AI makes them better at their jobs? It’s a strange, puzzling juxtaposition. Developers report higher satisfaction (67% say AI boosts work quality) than designers (40%), partly because code generation tools offer clearer utility. Designers, meanwhile, grapple with generative AI’s unpredictable outputs.

Hogan attributes this gap to the “limitations of how we as humans interact with these things,” not the technology itself. He cites Amara’s Law: We overestimate short-term change and underestimate long-term transformation. “Mobile took years to reshape industries,” he says, pointing to Uber’s evolution. Yet tools like ChatGPT sparked expectations of rapid, iPhone-level disruption—a bar AI hasn’t yet cleared.

Cooling expectations

Despite 85% of professionals calling AI essential for future success, expectations for its near-term impact are cooling. Only 27% predict AI will significantly influence company goals in the next year, unchanged from 2024. Hogan frames this as a recalibration, not disillusionment. “The hype gets ahead of what most people can do today,” he says, likening AI’s trajectory to the internet’s gradual adoption. Yet the “Cambrian explosion” of AI apps—like the one that happened with the iPhone’s apps—is yet to come. Sure, there are niche applications like medical document interpreters or predictive maintenance tools, but where are the truly transformative apps beyond being able to talk to glorified Wikipedia oracles? Where’s the Uber of AI?

The answer may lie in agentic AI, the fastest-growing product category. These tools, which automate multistep tasks, saw a 143% year-over-year surge in development (from 21% in 2024 to 51% in 2025). But Hogan warns they require rethinking design principles. “When should an agent check in with users? What information should it share?” Design’s role here is critical—52% of builders say design is more important for AI products than traditional ones, as intuitive interfaces bridge the gap between capability and usability.

AI’s paradox—ubiquitous yet underutilized, and underwhelming for a large part—stems from its adolescence. Designers and developers are caught between excitement, collective hysteria, and pragmatism, navigating a landscape where prototyping and iteration matter more than ever. The technology’s potential is real, yes. Code generation already accelerates development, and is used by 59% of developers. Agentic tools promise workflow revolutions, and adoption is rising. But without clearer goals, trust in outputs, and design-led refinement, AI risks becoming a toolbox without a blueprint.

As Hogan puts it, “We’re still early.” The challenge isn’t whether AI will reshape design, but whether teams can evolve their processes fast enough to meet its uneven promise. For now, the future belongs to those who treat AI not as a magic wand, but as clay—malleable, demanding, and far from fully molded.

View the full article