The State of AI in Product Development
Figma's second annual AI report, based on a survey of 2,500 Figma users, paints a picture of an industry in motion. The survey found that one in three respondents are launching AI-powered products this year—a 50% increase over last year—and they collectively shared more than 1,000 distinct AI products in development, from predictive maintenance analytics to medical document interpretation. The findings highlight both the accelerating adoption of AI and the persistent tensions around quality, purpose, and workflow integration.
Agentic AI Emerges as the Fastest-Growing Category
While text generation remains the most commonly shipped AI product, the fastest-growing category is agentic AI. Twice as many Figma users are now building agentic products compared to last year: 51% of users working on AI products are building agents, up from 21% in 2024.
The appeal of agentic tools lies in their ability to handle multi-step processes rather than single-shot content generation. Rather than producing copy or images from a prompt, agents digest information, reason through options, and take action. That capability comes with significant design challenges. Product builders must decide when an agent should check in with a user, how much information to surface, and whether a chat interface is the right format—or whether a series of button commands would be more intuitive. These decisions demand the kind of critical thinking and prototyping expertise that designers and developers bring to the table.
Design Loops Still Matter
Despite the novelty of AI-powered products, the teams that succeed are those that hold onto established best practices while adapting them to the new technology. The report defines "successful teams" as those that shipped an AI-powered project that met or exceeded expectations. Among those successful teams, 60% agreed with the statement "We explored multiple design or technical approaches to the problem," compared to just 39% of unsuccessful teams. Teams that treated the generative AI product process as fundamentally different from non-AI work were also more likely to succeed.
That might seem counterintuitive, but it aligns with broader sentiment: 52% of AI builders say design is more important for AI-powered products than for traditional ones, and 95% say it is at least as important. Company leadership reportedly agrees. In an increasingly crowded marketplace for AI tools, thoughtful design is becoming the critical differentiator. One respondent captured the iterative nature of the work well: building an AI product is like "running a restaurant with a menu that changes daily."
Small Companies Push Harder
AI investment is not evenly distributed across company sizes. The proportion of Figma users at small companies who said AI was essential to their product doubled compared to last year. Among companies with 1-10 employees, 61% of Figma users said AI is "very or critically important" to their market share goals.
Smaller companies are generally more nimble and may be able to experiment with AI more rapidly than their larger counterparts. There may also be a sense that AI can accelerate business growth more decisively for a small team, making the investment easier to justify.
The Developer-Designer Disconnect
Adoption of AI tools is deepening across workflows, but a gap in perception has opened between developers and designers. Developers report higher satisfaction with AI tools (82%) and are more likely to say AI improves the quality of their work (68%). Among designers, those numbers drop to 69% and 54%, respectively.
The divide appears to stem from how each group actually uses AI. While 59% of developers use AI for core development tasks like code generation, only 31% of designers use AI in core design work such as asset generation. Developers are also more likely to work with AI output directly: 68% say they use prompts to generate code, and 82% report satisfaction with the results. Designers, by contrast, are still determining how these tools fit into their processes.
Progress Meets Uncertainty
Survey respondents see clear value in AI, with 78% agreeing that "AI significantly enhances the efficiency of my work." Yet only 32% say they can rely on the output of AI. This tension between efficiency gains and quality concerns is one of the central contradictions of the current moment.
The expectation that AI will reshape work is holding steady—and surprisingly moderate. Projections for AI's impact over the next year are not much higher than they were last year. Also, despite higher adoption rates, AI still runs a minority of projects: only 20% of designers and developers say that most of their work is AI-powered. A further challenge is that many AI projects lack clarity of purpose. Oonly 9% of builders name revenue growth as their top goal; 76% cite vaguer objectives such as "experimenting with AI" or "improving customer experience," which makes measuring impact difficult.
What remains consistent is the long-term outlook. More than 80% of both designers and developers say that learning to work with AI will be essential to future success in their roles. However the specific contours of that AI-integrated future develops, it will hinge on how effectively leaders manage the gap between AI's promise and its practical, trustworthy use—and how well they leverage the human experience and skills still required to build quality products.
