[Edaily Reporter Lee So-Hyun ] While artificial intelligence (AI) has lowered the barrier to entry for app and content creation, allowing anyone to quickly produce plausible results, an analysis suggests that the number of products users actually use frequently and prefer has not increased proportionally. Critics point out that if users become satisfied with and accept AI-generated results, they may end up relinquishing their own judgment.

Yuki Yamashita, Chief Product Officer (CPO) of Figma, said at a press conference held on the 14th at the Conrad Seoul in Yeouido, Seoul, “The moment you look at an AI-generated result and think, ‘This is good enough,’ you give up your own judgment,” adding, “We call this ‘quiet surrender.’”

Citing the fact that while the number of new apps appearing on the App Store has surged recently, the number of apps that users frequently use and love has not changed significantly, he pointed out, “Even if it looks like the best result you could create yourself, it might just be a new ‘average’ that anyone can produce.” In other words, while AI has accelerated product development, it has also led to an increase in average results that are no better than existing products.

CPO Yamashita emphasized, “While the number of apps has increased, differentiation has decreased, and it has become harder to capture people’s attention,” adding, “We must not simply accept AI-generated results at face value; instead, we must incorporate human judgment and taste to create unexpected experiences and a higher level of polish.”

Figma is a cloud-based collaboration platform for designing interfaces for websites and mobile apps and creating prototypes. It is characterized by the ability for members of various disciplines to work on the same file in real time, and recently, it has expanded its scope to become a product development platform that integrates code and generative AI. The company went public on the New York Stock Exchange last year.

AI Blurs Job Roles… “Being Able to Do Something Is Different from Doing It Well”

With the spread of AI, the boundaries between the roles of designers and developers are rapidly blurring. According to Figma’s 2026 AI Report, the percentage of developers involved in design work rose from 44% in 2025 to 60% this year. The percentage of designers involved in development work also nearly doubled during the same period, from 21% to 41%.

Korean product developers also reported a high level of AI impact. Seventy-six percent of Korean product developers said AI had a significant impact on their work over the past year, surpassing the 64% reported in the U.S.

Woo Sang-hoon, Head of the Content Production Tools Lab at Naver (NAVER(035420)) and a Figma partner who participated in the event, stated, “Being able to perform the work of other job roles using AI is different from doing it well.” He added, “The more we live in an era where average results are easily produced, the more important it is to have a firm sense of taste and the ability to distinguish what is truly good.” He noted that Naver is testing a plan to enhance the organization’s overall capabilities by standardizing individual AI usage methods and prompts at the team level.

CPO Yamashita introduced new features unveiled at Figma’s annual conference, “Config 2026,” along with the company’s product development strategy for the AI era. As a solution to transcend the “new average” created by AI, he presented a strategy to connect design, code, motion, and generative AI on a single canvas.

One of the key new features, “Code Layers,” allows users to import working code into a design file as if it were a layer, enabling team members to copy and modify it collaboratively. It also supports bidirectional workflows, allowing users to convert code-based outputs back into editable designs and to reflect design changes back into the code. CPO Yamashita stated, “Code is not a realm that stands in opposition to design; it is a new material,” adding, “Rather than replacing developers, it is a means to describe desired results more precisely.”

Figma also introduced “Motion” for creating animations, “Shader” for generating visual effects from text, and “Weave” for connecting various image and video-generating AIs. A key feature is that users can directly modify and control AI-generated drafts rather than using them as-is.

Leader Woo explained that Naver used Weave to create an internal prototype that visually represents users’ data on saving, booking, visiting, and reviewing locations, adding, “Even with a small team, we were able to flesh out the concept and service process faster than expected.”

“Individuals Have Become Faster, but Teams Have Become Isolated”… Figma’s “Multiplayer AI”

CPO Yamashita pointed out that while AI has increased the speed of individual work, it has also created new “silos” where team members interact with their own agents without sharing their results. “In the AI era, we need to be able to see not only what my agent is doing but also what my colleagues’ agents are doing,” he said, emphasizing “multiplayer AI,” where humans and AI can view and edit each other’s work on the same canvas. While existing coding tools typically require individuals to write code and then submit their results, Figma’s vision is to enable multiple people to collaborate in real time from the early stages of idea exploration.

He also spoke highly of Korean companies’ commitment to AX (AI Transformation). CPO Yamashita remarked, “In the past, I felt that the maturity of design and product development at Asian companies, including those in Korea, lagged behind the U.S. by three to four years, but that has changed,” adding, “Korean companies are seeking to transition their work methods to be AI-native in order to go beyond catching up and become the best in the world.”

He advised that managing corporate AI costs requires direct human editing and the sharing of results within the organization. “There are times when it’s faster and cheaper for a person to make direct revisions than to keep asking AI to improve the results,” CPO Yamashita said, emphasizing that organizations must share the results of repetitive tasks and the relationships between them.

He also dismissed the so-called “SaaS-pocalypse” prediction that Agent AI would replace Software as a Service (SaaS). “As agents’ tasks become more complex, we need visual interfaces that allow humans to understand and review the process,” he said, emphasizing, “SaaS will evolve, but it won’t disappear.”