AI Redefines Design Roles and Product Velocity
Jenny Wen, Head of Design at Anthropic, explains how AI agents are collapsing the traditional design process. Learn why mocking is declining, how designers must pivot to execution and vision, and the new hiring archetypes for the AI era.
The Collapse of Traditional Design
The traditional design process, characterized by extensive research, discovery, and long-term visioning, is rapidly becoming obsolete. Jenny Wen, Head of Design at Anthropic, argues that the speed of AI engineering has forced a structural shift in the design role. Designers can no longer afford to spend months on beautiful mocks; instead, they must operate in a cycle of rapid prototyping and execution. The focus has shifted from creating a 5-year vision to guiding teams through 3-6 month directional prototypes that align with the current capabilities of non-deterministic AI models.
From Gatekeepers to Collaborators
The designer's role has stratified into two primary functions: supporting implementation and creating directional vision. Previously, 60-70% of a designer's time was spent on mocking and prototyping. Today, that figure has dropped to 30-40%, with the remainder dedicated to pairing with engineers, implementing code, and consulting on execution. This shift requires designers to become more technical, using tools like Claude Code to polish features and ship directly. The goal is not to block engineers but to help them navigate the rapid development cycle, ensuring cohesion and quality without becoming a bottleneck.
Strategy for Trust and Hiring
In an environment where products ship at high velocity, trust is built through speed and transparency. Launching features as 'research previews' with a clear commitment to iterate based on user feedback is a viable strategy for maintaining brand integrity. This approach allows companies to learn from real-world usage of non-deterministic models, which cannot be fully mocked. Regarding talent, hiring criteria are evolving. Wen identifies three key archetypes: strong generalists with multiple 80th-percentile skills, deep specialists with rare expertise, and 'craft new grads' who are eager learners unburdened by legacy processes. Resilience and the ability to adapt to changing tools are now more critical than mastery of static design methodologies.
Leadership and Cultural Dynamics
Effective leadership in this new era involves embracing 'low-leverage' tasks. Managers who personally test products, fix bugs, and engage in detailed feedback loops demonstrate high care and build cultural trust. This approach, combined with psychological safety that allows for candid feedback and even 'roasting,' creates high-performing teams. Finally, designers must adopt a VC-like mindset, using frameworks like 'legibility' to identify and decode confusing but promising internal prototypes, transforming illegible ideas into clear product directions.
Key insights
-
The traditional design process of long-term discovery and visioning is dead. AI speed forces a shift to 3-6 month directional prototypes that guide execution rather than dictate final aesthetics.
Impact: Teams that abandon rigid discovery phases will ship faster and adapt more effectively to non-deterministic AI capabilities.
-
Designers are transitioning from mock creators to code implementers. The proportion of time spent on mocking has dropped from 70% to 30%, with the rest dedicated to pairing with engineers and shipping code.
Impact: Designers who learn to code and use AI agents will become more valuable, bridging the gap between design intent and technical reality.
-
Trust is built through rapid iteration and transparent communication. Launching 'research previews' with a clear promise to fix issues based on feedback is more effective than delaying launch for perfection.
Impact: Companies that ship early and iterate visibly will retain user trust better than those that delay releases for unattainable polish.
-
Hiring priorities are shifting toward resilience and adaptability. The most valuable candidates are strong generalists, deep specialists, and eager new grads who can navigate rapid tool changes without being stuck in legacy processes.
Impact: Organizations that hire for adaptability over static skill mastery will build teams capable of surviving the rapid evolution of AI tools.
-
Managers should perform 'low-leverage' tasks like testing products and fixing bugs. These actions signal deep care, build cultural trust, and provide leaders with direct product familiarity that pure people management cannot offer.
Impact: Leaders who roll up their sleeves create higher-performing teams by demonstrating shared commitment and maintaining a direct connection to the product.
Action items
-
Reduce time spent on high-fidelity mocks and increase time spent pairing with engineers. Use AI coding tools to implement and polish features directly rather than handing off static designs.
Impact: This accelerates the feedback loop and ensures design intent is preserved through the implementation process.
-
Adopt a 'research preview' launch strategy for AI features. Clearly communicate that the product is early and commit to rapid iteration based on user feedback.
Impact: This manages user expectations and builds trust through visible responsiveness, turning early adopters into advocates.
-
Update hiring criteria to prioritize resilience and adaptability. Look for candidates who can demonstrate a willingness to learn new tools and processes quickly, rather than just deep expertise in legacy design methods.
Impact: This ensures the team can keep pace with the rapid evolution of AI tools and methodologies.
-
Implement a 'legibility framework' for internal prototypes. Encourage designers to identify and decode confusing but energetic ideas, acting as VCs to transform illegible concepts into clear product directions.
Impact: This uncovers hidden innovation opportunities and ensures that promising but poorly articulated ideas are not overlooked.
-
Encourage managers to engage in 'low-leverage' tasks such as testing the product, fixing minor bugs, and providing detailed feedback. Use these actions to build cultural trust and product familiarity.
Impact: This demonstrates leadership commitment and creates a culture of shared responsibility and high standards.
Quotes
“This design process that designers have been taught, we sort of treated as gospel. That's basically dead.”
“A big part of the design role now is helping engineers and teams execute, not just telling them here's the design.”
“At the end of the day, someone has to decide what is actually going to get built and what actually matters.”