The design process is dead. Here’s what’s replacing it. | Jenny Wen (head of design at Claude)

Mar 1, 2026 Episode Page ↗
Overview

Jenny Nguyen, Head of Design for Cloud Cowork and former Director of Design at Figma, discusses how the design process is rapidly changing due to AI, shifting from extensive mocking to supporting engineering execution and vision setting. She shares insights on new designer skills, maintaining craft, and effective management in this evolving landscape.

At a Glance
18 Insights
1h 17m Duration
17 Topics
4 Concepts

Deep Dive Analysis

Traditional Design Process is Obsolete

Two New Types of Design Work in AI Era

Widespread Impact of Design Process Shift

Day in the Life of an Anthropic Designer

Jenny Wen's AI Tool Stack

Figma's Continued Relevance for Design Exploration

Collaborating with Engineers in the AI Era

Maintaining Design Quality and Trust with AI Speed

AI's Potential for Taste and Judgment in Design

Durability and Evolution of Chatbot Interfaces

Transitioning from Design Director to IC

The Development Process of Claude Cowork

Three Archetypes for Hiring Designers

Career Advice for New and Senior Designers

High Value of 'Low Leverage' Manager Tasks

Encouraging 'Roasting' for Team Psychological Safety

The Legibility Framework for Ideas and Founders

Traditional Design Process

This process, often treated as gospel, involves extensive research, discovery, and iterative mocking/prototyping, but is now considered dead or dying due to the rapid pace of AI development and engineering capabilities. Designers no longer have the time to create beautiful mocks as before.

Two Types of Design Work

In the new AI era, design work is stratified into two main types: supporting implementation and execution (helping engineers ship quickly and cohesively) and creating vision or direction (setting a 3-6 month roadmap, often through prototypes, rather than long-term detailed plans).

Building Trust Through Speed

This concept suggests that releasing early-stage products, even with flaws, and then rapidly iterating based on user feedback, can build trust. The commitment to continuous improvement and responsiveness to users demonstrates that the company values their input and is actively working to enhance the product.

Legibility Framework

Developed by Evan Tana, this 2x2 framework categorizes founders and ideas as either legible or illegible. It suggests that truly novel opportunities often lie where ideas are 'illegible' (frontier, not yet understood by many), and a designer's role can be to spot these ideas and make them legible through storytelling and UX.

?
How is the design process changing with the rise of AI?

The traditional design process, which emphasizes extensive research and detailed mocks, is becoming obsolete. Design work is now split between supporting rapid engineering execution and setting short-term (3-6 month) strategic visions, often through quick prototypes.

?
What does a day in the life of a designer at Anthropic look like?

A significant portion of the day involves staying updated on internal developments, model advancements, and team prototypes. Designers also spend time thinking about future directions, jamming with engineers, and directly implementing polish or code tweaks.

?
Why does Figma still matter for designers in the AI era?

Figma remains crucial for exploring numerous design options and iterating on fine visual and interaction details. Coding tools are often too linear for broad exploration, making Figma ideal for brainstorming and micro-level design decisions.

?
How can designers maintain craft, quality, and trust in the AI era of rapid shipping?

Designers can maintain quality by clearly labeling early releases as 'research previews' and committing to continuous iteration based on user feedback. The key is to show users that their input is valued and that the product is constantly improving, building trust through speed and responsiveness.

?
Will AI ever develop 'taste' and 'judgment' in design?

AI is expected to get better at taste and judgment in design. However, humans will likely still be needed to make final decisions, be accountable for those decisions, and resolve disputes, as these are complex social and ethical challenges beyond current AI capabilities.

?
What is the future of chatbot interfaces?

Chatbot interfaces are likely to be durable, offering infinite flexibility for interacting with models. They will probably be combined with more tactile UIs and widgets, many of which will be AI-generated, providing a blend of conversational and direct interaction.

?
Why did Jenny Wen transition from a Director role at Figma back to an IC role at Anthropic?

Jenny Wen returned to an IC role to stay close to the rapidly changing design work and gain new hard skills, believing it's a critical time to be hands-on. This experience provides empathy and understanding for how the design process has evolved, which she feels is essential for future management.

?
What three archetypes is Jenny Wen looking for when hiring designers?

Jenny looks for 'strong generalists' (block-shaped, 80th percentile in a few core skills), 'deep specialists' (t-shaped with exceptional depth in one area, top 10% of the industry), and 'cracked new grads' (early career, wise beyond their years, humble, and eager to learn new tactics without ingrained processes).

?
What advice does Jenny Wen have for new designers trying to break into the industry?

New designers should focus on building a lot of actual things and trying out new technologies. Being unburdened by past expectations, they can explore possibilities freely, build a portfolio, and find a community that encourages continuous building and sharing.

?
What is the value of 'low leverage' tasks for managers?

Managers taking on 'low leverage' tasks, such as thoroughly dogfooding the product, fixing minor bugs, or creating thoughtful anniversary cards, can actually be highly leveraged. It demonstrates deep care, familiarity with the product, and a 'we're all in this together' attitude, fostering trust and respect within the team.

?
How can encouraging 'roasting' among team members contribute to building strong teams?

While not forced, a team's comfort with 'roasting' each other can be a sign of strong psychological safety and trust. It indicates that team members feel secure enough to push boundaries and that they don't fear their leaders, creating an environment where direct feedback and high standards can be applied effectively.

1. Adapt Design Process

Let go of the traditional, extensive design process focused on beautiful mocks and prototypes, as the rapid pace of engineering with AI tools now requires designers to primarily support implementation and execution.

2. Short-Term Vision Setting

When creating design visions, focus on shorter timeframes (3-6 months) and practical prototypes that point people in the right direction, rather than elaborate, long-term decks, because technology changes too rapidly.

3. Ship Early, Iterate Fast

For non-deterministic AI products, prioritize shipping early as a ‘research preview’ to gather real user data and feedback. Continuously iterate and visibly respond to user input to build trust and improve the product, rather than striving for perfection before launch.

4. Designers Embrace Coding

Actively use coding tools to engage in last-mile implementation, polishing, and prototyping in actual code. This allows for closer collaboration with engineers and faster feature delivery, reducing reliance on engineers for initial prototypes.

5. Managers Do IC Work

Design managers should engage in individual contributor (IC) work to gain practical skills and empathy for the rapidly changing design process. This hands-on experience is crucial for effectively guiding teams in the current environment.

6. Balance Safety & High Standards

As a manager, balance fostering psychological safety with maintaining high standards for work quality. When trust is established, direct challenging becomes easier and more effective, leading to better work without fear.

7. Spot Illegible Ideas

Actively seek out ‘illegible ideas’—those on the frontier that people don’t yet understand but have underlying energy. Dive deeper to comprehend their potential and then transform them into clear, actionable concepts through UX, form factor, or storytelling.

8. Hire Resilient, Adaptable Designers

When hiring designers, prioritize candidates who demonstrate resilience and adaptability, showing a willingness to learn new tools and methods rather than clinging to outdated processes, which is crucial for navigating rapid industry changes.

9. Hire Strong Generalists

Seek out ‘strong generalists’ who excel at multiple core design skills (80th percentile good). Their broad, deep skill set allows them to flex and expand their role as design responsibilities evolve.

10. Hire Deep Specialists

Look for ‘deep specialists’ who are among the top 10% in a particular design skill (e.g., technical design, visual design, icon design). Their unique expertise can significantly differentiate products in a world where AI can generate many things.

11. Hire ‘Cracked New Grads’

Don’t overlook ‘cracked new grads’—early-career designers who are wise beyond their years, humble, and eager to learn. Their blank slate and quick learning ability are invaluable in a rapidly changing field, free from ingrained outdated processes.

12. New Designers: Build Things

For new designers, focus on actively building and experimenting with technology to create actual products. This hands-on approach, unburdened by traditional expectations, is key to standing out and learning effectively.

13. Prioritize ‘Low Leverage’ Tasks

Leaders should intentionally take on seemingly ’low leverage’ tasks, such as thoroughly testing the product or personally creating thoughtful gestures for team members. These actions demonstrate deep care, build trust, and foster a strong team culture, ultimately having a high impact.

14. Explain Design Decisions

When consulting with engineers on projects, explain the ‘why’ behind design decisions to help them understand underlying principles. This equips them to make better design choices independently in the future.

15. Integrate Coding Tools

Designers should integrate coding tools into their toolkit and become proficient in using them, even if not aiming to be full-fledged technical coders. Awareness and practical use of these tools are essential for the evolving design landscape.

16. Foster Playful Psychological Safety

Encourage a level of psychological safety where team members feel comfortable playfully ‘roasting’ each other and their leaders. This indicates a high degree of trust and comfort within the team, fostering a more open and collaborative environment.

17. Equip Engineers with Design Systems

Actively guide engineers to use the existing design system in code, especially when AI tools are generating code that might not adhere to it. This ensures consistency and quality across the product.

18. AI for Introspection

Leverage AI tools like Claude Co-Work for personal introspection and analysis by feeding it diverse personal notes (e.g., one-on-ones, random thoughts, memos). This can help extract insights, create rubrics, and learn new things about oneself implicitly.

You as a designer actually like do not have the time to make these beautiful mocks anymore.

Jenny Wen

At the end of the day, someone has to decide what is actually going to get built and what actually matters.

Jenny Wen

I don't think chat is ever going away because this opened up this like new way of like infinite ways to to work with the model and to sort of like talk to the computer that we just didn't have before.

Jenny Wen

I think it can be a really good sign when the people on your team kind of feel comfortable just like kind of poking fun at each other a little bit.

Jenny Wen

I think that ends up being like super actually high leverage even though it's a lot of time of like nitty-gritty time because it creates this like familiarity with the product which I think is really good.

Jenny Wen
60 to 70%
Proportion of design work spent mocking and prototyping a few years ago Jenny Wen's estimate for her own experience
30 to 40%
Current proportion of design work spent mocking and prototyping Jenny Wen's estimate for her current experience
3 to 6 months
Typical timeframe for a design vision in the AI era Reduced from multi-year visions due to rapid technological change
10 days
Time taken to get Claude Cowork from internal prototype to external release Refers to the final push to ship, after a longer internal development period
12-15
Maximum number of designers Jenny Wen managed at Figma Including a few managers reporting to her