Dive Club 🤿 – "How AI is Changing Design Workflows"
Host: Ridd
Guests:
- Henry Modiset (VP of Design, Perplexity)
- Pranathi Perry (Lead Designer, V0 at Vercel)
- Nick Pattison (Brand Designer, Lovable, Craft Maven)
Date: October 29, 2025
Episode Overview
In this live panel, Ridd brings together three influential design leaders to dissect how AI is reshaping design workflows—from branding and product prototyping to cross-functional collaboration and the evolving technical expectations for designers. The discussion drills down into hands-on workflows, real tools, and how emerging tech is changing the very nature of design deliverables and what it means to be a designer in 2025.
Key Discussion Points & Insights
1. Building AI-Powered Tools for Branding
Speaker: Nick Pattison
[00:10 – 10:17]
-
Workflow Evolution: Nick shares how his team is now building internal generative tools for clients as part of the brand handoff, making scalable, fast visual systems that clients themselves can continue to tweak.
- Example: "For a payments processor called Flow Glad, inspired by Islamic geometry and subway lines, we built a pattern architect tool in about an hour. It lets us—and clients—play with patterns, radius, density, etc." – Nick [02:20]
- Delivering these tools turns a brand sprint into an ongoing, interactive experience for clients.
- Handoff is enhanced: “The most exciting thing is on day three, I'm like, we built a tool for you. If you like this, the tool is yours.” – Nick [08:23]
-
Experimentation Supercharged: Teams develop multiple pattern tools in days, then move quickly from experimentation to polished deliverables, reducing dependence on found inspiration.
“We're starting to find that we can produce things faster than saying, hey, what if we did it like these other designers who did it? So our mood boards—I'm just showing you—looked actually quite polished.” – Nick [04:37]
2. Prototyping with AI: Beyond the Designer Stereotype
Speaker: Pranathi Perry
[11:14 – 18:22]
-
V0 as a Bridge: Tools like V0 serve as a middle ground—less rigid than engineering, more interactive than Figma.
- Useful for PMs and non-designers to create functional mockups and "personal software"—not just aesthetics, but real, data-driven prototypes. [11:14]
-
AI for Dynamic Asset Creation:
- Pranathi demonstrates building dynamic SVG assets for her portfolio, using natural language prompts in V0 to tint SVG gradients, create transparency effects, and assemble sprite tools for future game design.
- “I just put in: ‘Hey, can you tint all of the gradients in this SVG to the color I want?’ It became this fun tool for generating different variations.” – Pranathi [13:03]
- Using natural language, designers can iterate and debug with AI assistance—no need for manual code changes unless necessary.
-
Iterative Prompting:
- Debugging flows by providing detailed feedback to the AI.
“I try to test the limits of V0. I won't edit the code at all, I'll do everything in chat just to see how close I can get from spoken word.” – Pranathi [15:52]
- Debugging flows by providing detailed feedback to the AI.
3. Expanding the Designer's Scope
Speaker: Henry Modiset
[19:01 – 24:04]
-
Designers as Multi-Hyphenates:
- “Being a designer actually means you have potentially a lot of different roles … sometimes it's problem solving, sometimes it's communication, sometimes it's both.” – Henry [19:01]
- AI-driven tools allow designers to own more of the process, from ideation to interactive storytelling.
- The spectrum of deliverables increases—designers can move between mockups, interactive prototypes, and even lightweight product development without engineering dependencies.
-
Volume & Velocity Principle:
- At Perplexity, brand teams use any combination of tools to maximize output and presence.
- "We want to be able to make a lot of stuff and put it everywhere it needs to be that Perplexity is a brand that is just totally ubiquitous." – Henry [21:24]
-
Blurring Lines:
- Designers writing production code is normal; some focus on communication/visualization, others push code to production.
- “Anyone that wants to learn [code] is embraced.” – Henry [34:36]
4. The New Toolkit – Stitching Tools Together
All Speakers
[25:47 – 28:44]
- Chaining Tools is the New Artform:
- Stitching together specialized tools—Cursor, V0, Lovable, Figma, Midjourney, Kreia, Suno, etc.—for distinct phases of work.
- “There's an art to knowing what works well for what. That's just late nights and exploration. Eventually … this stuff will get streamlined.” – Henry [27:20]
- The end deliverable often still passes through traditional tools like Figma, but many steps before are handled in AI-powered environments.
5. When to Use Which Tools
Speaker: Pranathi Perry
[28:44 – 33:36]
- Tool Decision Matrix:
- Pre-dev: V0 and Figma for ideation, "rarely using Figma as source of truth"—live code is becoming the final reference.
- Dev: “Huge Claude/code user, implementing most designs myself.”
- Post-dev: Tools like Versatile Agent leave PR comments and help polish.
- Lean into AI especially when dealing with edge cases or complex interaction flows—much easier to prompt the tool conversationally than to mock up every case manually.
"Once you're getting to the point where you're building the whole product in a prototyping tool, you might as well just implement it." – Pranathi [32:07]
6. Technical Depth and Collaboration
All Speakers
[34:36 – 38:18]
-
Owning Front-End:
- A spectrum exists from designers who fully implement in code, to those who use AI tools for sandboxed prototypes.
- "Anyone who wants to learn is embraced … even those who never coded before are writing Smith UI now, with help from Cursor." – Henry [34:36]
-
AI is Not Just Speed, It’s New Deliverables:
- “AI doesn't necessarily make you faster at doing the old thing you were doing in Figma. It opens the door to new types of deliverables.” – Ridd [38:18]
7. Career Signals and The Edge for Designers
Speaker: Henry Modiset
[46:09 – 47:54]
-
What Stands Out in Designers Today:
- Product intuition—knowing "why would someone use this" is more critical than ever.
- Basic graphic design craft never goes away—“AI will give you choices, a designer needs to make a choice.” – Henry [47:08]
- Bias toward action and experimentation: Most successful hires have been solo freelancers or founding designers, able to handle ambiguity and make decisive moves.
-
Evergreen Advice:
“The last thing you can do is sit still when everything’s changing so quickly.” – Ridd [47:54]
Notable Quotes & Moments
-
"We're actually giving these tools to clients when we're done. The tool is theirs now … It's not just how fast we work—it’s about handoff, which is tough for companies moving fast."
— Nick Pattison [08:23] -
“It’s not quite Figma, it’s not quite dev, it’s like this middle ground that gives a better idea of what you actually want to build.”
— Pranathi Perry [12:20] -
“It doesn't make any sense that we use a tool that makes images as a way to plan out how software should work.”
— Henry Modiset [41:36]
Timestamps for Important Segments
- [00:10] Nick begins case study: AI-powered brand toolmaking
- [02:20] Nick on pattern architect tool for Flow Glad
- [08:23] Handoff to clients via interactive AI tools
- [11:14] Pranathi on V0, bridging “Figma” and “dev”
- [13:03] Natural language-driven asset generators in V0
- [15:52] Debugging and prompt iteration process with AI
- [19:01] Henry on expanding deliverables and the new scope of design
- [21:24] Perplexity’s “velocity and volume” principle
- [25:47] Stitching together diverse design/AI tools
- [28:44] Pranathi’s process: pre-, during-, and post-dev phases
- [34:36] Who codes: Ownership and learning spectrum at Perplexity
- [38:18] AI and the new technical bar for designers
- [41:36] Blending of roles across design/PM/engineering
- [46:09] Henry on hiring signals and design values
Takeaways for Designers
- Generative AI is not just an accelerant—it is redefining the types of work designers can deliver and how they collaborate.
- Building or stitching together small AI-powered tools can be as core to branding as a logo or type system.
- AI-powered prototyping is less about saving time for mockups and more about unlocking higher-fidelity, interactive, and even code-level deliverables for everyone in a product org.
- Successful designers will be those who are comfortable with ambiguity, have strong product intuition, a bias for action, and who never stop experimenting and learning.
Closing Thought
“The last thing you can do is sit still when everything’s changing so quickly.” — Ridd [47:54]
End of Summary
