Prototyping with AI: The End of Figma?
At the inaugural Curiosity session hosted by Experience Haus, we were excited to welcome Parth Loliyania to the stage. A Senior Product Designer at Publicis Sapient and a Service Design instructor at Experience Haus, Parth has always been passionate about how emerging tools shape the way we create. His talk, provocatively titled “Prototyping with AI: End of Figma?”, explored how a new generation of AI-powered tools is challenging long-held assumptions about design workflows—and perhaps even redefining the role of designers themselves.
The Search for the “Perfect” Prototyping Tool
Parth opened by reflecting on the long-standing ambition of the design community: to find a single tool that can fully encapsulate a product’s value, communicate its look and feel, and clearly convey specifications. Over the years, we’ve migrated from makeshift solutions like Microsoft Paint and PowerPoint, to graphic design tools such as Photoshop and Illustrator, and then into platforms tailored specifically for UX/UI work—Sketch, Adobe XD, and finally, Figma. Figma, in particular, brought collaboration and developer handoff to new levels, becoming the de facto standard for many design teams.
Yet, despite these advancements, something still feels incomplete. Designers are still chasing prototypes that don’t just look like the real product, but actually feel like it. This gap between the visual design and a truly functional experience is precisely where Parth sees AI stepping in—not just generative AI, but a more evolved form: agentic AI.
What Is Agentic AI—and Why Does It Matter?
Unlike generative AI tools such as ChatGPT or Midjourney, which respond to prompts by producing text, code, or static images, agentic AI has a purpose-driven architecture. These tools—like Vercel, Replit, Relume, and v0 (formerly VZ)—are designed to execute on objectives. Give them a prompt to create a quiz app, and they won’t just spit out some code; they’ll generate a working, hosted experience. It’s not just design; it’s delivery.
Parth demonstrated this difference with a compelling example. When he asked ChatGPT to help him build a quiz app, it returned code snippets and text explanations—useful, but incomplete. When he gave the same prompt to an agentic AI tool, it generated an entire working app, hosted online, complete with questions, interactions, a scoring system, and UI. It was, in his words, “shockingly usable”—and it was created in minutes, not hours.
Speed, Consistency, and the Designer’s Role
The conversation quickly turned to speed and efficiency—two of the most immediate advantages of agentic AI. Tasks that might take hours in Figma, such as building a login screen with working text fields and proper error states, can be completed by these AI tools in under two minutes. Not only are these outputs fast, but they’re also surprisingly consistent. That’s because the tools draw from existing UI libraries like Tailwind and Catalyst, creating predictable and standardized designs.
This consistency extends beyond visual style. Parth noted how these tools also respond intelligently to requests for interactivity and responsiveness. For example, when he asked v0 to change a top navigation bar into a bottom navigation for a mobile view, it made the adjustment without compromising the layout on tablet or desktop—demonstrating a level of context awareness and responsiveness that traditional tools struggle to match.
What’s more, these prototypes are live. Designers can send users to functioning websites instead of static mockups. This means real-time feedback, especially on things like error states or dynamic content, is now possible in user testing—something that Figma can’t replicate without extensive workarounds.
But It’s Not All Perfect
Despite the potential, Parth was quick to acknowledge that agentic AI tools are not ready to replace Figma across all scenarios. They excel in short, sharp engagements—concept pitches, MVPs, RFPs—where speed and functionality are paramount. But for longer-term projects with complex systems, design tokens, and stakeholder alignment, these tools are still finding their footing.
The output also isn’t flawless. Parth shared moments where the AI-generated designs required multiple rounds of prompting, and even then, didn’t always align with his original vision. But even here, there was a silver lining: the tools were getting better at correcting themselves. Mistakes weren’t the end of the road; they became part of the process.
Another limitation is the need for thoughtful guidance. Designers need to prompt with precision and intention. Parth likened it to being part-business analyst, part-designer—you must clearly define what you want the system to do, and be ready to adapt when it misses the mark.
A Glimpse Into a New Workflow
One of the most impressive demonstrations in Parth’s talk came when he used screenshots from a dashboard concept and fed them into v0. The AI tool was able to create a live prototype with interactive elements, real-time filters, and responsive layouts. Features that would’ve taken hours to wire up in Figma—like sliders that actually control chart outputs—were up and running within minutes.
That’s when the penny dropped: perhaps the future isn’t Figma or AI, but a workflow that ends with AI. Figma might still be where designers ideate, explore, and visualize. But when it comes to producing testable, interactive outputs, agentic AI might be the new finish line.
So What Happens to Designers?
This is where Parth got reflective—and perhaps even philosophical. If AI tools are accelerating delivery, what does that mean for the designer’s role?
He offered two possible paths. One is a return to craftsmanship, where designers focus on the emotional fidelity of experiences—think of UX artists building hyperrealistic 3D prototypes that feel indistinguishable from the final product. The other is a shift toward strategy. With execution automated, designers could become research-driven decision-makers, shaping products at the conceptual level rather than just building screens.
In both futures, Parth argued, design becomes more valuable—not less. We might not be the ones drawing every screen, but we’ll be the ones guiding what those screens mean, how they behave, and what problems they solve.
Closing Thoughts
By the end of the session, one thing was clear: AI isn’t here to replace designers. It’s here to augment us. The challenge now is to learn how to work with these tools—how to prompt, iterate, and think strategically about where we add value.
As Parth put it, the question isn’t who is a designer in this new world, but what is a designer. What skills matter most? What mindset do we need to embrace? And how do we stay curious in a field that’s evolving faster than ever?
If you missed the session, we highly recommend keeping an eye on future Curiosity talks at Experience Haus—because the future of design isn’t just coming. It’s already here.
