Skip to main content
How I AI

The internal AI tool that’s transforming how Stripe designs products | Owen Williams

54 min episode · 2 min read
·
Owen Williams

Episode

54 min

Read time

2 min

Topics

Productivity, Relationships, Startups

AI-Generated Summary

Key Takeaways

  • Design system MCP integration: Connect your internal design system to an MCP server and bundle cursor rules that instruct the LLM to query that server before writing any code. This prevents the model from hallucinating generic Tailwind components and instead generates on-brand UI at roughly 90% fidelity from a single prompt, eliminating the "indigo blurple slop" problem common in tools like v0.
  • Dev box deployment over local setup: Hosting prototypes on internal dev boxes rather than local machines solves two problems: designers often lack high-RAM engineering laptops, and shareable URLs replace screen-sharing in design reviews. At Stripe, spinning up a ProtoDash dev box takes under two minutes via a single internal URL, making prototypes immediately clickable by everyone in the room.
  • Data state prototyping in code vs. Figma: Prototyping data-heavy dashboards in code unlocks states that Figma cannot practically replicate — zero data, high volume, internationalized text, different business models (startup vs. enterprise), and error states. What previously required duplicating dozens of Figma frames can now be generated by prompting a single variant request directly in the browser-based tool.
  • Embedded annotation-to-fix pipeline: ProtoDash Studio includes a canvas annotation mode where reviewers click UI elements, leave comments directly on the prototype, and queue all feedback for the LLM to action in one batch. This replaces post-review Google Doc summaries and allows designers to send stakeholders a "fixed" confirmation with receipts immediately after the meeting ends.
  • PM self-service prototyping changes designer relationships: When PMs can generate on-brand prototypes by pasting a PRD link into the chat, they unblock early-stage exploration without waiting for designer availability. Rather than displacing designers, this shifts conversations from staffing arguments to craft elevation — PMs arrive at design reviews with a working baseline, and designers focus on refining quality rather than explaining concepts from scratch.

What It Covers

Owen Williams, design manager at Stripe, built ProtoDash — an internal AI prototyping tool that connects Stripe's design system (Sail) via MCP server to generate realistic, on-brand dashboards in a browser. The tool has shifted design reviews from static Figma JPEGs to clickable, data-rich prototypes used by both designers and PMs.

Key Questions Answered

  • Design system MCP integration: Connect your internal design system to an MCP server and bundle cursor rules that instruct the LLM to query that server before writing any code. This prevents the model from hallucinating generic Tailwind components and instead generates on-brand UI at roughly 90% fidelity from a single prompt, eliminating the "indigo blurple slop" problem common in tools like v0.
  • Dev box deployment over local setup: Hosting prototypes on internal dev boxes rather than local machines solves two problems: designers often lack high-RAM engineering laptops, and shareable URLs replace screen-sharing in design reviews. At Stripe, spinning up a ProtoDash dev box takes under two minutes via a single internal URL, making prototypes immediately clickable by everyone in the room.
  • Data state prototyping in code vs. Figma: Prototyping data-heavy dashboards in code unlocks states that Figma cannot practically replicate — zero data, high volume, internationalized text, different business models (startup vs. enterprise), and error states. What previously required duplicating dozens of Figma frames can now be generated by prompting a single variant request directly in the browser-based tool.
  • Embedded annotation-to-fix pipeline: ProtoDash Studio includes a canvas annotation mode where reviewers click UI elements, leave comments directly on the prototype, and queue all feedback for the LLM to action in one batch. This replaces post-review Google Doc summaries and allows designers to send stakeholders a "fixed" confirmation with receipts immediately after the meeting ends.
  • PM self-service prototyping changes designer relationships: When PMs can generate on-brand prototypes by pasting a PRD link into the chat, they unblock early-stage exploration without waiting for designer availability. Rather than displacing designers, this shifts conversations from staffing arguments to craft elevation — PMs arrive at design reviews with a working baseline, and designers focus on refining quality rather than explaining concepts from scratch.

Notable Moment

During a live build, Williams prompted ProtoDash Studio to create a Black Friday dashboard for a pet store with a sales ticker and trending products. The tool autonomously queried the Sail design system, built the page, detected its own rendering error via screenshot, and self-corrected — all without any manual intervention.

Know someone who'd find this useful?

Episode Transcript

My dream was, I want something that's like v zero but for us. We have all of these tools internally that are really cool. We can connect different data sources together. Why can I not just do this in my browser? Like, why do I need Cursor? You're seeing a lot of designers use it, but maybe even more PMs. I started seeing PMs use it and got a little nervous. Oh my goodness. PMs designing. It's like, what's gonna happen? Is it how painful is it to prototype a data dashboard with all its interactions, all its filters, all its states, different states, zero data, a bunch of data. It is nearly impossible to do that in Figma. It's sort of been this very transformative thing because all of a sudden, I'm sitting in these design reviews, and it's so convincing that I'm like, is this the real product, or am I looking at something fake? Welcome back to How I AI. I'm Claire Vow, product leader and AI obsessive, here on a mission to help you build better with these new tools. Today, I have Owen Williams, design manager at Stripe, and he's gonna show us how he vibe coded his own vibe coding platform for their internal design prototypes. It is one of the most impressive internal tools I have seen and such an awesome way to rethink how your product builds products with tools. Let's get to it. This episode is brought to you by Siligo. Every company today wants AI to improve how work gets done. The fastest way is building it directly into everyday business processes, automating employee onboarding, keeping customer data accurate, managing orders and inventory, or resolving finance and operations issues. When AI lives inside the flow of work, it can update records, trigger approvals, route work, and kick off the next step across systems. That's how teams operationalize AI and deliver measurable results. Soligo makes this possible. And now, with Soligo Aura, it's never been easier. Soligo Aura gives you access to the entire platform through natural language, connecting your systems and turning intent into action. All of it under your control. Companies like Databricks, PayPal, and Olipop rely on Soligo to run critical business operations at scale. Ready to operationalize AI? Visit celigo.com/howiai. That's celigo.com/howiai. Owen, thanks for joining How I ai. Hey. I'm happy to be here. I know that you're stepping away for a few minutes from parental leave with your second. So I appreciate you giving us the time. And what I love about what you're gonna show us is how to we were joking before we started recording. How to get prototypes that don't look like generic tailwind indigo slop. So tell me about tell me what the problem you're facing and kind of how you came to the solution that you came to. You know, I'm a design manager, so I'm I'm sitting in a lot of design reviews. And all of the designers at …

Get the full transcript (10,926 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all How I AI transcripts →

You just read a 3-minute summary of a 51-minute episode.

Get How I AI summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links.

Tools

  • Celigo is listed as a sponsor with url https://celigo.com/howiai
  • Cursor is listed as a sponsor with url https://chatprd.ai/howiai
  • ProtoDashBy guest

    by Stripe

    Owen Williams, design manager at Stripe, built ProtoDash — an internal AI prototyping tool that connects Stripe's design system (Sail) via MCP server to generate realistic, on-brand dashboards in a browser.
  • This prevents the model from hallucinating generic Tailwind components and instead generates on-brand UI at roughly 90% fidelity from a single prompt, eliminating the 'indigo blurple slop' problem common in tools like v0.
  • This prevents the model from hallucinating generic Tailwind components and instead generates on-brand UI at roughly 90% fidelity from a single prompt, eliminating the 'indigo blurple slop' problem common in tools like v0.
  • SailBy guest

    by Stripe

    Owen Williams, design manager at Stripe, built ProtoDash — an internal AI prototyping tool that connects Stripe's design system (Sail) via MCP server to generate realistic, on-brand dashboards in a browser.

More from How I AI

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best AI Podcasts (2026) — ranked and reviewed with AI summaries.

Read this week's Startups & Product Podcast Insights — cross-podcast analysis updated weekly.

You're clearly into How I AI.

Every Monday, we deliver AI summaries of the latest episodes from How I AI and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime