Platform design at scale
I scaled Intuit's AI+HI expert platform across QuickBooks and built the tooling that made delivery 10× faster.
AI Native IEP — Intuit's AI+HI expert platform — started as a tool for one business unit: single-session, high-volume, resolution in minutes. The next business unit up was the opposite. Multi-session, relationship-driven, customer lifecycles measured in months. Porting the platform wasn't a reskin. I had to separate the durable patterns — context, handoff, expert assist — from the assumptions baked in by the original single-session use case, then redesign the rest.
Challenge.
Three new services were on the roadmap — Expert-Led Onboarding, Bookkeeping Review, and Quarterly Tax Estimates — each with its own engagement model, its own data shape, and a customer relationship measured in months instead of minutes. None of them mapped cleanly to the patterns the original platform was built around.

What used to take days of manual Figma configuration now takes a conversation.
Approach.
I led design on three new services that required new platform patterns: a cross-engagement context model so experts see a customer's full journey across sessions, a parent-child engagement architecture for embedded follow-up services, and cross-brand handoff patterns that preserve context as customers move between QuickBooks and TurboTax.
In parallel, I built IEP Studio — an AI-powered design and prototyping tool where describing a service in plain language generates a pixel-accurate, design-system-compliant IEP screen. What used to take days of manual Figma configuration now takes a conversation. Any designer, PM, or stakeholder can iterate on a live workflow preview without touching Figma or writing a line of code. AI composes inside the design system; the team signs off on the service.




Outcome.
Three new services — Expert-Led Onboarding, Bookkeeping Review, Quarterly Tax Estimates — are in active design and build, all running on a shared cross-engagement context model that's now the foundation for every future service on the platform. IEP Studio collapsed the work of populating a single IEP instance from 45 minutes of manual Figma configuration to a one-second paste — and it's live at a public URL with no access reviews, no admin rights, no setup. The first time I scaled this platform to a new business unit took months of figuring out which behaviors generalized and which were specific to where it grew up. The next designer who does it starts with the playbook already built.
Design tokens.

What changed.
3
New services designed and in development.
2,700×
Faster to populate an IEP instance — 45 minutes of manual Figma configuration collapsed to a 1-second paste.
92 sec
Blank slate to populated IEP instance — 4 hours of rough-draft writing and 45 minutes of Figma configuration collapse to a single FigJam URL paste.
0
Login, install, or build step to run IEP Studio. Live at a public URL — open to anyone, not just designers.

The work that compounded wasn't any one of the three new services. It was the playbook left behind — the patterns that decide where AI composes and where the team signs off on the service, the tooling, the public-URL prototype anyone can run without an access review. The next designer who scales this platform starts where I finished.
[Next case study]