Staff Product Designer at Intuit
Case Studies
A multi-entity financial platform for franchise brands running dozens of locations.
Franchise brands had no unified way to manage their finances. Every franchisee operated in isolated books with an inconsistent chart of accounts, while franchisors stitched together spreadsheets to get even a partial view of network health.
Month-end close dragged. Benchmarking one location against another was effectively impossible. And Intuit had no product built for a fast-growing mid-market segment that was asking for one.
AI that doesn't answer questions — it does the account work a team used to do by hand.
Lighthouse is a select group of enterprise customers on Intuit Enterprise Suite, our new platform. They show us the way — co-creating solutions with us, holding direct relationships with Intuit's C-suite, and shaping a segment that's central to the company's growth.
Since the program launched, those relationships had been run out of five or more Google Sheets. One tracked monthly check-ins. Another held basic customer profiles. A third tracked executive engagements and GTM-ready status. The most strategically important accounts at the company were being managed by hand, in tabs, with no single view of who a customer was or how they were actually doing.
As project lead and product designer, I defined the customer 360 that would replace those manual processes — and designed the AI to do the work, not describe it. Every automation replaced a task a person was previously doing by hand.
An agent that does real financial work — and earns the trust to touch money.
Our customers paid their bills through third-party tools — Ramp, BILL, and others. The company's bet was to become the all-in-one platform: manage bill pay end to end without leaving our system.
The gap was AI. Competitors already offered automated bill entry and approval workflows; we didn't. And the deeper problem underneath that gap was trust. Paying a bill is irreversible. A finance team will not hand money to a system that can't show its work, no matter how fast it is. Automation that can't be supervised doesn't get adopted — it gets turned off.
I led design on the Bill Pay agent for mid-market customers. Rather than retrofitting a chatbot onto the existing AP flow, I designed the full loop of delegating consequential work to an agent — and staying in control of it.
