Cars311: desking tool for US car dealers
Dealers were building deals in spreadsheets and praying the taxes were right. I designed a desking platform that handles Lease, Finance, One Pay Lease and Cash for all 50 states, on one screen.
TL;DR
- Problem
- Building one deal took ~15 minutes across spreadsheets and calculators, and tax mistakes cost dealers money and trust.
- Solution
- A desking tool that compares deal types side by side, applies state tax rules automatically and shows dealer profit in real time.
- Impact
- Deal build time down to under 3 minutes · tax logic for 50 states in one flow · predictive profit hints for managers.
Context & problem
Business goal
Win US dealerships away from clunky legacy tools by making deal structuring faster, safer and more profitable. The target: replace manual spreadsheets and outdated software with a calculator that is accurate to the cent, from trade-in equity to the final margin.
User pain
Sales managers juggle rebates, money factors, residuals, fees and state taxes while the customer is sitting right there. One wrong field and the deal falls apart at the F&I desk.
Why it’s hard
One deal pulls together scattered data: credit scoring, dealer incentives, rebate programs and the exact vehicle condition (New, CPO, mileage). On top of that, taxes and fees depend on the dealer’s and the customer’s addresses, in any of 50 states.
My role
Lead product designer from discovery to launch. I owned the market and domain analysis, information architecture, the calculation UX, predictive profit features, the design system and UX QA, working daily with the product owner and engineers.
Discovery & research
Domain deep-dive
Sessions with stakeholders and a dealer-side domain expert to untangle how desking really works across four transaction types: Lease, Finance, One Pay Lease and Cash. Key insight: managers don’t trust a number they can’t trace back.
Competitor audit
DealerSocket, VinSolutions, Dealertrack: powerful, but everything is buried behind 10+ tabs.
Tax logic mapping
Mapped how taxes, fees and rebates behave per state and per deal type into one decision table the devs could build from.
Design strategy
User journey
Mapped the end-to-end deal flow: vehicle → customer → structure → compare → print or send to F&I.
A modular Worksheet
To handle the sheer number of variables, I designed a modular Worksheet that puts flexibility first without losing clarity. Its core is a deal structure dashboard where managers tune program factors in real time: residual, rates, mileage adjustments and capitalized costs.
Hi-fi explorations
Went straight to hi-fi in the design system with real deal data. Three layout directions; the tabbed one lost because managers kept losing context mid-deal.
Key decisions
1. Deal types side by side, not in tabs
Problem: customers ask “what if I lease instead?” mid-conversation. Options: tabs, a toggle, a column grid. Choice: the grid. Why: comparing is the whole job, and tabs hid it. Switching a deal from Lease to Finance keeps the relevant data and re-prices the profit instantly.
2. Every number is traceable
Click any payment to see the math behind it: taxes, fees, rebates. Trust went up, and “let me double-check in Excel” went away.
3. Profit visible only to the manager
A role-based view, so the screen can face the customer without exposing dealer margin, plus predictive hints on where the deal can earn more.
4. Upfront and capitalized costs, clearly split
A strict visual line between what the customer pays upfront and what gets capitalized into the deal gives F&I managers full control over the structure.
5. A trade-in that knows owned from leased
The Trade-In module adapts its inputs to the vehicle’s status, owned or leased, and recalculates payoff and equity automatically.
Final design
Design system & UI
A scalable component library with W3C DTCG tokens synced to code. Dense, number-heavy tables got their own type scale and states.
The layout is dense but organized, the way financial terminals are: key profit indicators and the final price stay in view while a manager digs into mileage penalties or regional fees.
How AI helped
Used Claude to turn state tax rules into structured test cases and to generate edge-case deal data for prototypes. Prototypes went to the team as code, not just Figma links.
Outcome & impact
- Deal build time: ~15 min → under 3 min
- 50 states of tax logic in one flow, no spreadsheets
- Calculation errors designed out: lender program terms live right in the UI, so dealers quote accurate, optimized payments on the spot
- Predictive profit hints that help managers protect margin
- A foundation for the whole car-sales ecosystem: usable design with the financial accuracy the US market requires
What I learned
- In FinTech UX, trust comes from showing the math, not hiding it.
- A domain expert in weekly reviews saves months of rework.
- Dense doesn’t mean cluttered: hierarchy matters most when every number counts.
Next case
