Challenge

A worn-down team, a broken process, and a 12-week forcing function

Capital One Auto Navigator redesign — before

The experience was stale, with known pain points. Concepts looked great in reviews but never shipped, until a national campaign made the status quo untenable.

I inherited a design team that was worn down. Their bold concepts solved real usability problems, but every quarter product and engineering gave the same answer: too hard to build, too risky to change, other priorities. Meanwhile, the UI was 7 years old and customers kept hitting the same frustrations.

Talking to the team, I found a set way of working that wasn't working: long review cycles, a waterfall handoff to engineering, and late legal reviews that sent designs back to square one.

Then the business handed us a forcing function. Auto Navigator was launching its first national ad campaign in 12 weeks. We couldn't send that traffic to an experience we knew was broken. My manager called: "We need to launch the new experience before the campaign goes live." We knew what to do, but not how.

Solution

A 10-week plan from design to deployment

Capital One Auto Navigator redesign — after

Higher quality, faster, with product, engineering, and legal in the room together.

I built a 10-week plan to get from design to deployment, and set the tone early:

"We need to get to higher quality, faster."

In the same breath, I laid out how: a real agile cycle, with two-week cross-functional reviews and engineers picking up designs before they were 100% locked. I coached designers to learn the tech stack so our decisions only touched the front end. Every review checked that, and anything needing backend work got caught early.

Engineers were the hardest to win over. They wanted locked designs before they'd build. I convinced them to work from designs that were 80% done, with every major decision final and small details still moving, so we didn't lose weeks to fine-tuning before dev could start.

It's important to me that designers are good partners to engineers. They started trusting the process once they saw us deliver on our promise.

We grounded every decision in research we already had, and broke down the silos between designers. I called it "playing jazz": when someone got stuck, another designer jumped into their Figma file and iterated with them, live on Zoom.

Results

Shipped on time — without touching the backend

+4.5%increase in funded loans
+16%searches
+13%completed applications

Every metric that mattered went up, and the process stuck as a better way of working.

We shipped on time, without touching the backend.

Most importantly, a tense project that could've easily become a "fire drill" was instead treated as a new, strategic and effective way of working smarter, faster and more deliberately alongside cross-functional partners.

Saved cars per customer also tripled. With only front-end changes behind results that strong, this project has always stood out to me as a clear example of why design matters. We sacrificed very little of the design. The only real sacrifice was an old way of doing things.