Disclaimer: Shared with permission from America's Car-Mart. Screens are recreated with sample data.

If you only have a minute

• The product: America's Car-Mart sells used cars and finances the loans itself, so collections is central to the business. After 15+ years on ALIS, a legacy system, the company moved collections to Salesforce, partnering with Salesforce's implementation team.

• My role: UX Researcher, the company's first. I co-led discovery with Salesforce's UX designer. When she left half a sprint into the build, I became the only UX person on the project.

• The challenge: 150+ dealerships each ran collections their own way, leaders couldn't see what was working, customers weren't responding, and associates' time went into paper and workarounds. I joined two weeks before the project began, in an industry that was new to me.

• What I did: traced each business problem to its root cause on the dealership floor, turned discovery into 65 requirements across 7 roles, caught steps the user stories had missed, ran two rounds of usability testing, restructured the screens reps use most, and studied the platform after launch.

• Results: within three months of launch, delinquency fell by ~14% and ~35% more customers responded. Personal visits fell by ~75%.

• Status: launched at about two dozen pilot dealerships, then rolled out company-wide.

Why collections needed a new foundation

Because Car-Mart carries its customers' loans, collections isn't a back-office function. When customers fall behind, the business feels it directly. For 15+ years the work ran on ALIS, a legacy system the business had outgrown, and the strain showed at every level.

So the company set out on a multimillion-dollar move to Salesforce, to build a foundation that scales with the business.

Plus: associates' time back on customers, and visibility at every level.

How the work unfolded

Seven phases over about ten months, from the kick-off workshop to post-launch studies. Each phase answered one question. This page tells the Discover story in full; the rest is summarised at the end.

Who we heard from

In every round I sampled both ends of three dimensions, so we'd hear where needs differed, not only where they agreed.

Step 1 · Understanding the business before meeting its users

I joined two weeks before the project began, in an industry that was new to me. Before I could ask associates the right questions, I needed to understand how the business works: how Car-Mart makes money, what leadership expected from Salesforce, and how success would be measured.

I joined a three-day workshop with business stakeholders, product and tech leaders and Salesforce's team, to confirm the vision, goals and measures together. With Salesforce's UX designer, I then drafted research plans for the stakeholders and for each user role.

Step 2 · Following the work to the dealership floor

Leadership could describe the problems, but not why they happened. The answer was at the dealerships, so that's where we went. With Salesforce's UX designer, I co-led 25 interviews across 12 dealerships, from Account Reps to the VP of Collections. Most were in person, and we shadowed associates as they worked with customers.

We didn't interview customers directly. We learned about them from the associates who speak to them every day: their questions, hesitations and fears. We also watched those conversations happen.

One pattern showed up early: frustrations depended on each associate's tenure, background and comfort with technology. Telling training needs apart from real system problems became a thread through the whole project.

Methods: interviews · shadowing · affinity mapping · personas · journey mapping · SWOT

Step 3 · Behind every business problem, a root cause on the dealership floor

I took each of the business's challenges to the people living with them, and traced why it happened, how it showed up, and how people coped. Each cause is tagged with the challenge it explains.

Dealership leadership: the numbers couldn't say why. Leaders could see results, but not the work behind them.

Reps and the Resolution team: the system held data, not decisions. Reps carried the history, the rules and the proof in their heads, in notes and in their own spreadsheets.

Customers: it wasn't only the channel. Every way to fix a late payment cost the customer time they didn't have.

What research changed: the project set out to replace personal visits with texting. Discovery showed a significant group of customers, often long-standing ones, don't text or pay online. So the goal became reducing personal visits, not eliminating them.

Step 4 · Seeing the whole workflow, and the desk behind it

No single role could describe collections end to end. I mapped how an account moved, from the first missed payment onward, so the team could see exactly where time, trust and customers were being lost.

Redrawn from my journey maps and persona workflows, without names. Quotes paraphrased.

A normal day at a rep's desk. I drew this scene to keep the team focused on the experience we had to fix. Everything on this desk was information the system should have held, and reps were carrying it while they served the customer in front of them.

Redrawn from shadowing at the dealerships.

What Discover set out to fix

Discovery ended with one goal for each group, and every requirement that followed had to serve one of them.

From discovery to launch, in short

• Define: I turned discovery into 65 requirements across 7 roles. In parallel I researched texting: 31 requirements, 33 KPIs, and a workshop where stakeholders agreed the rules for texting customers.

• Build: when Salesforce's designer left half a sprint into the build, I found steps the user stories had missed, such as digital escalations. I then walked both teams through the discovery workflows.

• Test and iterate: hands-on access raised alarm, so I ran a usability study (SUS 71, 33 issues) and separated product fixes from training needs. I then restructured the screens reps uste mos, using card sorting and tree testing.

• Validate: a second usability round led to dashboards that drill down from dealership to rep to customer.

• Launch and learn: after the pilot, a gap analysis and an impact study showed what still got in the way. Every issue raised was fixed before the company-wide rollout.

The full walkthrough is available in interviews.

What I learned

• Show, don't tell. UX was new to the company, and a short demo of how a usability study works got a build paused to make room for one.

• Raise gaps as questions, not corrections. That's how they got escalated instead of defended.

• A platform change is a change in people's working lives first. Long-tenured reps needed support, not just training.

• Next time, I'd set baselines before launch, so every outcome could be measured against a starting point.