Unifying six tools into one workflow for a $1.1B collections operation
Prestige Financial Services
UX Researcher & Designer
Prestige Financial Services
UX Researcher & Designer
Prestige Financial Services was an auto-finance debt collection agency servicing over 43,000 accounts valued at more than $1 billion. Their collection agents worked across a fragmented ecosystem of six-plus legacy tools to manage calls, messaging, payments, documents, account details, and compliance requirements simultaneously. As the principal designer on $5 billion in serviced accounts, I was brought on to design an internal case management application that would replace this sprawl with a unified, intelligent workflow.
This case study covers the end-to-end UX process: contextual research in a live call center, synthesis and persona development, design and usability testing, and the organizational realities of advocating for user needs against competing business constraints.
Collection agents at Prestige worked accounts across multiple delinquency stages — from 0-29 days past due through accounts headed toward repossession. Their daily work involved a demanding mix of outbound calls, inbound responses, text messaging, email, and payment processing, all while navigating strict compliance requirements around disclosures, contact frequency, and opt-outs.
The problem was that no single tool supported this workflow. Agents routinely operated across six or more applications simultaneously: Phoenix (core account system), MuTrak (phone system), Invotra (messaging), TalkDesk (another phone system), Word (letter templates), and Power BI (reporting). Each held a piece of the picture; none provided the full view an agent needed to efficiently work an account.
Compounding the challenge, the business was facing a scalability crisis. The existing spread of siloed systems, manual spreadsheets, and inconsistent information couldn't support the planned growth from 100 to 500+ serviced accounts. Employee turnover made things worse — in 2022, Prestige hired 50 new agents and only 3 remained by year-end. The design opportunity was to build something that was both unified and learnable quickly, and supportive of agents still building domain knowledge.
Agents were skilled communicators buried under fragmented tools, redundant processes, and information noise. The system needed to get out of their way so they could focus on what they did best: resolving accounts through human conversation.
Given the complexity of this domain, I knew surveys or interviews alone wouldn't capture the full picture — I needed to see how agents actually worked, not just how they described their work. I prioritized contextual inquiry and job shadowing as the foundation of the research plan, observing agents in live work environments across the outbound (ICT) and inbound (INT) teams. I also conducted swim-lane/journey mapping with agents, team leads, and training staff.
The research surfaced insights that self-reporting never would have. Agents routinely had multiple applications open simultaneously, switching contexts almost constantly. The notes feature in Phoenix, universally described as the most important part of the tool, was also the most frustrating: system-generated notes were mingled with actual conversation notes, forcing agents to spend time hunting for relevant information. Compliance tasks required agents to manually sync across systems, and the call management system created entries every 30-45 minutes.
Throughout the project, there was recurring tension between what research indicated would best serve agents and what the business was willing to implement. The research strongly supported a user-driven approach — customizable views, flexible queue management, agent-controlled templates. While leadership preferred a more controlled, streamlined experience.
In retrospect, I would have involved business stakeholders in research sessions from earlier, having leadership observe agents struggling with workarounds firsthand — rather than presenting findings in a slide deck after the fact. Building shared empathy for user needs at the organizational level is as important as the design work itself.
The project delivered a research-informed design that addressed the core pain points identified through contextual inquiry, synthesis, and usability testing. Prestige ultimately closed its operations before the full vision could ship, but the work surfaced principles that continue to shape my practice: that research in context reveals what interviews cannot, that advocating for users requires organizational strategy alongside evidence, and that the most meaningful design decisions in enterprise UX aren't about visual design — they're about rethinking how information and actions are organized around the user's natural rhythm of work.
