Payments

Discovery

After the core components of CustomerConnect were built, my team began a thorough investigation into why members were calling, which call types were taking longest, and how to reduce cost for the enterprise.

Analysis of Caller Intent data revealed that over 30% of all calls were for billing and payments-related issues — the single largest call driver. Despite this, advocates still had to jump between multiple legacy systems to answer even basic payment questions.

Research

My team organized interviews with phone advocates to identify pain points and document the gaps between what CustomerConnect offered and what the legacy tools provided.

After compiling notes, I conducted side-by-side shadowing sessions with advocates on live calls. This is the most valuable research method available — it lets you observe exactly how quickly advocates move through the application and, more importantly, it emotionally attaches you to the work because you're watching real people who need real assistance.

I also did demographic analysis of the advocate user pool and traveled to Florida Blue callcenters in Miami, Orlando, San Antonio, and Raleigh to test my hypothesis about legacy tool preference firsthand.

Pain points identified:

• Data gaps — missing information not carried over from legacy systems
• Invoice breakdown confusing — labels inconsistent and unclear
• Missing adjustment reasons
• Delinquent status shown in the wrong color
• Takes 3 days for a posted payment to reflect

Testing & Prototyping

I compiled all research into documents for each interview and shadowing session, including screenshots of the legacy systems. The goal was to organize information in a way that felt familiar, but with a modern approach to navigating complex billing data at scale.

I created visual artifacts in Adobe XD and exported them to inVision to build clickable prototypes. Prototyping with non-coded assets was new for the company — it saved months of development time and became the universal method for gathering early feedback without needing engineering resources.

We interviewed advocates with both the old tools and the new prototype, iterated through several rounds, and landed on a vision now in production. This exercise drove enterprise-level change: better design tools, greater access to users for research, and hundreds of thousands of dollars saved by targeting exactly the right enhancements.