Problem Scoping
A process that worked - until it didn't
"The software used is dated, has a steep learning curve, and makes it difficult to train new technicians." — Consistent feedback across all staff interviewed
Current - Home Screen
Grayed-out menu items inaccessible to technicians create visual clutter and confusion.
Current - Checkout Screen
On-screen keyboard on a non-touch terminal. Patient info lives on a separate screen entirely.
1: Wasted Interface Real Estate
Pharmacist-only features are visible but inaccessible to technicians — dead buttons that add confusion, not functionality.
2: Checkout ↔ Profile Navigation Has Dead Ends
Updating patient info mid-checkout requires exiting the process entirely and restarting — adding minutes to every complicated pickup.
3: No Cost Transparency for Patients
Insurance details and price breakdowns don't appear in the patient-facing communications, generating frequent "how much will this cost?" or "does my insurance cover this?" calls.
Research
The scope expanded the moment I started watching
What I thought was a UI problem turned out to be a service design problem. The software was frustrating, yes — but the underlying issue was that the entire pickup journey had gaps: gaps in communication, in transparency, and in the handoff between digital and physical.
I used three methods to understand both sides of the experience: contextual inquiry with technicians and pharmacists over several weeks, a survey of pharmacy customers, and journey mapping to trace where friction accumulated for both groups.
Wait time was the #1 customer complaint
Survey data confirmed what observations suggested: customers weren't upset about prices or availability — they were upset about standing in line.
Technicians search across multiple systems
Patient info was scattered across screens and tools, forcing mental context-switching during every checkout — especially when something went wrong. This is due to the checkout system and pharmacy software being two completely separate softwares.
Notification gaps created confusion
Patients received "there's an issue with your prescription" texts with no explanation of what the issue was or what to do next.
HIPAA shaped every customer-facing idea
Any kiosk or queue display needed to be completely anonymous — names couldn't appear on shared screens.
Users
Two very different people, one broken system
I created personas based on direct observation and survey data — not assumptions. These two users have conflicting needs: Virginia needs fewer interruptions during checkout; Greg needs more visibility into what's happening before he ever reaches the counter.
Pharmacy Technician
Virginia wants to provide fast, friendly service — but the software fights her at every turn. She's memorized workarounds that newer technicians haven't learned yet, creating uneven service quality across shifts. Her biggest blocker: patient info is never where she needs it when she needs it.
Pharmacy Customer
Greg
Greg picks up prescriptions on his lunch break. He doesn't know how long the line will be, he doesn't know why his prescription has a "problem," and he doesn't know what insurance was billed. He feels like the system wasn't designed with him in mind — because it wasn't.
Design
From a clunky checkout to a unified workspace
Before building, I did a competitive analysis of Walmart.com, Amazon, and Shopify's checkout flows to find patterns worth borrowing. All three share a consistent structure: cost breakdown on the right, item details on the left. The Connexus redesign follows the same logic — familiar mental models reduce training time.
Before - Fragmented
Patient profile and checkout are separate screens with no shared navigation.
After - Unified Workspace
Patient info, prescriptions, and checkout live in one view. No exiting, no restarting.
Vertical Sidebar Navigation
Keeps consistency with the existing UI while making all key features accessible in one tap — no hunting through menus.
"Next due date" auto-calculation
Technicians currently estimate refill dates manually. Surfacing this automatically reduces cognitive load and speeds up one of the most common customer questions.
Bulk insurance update
One click to apply a new insurance plan to all prescriptions simultaneously — vs. the current process of updating each one individually.
Walmart-branded checkout layout
Cost breakdown, payment selection, and prescription list follow e-commerce best practices — immediately familiar, minimal training needed.
Service Design
Fixing the screen wasn't enough
The software redesign addressed the technician side — but customers were still standing in winding lines with no visibility into wait times or prescription status. I explored three in-store additions: a waiting area with seating, a queue display, and a self-service kiosk.
Sketch exploring store layout with dedicated pickup zone and kiosk placement
Kiosk flow: prescription status check → expedite → numbered ticket (HIPAA-compliant queue)
The kiosk proposal addressed a specific gap: customers without the Walmart app had no self-service option at all. The kiosk gives them prescription status checks, vaccine scheduling, and a queue ticket system — all without displaying names on shared screens, keeping it HIPAA-compliant.
Testing
Testing with the people who actually use it
I tested the Connexus redesign with pharmacy technicians using a task scenario: check out a prescription for a patient who needs to update their insurance mid-checkout. For the kiosk, I ran concept tests with both staff and customers.
Single-window workflow was immediately preferred
All technicians appreciated having patient info and checkout in one view — no toggling, no restarting.
Next due date" was a big time-saver
Technicians said calculating refill dates manually is one of the most common time sinks. Automating this simple need was immediately valued.
Bulk insurance update was "very appreciated"
The ability to apply insurance changes to all prescriptions at once — vs. one by one — directly addressed their most common checkout interruption.
Auto-refill toggle caused confusion
Users mistook the auto-refill toggle on the patient profile for the prescription selector on the checkout screen. Next step: relocate or redesign for clarity.
Kiosk use is context-dependent
Customers said they'd use the kiosk when it's busy, but prefer talking to staff when it's quiet. Multi-language support was flagged as a priority addition.
Reflection
What this project taught me
I was wrong about the problem — and that was the most important discovery
I came in convinced this was a UI problem. The software looked old, technicians complained about it constantly, and I figured a cleaner interface would fix everything. What I actually found was a service design problem: the software issues were symptoms of deeper gaps in the pickup journey itself. That pivot happened because I kept watching and asking instead of jumping straight to solutions.
Observation reveals what interviews can't
Technicians couldn't always articulate why checkout was slow — they'd adapted to the workarounds so completely that the friction felt normal. Watching the process in real time showed me things nobody would have thought to mention in an interview.
The best design feedback isn't always about the design
The most meaningful moment from testing wasn't a usability win — it was a technician saying the redesign made her feel more capable at her job. That shifted how I think about what "success" means in this kind of work.
Next Steps
With more time, I'd expand the prototype to a fully interactive flow covering edge cases (late prescriptions, insurance rejections, multi-patient pickups), translate the kiosk features to a Walmart app prototype, and broaden testing to include non-English speakers and customers with disabilities — two groups the current system underserves most.
