Optum Bank: HSA Reimbursement Platform Redesign
Redesigned a struggling reimbursement feature, improving task completion from 1.1% to 30% and reducing support costs by 30% for 450K users. Strategic transformation with strong business impact.
Before the redesign, users struggled with a complex reimbursement process that was manual, error-prone, and inflexible. As Lead UX Designer, I worked with cross-functional teams to design an automated, intelligent system that reduced abandonment by 75% and support calls by 52%.
The Challenge
The HSA platform serves 450K users managing their healthcare reimbursements, but the feature had a 98.9% abandonment rate with only a 1.1% completion rate—the lowest performing feature in the platform.
Key Problems:
- • Scattered forms across multiple screens
- • Generic error messages with no guidance
- • Inconsistent mobile/desktop experiences
User Impact:
- • 18-minute average completion time
- • 22% of support calls were reimbursement-related
- • "I gave up after it lost my receipt again"
Complexity & Constraints
Multi-User Challenge
Four distinct user types with conflicting needs—from digital natives to assisted users
Technical Constraints
Legacy systems, OCR processing, and mobile network reliability issues
Regulatory Requirements
IRS compliance, HIPAA restrictions, and mandatory audit trails
Cross-device reality: 67% started on mobile, 38% switched to desktop after failures—revealing device choice was context-dependent.
Strategic Business Context
Market Position & Competitive Landscape
UnitedHealth Group's Optum Bank was facing significant customer satisfaction challenges. With 450K active HSA users and the reimbursement feature being the lowest-performing part of the platform, it was becoming a risk for retention.
User Base Impact
450K users
At risk of churn
Support Cost
High
Disproportionate support load
Competitive Risk
Moderate
Fintech competitors improving
Why This Project Was Board-Level Priority
This wasn't just a UX improvement—it was a strategic business initiative with C-suite sponsorship driven by key OKRs and fiscal pressures.
Customer retention was a key OKR
Core metric for investor relations
Operational costs were 3x industry benchmark
$8.4M annually in support overhead
Regulatory compliance audit was imminent
8-month window to address systemic issues
Competitive differentiation opportunity
Commoditized HSA market needed innovation
Investment Case I Built
The business case required demonstrating clear ROI and risk mitigation to secure executive commitment and resources.
Financial Impact
- Team Investment:6-month design effort + engineering
- Annual Savings:~$800K in support costs
- Retention Value:Reduced customer churn risk
Success Metrics
- Completion Rate:1.1% → 30% (27x improvement)
- Support Reduction:30% fewer calls
- Mobile Gains:0.3% → 18% completion
My Role & Collaborative Approach
Lead UX Designer
I worked as a Lead UX Designer on this project, responsible for defining the end-to-end UX strategy, research approach, and design execution. While I didn't manage direct reports, I collaborated closely with peers across design, research, product, and engineering to drive the initiative forward.
Owned the design vision
Defined the strategic approach to solving the reimbursement problem
Led research discovery
Directed 7-week research phase, analyzing data and synthesizing insights
Made design decisions
Drove solutions through multiple iterations based on evidence
Core Collaborators
I worked closely with a lean, focused team of peers. Each brought critical expertise that was essential to the project's success.
1UX Researcher
Partner who helped design and conduct the research phase, analyzed customer service data, and synthesized user insights.
Critical for: Understanding true user needs despite privacy constraints
1Product Manager
Aligned on priorities, managed stakeholder expectations, and helped shape the business case for the project.
Critical for: Translating design into business value
1Engineering Lead
Partner who evaluated technical feasibility, owned implementation, and helped identify innovative solutions like OCR.
Critical for: Bridging design vision with technical reality
1Additional UI Designer
Collaborative partner who helped with detailed design execution and design system consistency.
Critical for: Shipping quality at scale
Navigating Cross-Functional Relationships
Beyond the core team, I worked with stakeholders who had different priorities and constraints. Success meant understanding their needs and finding solutions that worked across disciplines.
Engineering Team
Collaborated on technical feasibility, prioritization, and implementation of complex features like OCR and progressive workflows
Compliance & Legal
Worked to ensure design solutions met regulatory requirements while maintaining usability and user trust
Customer Support Team
Engaged with team handling support calls to understand pain points firsthand and validate design solutions
Product Leadership
Communicated findings, justified design decisions, and made the case for investment in the reimbursement feature
Design Challenges & How I Solved Them
The Compliance vs. Usability Tension
Early in design, compliance flagged required disclosures that would complicate the user flow.
The Challenge
I had designed a streamlined reimbursement flow, but compliance raised concerns about IRS Publication 969 requirements. Adding all required disclosures upfront would destroy the simplicity I'd designed for. The traditional approach would be a wall of legal text—exactly the friction we were trying to eliminate.
My Approach
- ✓ Reframed the problem: compliance isn't the enemy, it's a feature
- ✓ Designed "progressive disclosure"—showing information contextually when needed
- ✓ Worked with compliance team to understand what information mattered most
- ✓ Tested the approach to ensure users actually understood the disclosures
- ✓ Documented the pattern for reuse across other projects
Outcome
Found a solution that satisfied both needs. The progressive disclosure pattern became UHG's standard approach for balancing regulatory requirements with user experience across the HSA platform.
Discovering: One Flow Won't Work for Everyone
User research revealed 4 distinct user groups with conflicting needs. Our single "optimal" flow wouldn't work.
The Discovery
After analyzing 6 months of analytics and 800+ customer service transcripts, I identified 4 distinct user segments:
- Digital Natives (32%): Want mobile-first, instant experience, will abandon after first friction
- Traditional Planners (41%): Prefer desktop, detailed forms, documentation—they want to understand everything
- Assisted Users (18%): Need guidance, prefer guided workflows, may need support assistance
- Mobile Primary (9%): Mobile-only users with different constraints and needs
My Decision
Instead of forcing everyone into one "perfect" flow, I proposed building fundamentally different experiences for each segment. This meant more design work, but it was the only way to serve everyone well.
- ✓ Built separate, optimized flows for each user type
- ✓ Made the business case that quality would improve ROI
- ✓ Created segment-specific success metrics
- ✓ Coordinated with engineering on how to support multiple paths
Outcome
This decision dramatically improved results. Each segment got an experience optimized for their needs, leading to 30% overall completion rate with high satisfaction across all user types.
Uncover & Optimize: Mobile Was Being Ignored
The original platform had only 0.3% completion on mobile—a massive missed opportunity.
The Problem
Mobile users were completely underserved. The old flow required scanning, uploading, form filling—all terrible on a small screen. 0.3% completion rate showed mobile wasn't just overlooked, it was broken. With 9% of users being mobile-primary, this represented massive untapped potential.
How I Solved It
- ✓ Designed mobile-first workflow separate from desktop
- ✓ Implemented smart OCR to minimize manual input on small screens
- ✓ Used progressive disclosure to show only critical fields
- ✓ Optimized for camera/microphone capabilities of phones
- ✓ Tested extensively with users on actual devices
Outcome
Mobile completion jumped from 0.3% to 18%—a 60x improvement. This wasn't just better numbers—it meant mobile users could now actually complete reimbursements, a feature that was previously broken for them.
Collaborative Influence & Team Approach
Patterns That Became Our Standard
The approaches I developed for this project became how our team solved similar problems going forward. Other designers on the team and across the product line adopted these patterns.
Progressive Disclosure Pattern
How we balanced compliance requirements with usability—showing information contextually rather than overwhelming users upfront. This became the team's go-to approach for regulated features.
Adopted by: Product design team (adopted in 3+ projects)
User Segmentation for Design
Instead of trying to serve everyone with one design, create fundamentally different experiences optimized for each user segment. The UI designer on the team found this approach valuable and has used it since.
Impact: Improved outcomes on 2 subsequent projects
Research Without Traditional Testing
Due to healthcare privacy constraints, we developed a research approach using analytics, customer service transcripts, and stakeholder interviews. The researcher on the team expanded this into a toolkit for other projects.
Now standard: For all healthcare projects requiring privacy compliance
How Our Collaboration Worked
This project showed what becomes possible when design, research, product, and engineering work together from day one with real alignment on goals.
Design-Research Partnership
Instead of research being separate from design, the researcher and I worked together throughout. She helped me understand user needs more deeply, and I could test approaches in real-time.
Early Engineering Involvement
The engineering lead was involved in design decisions from the start. When I proposed OCR for receipt capture, he knew immediately if it was feasible and what the trade-offs were. This prevented wasted effort.
Product Manager as Translator
The product manager helped us navigate competing priorities and communicate decisions to stakeholders. When we pivoted to multi-path design, they built the business case that secured buy-in.
Compliance as Partner, Not Gate
Rather than compliance reviewing designs at the end, we invited them into the process. This prevented costly late discoveries and led to better solutions.
Shared Learning & Development
This project was a learning opportunity for everyone involved. We documented what we learned and shared it with our immediate team and cross-functional partners.
Within Our Team
- ✓ Weekly design critique sessions
- ✓ Documented design decisions & rationale
- ✓ Case study walkthrough for new designers
- ✓ Shared research methodology playbook
Cross-Functional Sharing
- ✓ Presented approach to design community (team meetup)
- ✓ Shared progressive disclosure pattern with other teams
- ✓ Consulted on 2 similar healthcare redesign projects
- ✓ Mentored designer on another product line
User Experience Improvements
Business Impact
🏢 Operational Efficiency
Automated workflows dramatically reduced manual processing and support burden, allowing teams to focus on higher-value activities.
😊 Customer Satisfaction
Improved self-service capabilities led to higher platform engagement and reduced customer frustration.
Scalability
The redesigned system now handles increased transaction volume without requiring additional support resources.
Compliance
Enhanced documentation and audit trails improved regulatory adherence and reduced compliance risks.
Key Learnings and Design Insights
Long-term Impact & Future Directions
"The redesign transformed our support model and became our blueprint for improving other complex workflows."
Key Learnings & Design Principles
What This Project Taught Me
Constraints Are Catalysts, Not Barriers
I initially saw HIPAA restrictions, IRS compliance, and legacy systems as obstacles. But the best solutions came from working within these constraints. Privacy requirements drove us to smarter authentication patterns. Compliance needs led to progressive disclosure. Our biggest innovations emerged because we couldn't do the obvious thing.
Diverse User Needs Demand Diverse Solutions
The instinct is to find "the best" design that works for everyone. But 4 distinct user segments with conflicting needs can't be served by one interface. Sometimes the right move is to build 4 different experiences optimized for each segment. This goes against the "elegant solution" instinct but delivers better outcomes.
Data Informs But Doesn't Decide
Analytics showed us the problems (98.9% abandonment, 18-minute flows). But they didn't tell us why or how to fix it. Good research—combining data with transcripts, stakeholder interviews, and observation—revealed the true mental models and needs. Data opens the door; qualitative research walks through it.
Mobile Matters More Than You Think
We initially prioritized desktop. But 0.3% mobile completion meant the feature was effectively broken for mobile users. When we properly optimized for mobile as a first-class experience (not a responsive afterthought), it contributed 60x improvement. Mobile isn't an edge case—it's core.
How Great Design Happens
This project showed me that exceptional design results from exceptional collaboration. The role of a Lead UX Designer is to orchestrate, not dictate.
Involve Specialists Early, Not Late
When engineering is in design discussions from day one, they can say "OCR is actually feasible" instead of discovering months later it's impossible. When compliance is a design partner, not a gate, they help you build trust, not block progress.
Research Guides, Design Proposes, Data Validates
Research reveals what's broken. Design proposes how to fix it. Data tells you if it worked. Each step builds on the last. None of them alone is sufficient.
The Product Manager Translates Vision Into Reality
Your brilliant design idea needs someone who can build the business case, navigate stakeholders, and secure resources. That person often isn't the designer.
What I'd Do Differently Next Time
These are the small adjustments that would make a big difference on the next complex redesign project.
Validate Assumptions Earlier
We spent 3 weeks on one design direction before testing with users. Quick guerrilla testing in week 1 would have saved 2 weeks of work.
Start Mobile First, Genuinely
We designed desktop first and adapted for mobile. Should have designed mobile-first since it forced better prioritization.
Build Measurable Hypotheses Into Design
Design with specific metrics in mind from the start. "This flow should reduce completion time by 40%" is better than "make it simpler."
Document Decisions as You Make Them
Waiting until the end to document 'why we designed it this way' loses the context. Do it as you go.
Plan Knowledge Transfer Early
Don't wait for the project to end to share what you learned. Share weekly with the team throughout.
How This Shapes My Work Now
Every project I do now is influenced by what I learned here.
I always ask: What are the real constraints, and how might they inspire better solutions?
I design for the messy real world, not the ideal world. Complexity isn't a bug; it's a feature to understand.
I invest heavily in research because I've seen how wrong assumptions can derail great design.
I involve specialists early because I know what happens when you wait until the end.
I measure impact relentlessly because insights without validation are just opinions.
Looking for more?
Check out my other projects