Grounding Design in Real Friction

To understand the core challenges, I didn’t just rely on metrics. I conducted extensive field research, participating in in-person COVE and C&S Listening Sessions in Seattle, and logged hours listening to live call queues at the Producer Help Desk (PHD). 

Key Insights Uncovered

• The Support Bottleneck: Burying help contacts drove agents to call the support desk immediately.

• Knowledge Fatigue: The Knowledge Center was a wall of unorganized, text-heavy tabs.

• High Design Debt: Multiple siloed teams built inconsistent interfaces, driving up developer effort and cognitive load.

The Strategic Solution

To solve this, I co-designed a standardized UX Design Roadmap and Project Brief Template to align product, content, analytics, and engineering across Credera, MPP, and Onboarding teams. 

Re-Architecting Onboarding and System Logic

Before moving to UI design, I mapped out the system logic for how our agentic features would operate. I designed three distinct system flowcharts to explore alternative structural models: 

• The Orchestration Flow: Defined how automated onboarding tasks sync responsively across web, mobile, and tablet platforms.

• The Personalization Engine: Created rule-based experiences that adapt default dashboards and shortcuts dynamically based on the agent’s specific sales region and whether they are dealing with new vs. renewing members.

• The Timing & Cadence Model: Planned progressive disclosure rules, separating tasks that require immediate attention during first login from those that can be safely deferred.

Copy is the Interface

When designing conversational experiences, words dictate the user journey. I partnered closely with content strategy experts (Carla and David) to iteratively draft, test, and refine natural, friendly, and plain-language microcopy. 

Our Responsible Healthcare AI Checklist

To avoid crossing the line from "helpful" to "intrusive," I established three core trust-building UI patterns: 

• Clear Data Citations: Every AI summary featured explicit citation links to the original plan PDF, avoiding "black box" decisions and preventing errors.

• System Transparency: Provided clear, visual waiting states to show what the AI is retrieving or analyzing.

• Consent & Friction Control: Designed explicit user controls to let agents easily edit, dismiss, or opt out of AI suggestions.

Unifying the Interface

I translated our conversational flowcharts into a high-fidelity interactive mobile dashboard mockup featuring clear visual checklists, step-by-step progress bars, and contextual help cues.

Maintaining Rigid Pixel Hygiene

All responsive layouts strictly aligned to the Abyss Enterprise Design System (Abyss v2) and mobile governance guidelines. I built all components using advanced Figma Auto Layout and structured variant sets to ensure flawless multi-platform consistency.

Clean Handoff & Accessibility

I led a major design system cleanup of our Notification components, authoring execution-ready handoff redlines and Dev Mode annotations:

• Touch Target Optimization: Guaranteed minimum 48x48 pixel touch targets for mobile.

• A11y Backlog & Compliance: Conducted WAVE and AXE contrast and screen-reader tab-order audits, collaborating early with our Accessibility team to resolve gaps prior to development.

Hard Metrics

• Support Deflection: Redesigning self-service contact routing and adding smart FAQ placements directly deflected high PHD call volumes, reducing operational costs.

• Iteration Time Reduced by 40%: The combination of a standardized Figma Design System, AI templates, and structured components cut cross-team design iteration times.

• Developer Hand-off Friction Dropped: Implementing structured annotations and Dev Mode specs reduced developer clarify-and-rebuild cycles.

Key Learnings & Future Roadmap

• Transparency Builds Trust: Personalization in healthcare must always explain why a suggestion is being made.

• Next Steps for 2026: Evolving our guidelines into Phase 2 AI Summary Standards, focusing on robust user-consent, undo flows, and data minimization techniques to ensure UHG remains a trusted leader in human-centered AI.