What began as an initiative to centralize fragmented data and reduce call handle time for member support staff evolved into a broader, multi-pronged effort to drive efficient and impactful patient care. Throughout this transformation, I helped uncover business opportunities, shape UX strategy and execution, and define standards for future AI-powered innovation at SmithRx.
Company
SmithRx
Year
2025
Role
Lead Product Designer
Skills
AI, team alignment, UX strategy, prototyping, high-fidelity delivery
The primary goal of the Clinical Assistant was to decrease the time pharmacists spent reviewing Prior Authorizations (PAs), thereby improving both clinical efficiency and member outcomes.
For those unfamiliar, a PA is a form of pre-approval between the provider, patient, and insurance company to validate the necessity of a medication. These are notoriously complex and often cited as a top pain point across the healthcare industry. At SmithRx, the PA process was especially strained by legacy tooling and the lack of a unified view of the data needed to make fast, accurate determinations.
Building on recent momentum from designing an AI-powered application for our Member Support team, I collaborated closely with Clinical Operations and Engineering to map out pharmacist workflows and identify opportunities for intelligent automation. Together, we identified the decision points where AI could augment—not replace—clinical judgment, allowing pharmacists to focus on edge cases while routine decisions were accelerated by machine learning.
To ensure we maintained a high standard of usability and ethical responsibility, I developed a set of AI design principles that guided our interaction models. These principles emphasized clarity, consistency, and user oversight. This foundation was critical for trust and adoption, particularly in a high-stakes domain like healthcare.
One of the core design challenges was finding the right balance between simplicity and transparency. While it was tempting to abstract away complexity, I recognized that clinical users needed insight into model reasoning and underlying data sources to confidently validate system outputs. To address this, I designed layered interfaces that prioritized key information but allowed deeper exploration for those who needed it—what we internally called “progressive explainability.”
Beyond the product itself, I worked cross-functionally to influence how SmithRx approaches AI innovation at a broader level.
I facilitated working sessions to align technical constraints with user needs, and introduced collaborative rituals between design, data science, and clinical stakeholders to support faster iteration and shared ownership. These practices not only elevated the product’s quality, but also laid groundwork for future AI applications across the company.
Reflections and key learnings
This project challenged and expanded my capacity as a designer operating at the intersection of emerging technology, regulatory constraints, and real human impact. Three key learnings stood out: 1. Designing for AI is designing for trust —Transparent systems build confidence and adoption, especially in sensitive domains like healthcare. 2. Cross-functional fluency is essential — Embedding with technical and clinical teams helped me advocate for user needs while aligning with feasibility. 3. Scalable thinking pays off — Establishing AI design principles and collaborative processes extended the impact of this project to future projects and company culture.
Credits
Next Project
