Overall

🧠 The Challenge

When our team arrived at Epic Systems as part of our CS Capstone project, we were given a critical mission rooted in a growing challenge across healthcare:

​

"In recent years, healthcare has moved toward greater transparency by giving patients access to their medical data through online portals. However, much of this information—like test results and physician notes—is still written for healthcare professionals, making it difficult for patients to fully understand or act on their own health information."

​

We were asked to close this market gap by designing an AI-driven plugin—a “Friendly Mode”—to be embedded into patient portals like Epic’s MyChart. This plugin would process complex medical information and simplify it for patients directly within the browser, enhancing comprehension without altering the original clinical content.

​

Our solution:

SmartChart.ai, a conversational, AI-powered chatbot that guides patients through their health data using natural language explanations, reassuring context, and plain-language definitions. We were tasked with delivering a working prototype in just 12 weeks.

​

🔍 Understanding the Problem

We kicked off with qualitative research—interviewing MyChart users to understand where they struggled.

Key insights:

• “I don’t understand half of what’s written, so I just Google it.”

• “I’m not sure if something’s serious or routine.”

• “I wish I could just toggle a simple version of the results.”

Patients didn’t want more data—they wanted clearer meaning and reassurance. Their needs shaped our design focus.

​

🔄 Our Process

To manage design + dev in parallel, we combined Scrum + Kanban frameworks.

• Scrum: Weekly sprints with standups, retros, and rotating Scrum Masters.

• Kanban (Jira): Each team member pulled from prioritized tasks. I maintained the product backlog and coordinated between design and development.

As both the UX Designer and Product Owner, I ensured our design decisions reflected real user pain points, and that those insights translated into feasible dev milestones.

​

✍️ Ideation & Concept Development

With time constraints in mind, we scoped three MVP features:

1. “Friendly Mode” toggle – for simplified test results and doctor’s notes.

2. Contextual hovers – offering plain-language popups for complex medical terms.

3. Conversational chatbot – for guided walkthroughs and emotional support.

I led a collaborative ideation sprint using Figma and Miro, sketching wireframes and user flows based on user scenarios from our interviews.

​

🎨 Wireframing the Vision

I created a low-fi prototype in Figma, then tested with 6 users (including real MyChart users).

​

What worked:

• The toggle between simplified/original views felt empowering.

• Summaries eased user anxiety.

• Chatbot interactions made the experience feel more personal.

​

What we improved:

• Made the chatbot draggable and resizable.

• Added hover-based definitions for tricky terms.

​

Building the Experience

While the engineering team built the app in React.js, I continued to evolve the UX.

• Integrated Gemini AI plugin for summarizing health data

• Designed mobile-first, responsive screens

• Shifted from web scraping to file-based inputs for quicker implementation

• Conducted handoff with detailed design specs and Figma tokens

​