CONTEXT
Making fragmented healthcare data actually useful
Montecito Medical's practice executives were making high-stakes decisions: tracking competitor rates, monitoring referral patterns, and planning real estate strategy by using data stitched together from three different platforms into PowerPoints.
It was slow. It was siloed.
And there was no single place to see the full picture.
Every analysis = manual effort, every time
One Platform. Executives self-serve
Old Model
Every analysis = manual effort, every time
New Model
Every analysis = manual effort, every time
One Platform. Executives self-serve
Every analysis = manual effort, every time
One Platform. Executives self-serve
WHERE WE STARTED
From five demos to a clear direction
The client walked us through five products to give us full context on the space and ambitiously laid out everything they'd need help with. The challenge wasn't a lack of vision. It was that everything felt equally important.
So before I touched a single screen, I presented our approach back to them. We aligned on business goals, defined success metrics, and mapped out high-level user journeys and personas together. That conversation is what gave us our priorities and a clear place to start.
MAKING SENSE OF EVERYTHING
I fed all the information and notes from demos and conversations into NotebookLM to extract themes, spot patterns, and identify what actually mattered. What would have taken days of synthesis took hours.
Screenshot from my NotebookLM
BUILDING THE STRATEGIC FOUNDATION
I used Claude and Perplexity to accelerate strategy: researching the healthcare market intelligence space, testing persona assumptions, and drafting user journey frameworks to take straight to stakeholders for validation.
Business Goals
Opportunity Areas
Persona Summary
CREATING A VISUAL FOUNDATION
Using the company's brand guidelines, I created mockups in Figma Make to get a visual feel for the product and used that exploration to build out the design system components. The Gemini Chrome extension helped structure the components and align naming conventions with Angular Material.
Figma Make initial mockups
Gemini Plugin to structure design system components
DESIGNING WITH AI
AI didn't replace my process, it fueled it
Most of this project didn't follow a traditional research playbook. There was no time for extensive user interviews or field research; we had to move fast and validate as we built. AI tools became how I made sense of information quickly, filled research gaps, and kept momentum without sacrificing rigor.
Phase 1: Foundation
Phase 2: AI Prototyping
MAPPING THE PLATFORM ARCHITECTURE
Navigation across modules
Simultaneously, I defined how executives would move across modules by designing a two-level navigation system: primary navigation to move between modules and secondary navigation to explore content within each one.
SOLUTION FOCUS
Two modules built for executive decision-making
1. C-Suite AI Dashboard
A real-time health check for practice executives, surfacing performance across four core areas with AI-generated insights and recommendations, all without digging through a single report.
Decision 1: Health score cards and gauges
We moved from data density to glanceable truth, using a traffic-light system so an executive can assess practice health in under 10 seconds.
Decision 2: Inline view over other interactions
We explored several interaction patterns — tabbed priority views, floating widgets, split panels, and progressive disclosure. We landed on an inline view+priority view because it required the least navigation and gave executives the fastest path from score to context.
Exploring interaction patters
Decision 3: Tabbed priority (Immediate Action, Monitor Closely, Healthy)
Organizing KPIs by urgency rather than a flat list means executives know exactly where to look first. If asset utilization is low, the immediate action tab shows precisely what's pulling it down.
Decision 4: Structuring the AI detail view
The AI was generating dense summaries. I identified consistent patterns across all outputs — 30/7/1 day views, a short summary, and top and bottom contributor comparisons and turned them into a repeatable card structure that works across every health area.
With AI-generated data, I also designed for uncertainty — loading, partial data, and empty states so the interface always communicated honestly regardless of what the data layer returned.
2. Atlas: Healthcare market intelligence
Old workflow
Previously, delivering market insights to a practice meant pulling data from ESRI, Trilliant, and Google, filtering it in Power BI, building charts manually, analyzing opportunities, and packaging everything into a PowerPoint deck. Every time. For every practice.
PAIN POINTS
Entirely Manual and Labor-Intensive
Lack of Scalability
Data Silos and Fragmentation
Static and Inflexible Deliverables
Cumbersome and Complex Analysis
The six modules at a glance
Atlas automates the entire pipeline, giving practice executives direct access to their market data across six modules without waiting for a report.
Decision 1: Defining the data visualization
Each module required a different answer to the same question: what's the best way to show this? Some data was spatial; it needed a map. Some was comparative; it needed a chart. Some was granular; it needed a filterable table.
Decision 2: A filtering system
Filters were central to every module. Executives needed to slice data by geography, specialty, provider, procedure type, and time period.
Filter states: closed → open → applied
Built to grow
These six modules form the foundational layer of Atlas. The next phase involves refining map interactions, deepening filters, and adding an AI recommendation layer.
Patient Migration Trends
Market Share
Turning healthcare market data into executive strategy
Montecito Medical is a medical real estate company that provides physical space to medical practices. Beyond that, they want to become a true operational partner for the executives running those practices, giving them tools to monitor performance, understand their market, and make faster decisions.
The data existed. The platform didn't. I was brought in as design lead to build it from the ground up.
Unifying fragmented data for 380+ medical practice executives across a $6.5B real estate portfolio.
CLIENT
Montecito Medical
ROLE
Design lead, strategy through delivery
TIMELINE
4 months
FOCUS
Setting design direction, building the design system from scratch, and guiding a second designer throughout the process
COLLABORATION & EXECUTION
I worked directly with business leaders on product vision throughout to translate ambiguous goals into design decisions they could evaluate and commit to. With engineers, I collaborated on component behavior and made sure every pattern was buildable without rescoping.
With the product manager, I worked through exactly what AI-generated data should surface, what to abstract, and how to make users trust outputs they didn't generate themselves. I also guided a second designer on the project, providing direction on design system structure, component decisions, and how to approach modules where the scope was still evolving.
KEY TAKEAWAY
In an AI world, decisions became the real work.
This project changed how I think about design in an AI-augmented environment. When you can generate ten credible options in the time it used to take to sketch two, the bottleneck stops being ideation and starts being judgment. My value shifted from how fast I could produce screens to how effectively I could direct the AI, using critical thinking to filter and refine ideas into something that actually worked for the business and the users.
The real challenge was designing for trust. I realized that if an executive doesn’t trust the data, they won’t use the tool. I focused on making the AI feel like a transparent colleague, using clear loading and missing information states, and layering information so users could get a quick summary or dive deep into the why whenever they needed to.
Executive AI Summary