Back to Projects

Pulse

HiLabsShipped

Redesigned for the physicians who stopped using it. Delivered in 2 weeks. Contributed to a $2M deal.

Clinical AIHealthcareRedesign

The Numbers

$2M

Year-one contract value

2 weeks

Discovery to delivery

3

Personas served

0

Physicians using old system

AI

Woven through every touchpoint

01 · Context

A tool everyone paid for. Nobody used.

A major US health plan had been running a provider care management system for years. On paper it did everything: tracking care gaps, managing patient populations, coordinating between physicians and care managers. In practice, physicians were not logging in.

This was not unusual. Provider resistance is the most documented barrier to health IT adoption in the US, and for good reason. Most clinical systems are built around what the health plan needs to report, not around what a physician needs to do their job. The result is a tool that adds burden without giving anything back.

The client knew the system was broken. They wanted a rebrand, a redesign, and most importantly they wanted providers to actually use it.

I was brought in as the sole designer. I immediately pushed to go beyond the UI surface and got direct access to SMEs and client stakeholders to understand the real problem. That access changed everything.

What was wrong with the old system:

  • Legacy grey UI from the early 2000s
  • Tiny, unreadable fonts
  • No clear information hierarchy
  • Overwhelming, irrelevant data
  • No AI or clinical intelligence
  • Built for the health plan, not the physician
  • No feedback loop for providers
  • Required navigating multiple screens for basic context
Old system illustration · Dense table layout, no hierarchy

02 · Discovery

The UX problem was really a value problem.

I did not start in Figma. I got in the room with the client, with SMEs, and critically with a physician who actually used or tried to use the system.

The UI was bad. But bad UI alone does not explain zero logins. Physicians navigate genuinely terrible systems every day when they have to. What the old system lacked was any reason for a doctor to come back tomorrow. No personal feedback loop. No clinical intelligence. Nothing that made their day easier or their practice better.

Then one conversation changed the brief entirely.

"Doctors are naturally competitive. They want to know how they're doing compared to other doctors."

Physician SME, during discovery session

This was not a UX insight. It was a behavioral one, and behavioral science backs it up. Peer comparison is one of the most powerful engagement drivers in medicine. Physicians who can see their performance relative to peers are measurably more motivated to close care gaps, improve outcomes, and return to the platform that shows them their score.

The system did not need to just become easier to use. It needed to become something physicians wanted to open.

The Strategic Reframe

Before

A reporting tool for the health plan. Data organised around what the payer needed to track.

After

A clinical intelligence platform for the physician. Designed around how physicians actually work.

Strategic reframe diagram

"I did not just redesign the UI. I reframed what the product was for and who it was actually designed to serve."

Stacy Carvalho, Lead Designer, Pulse / PCMS

03 · Process

Two weeks. Every hour counted.

With a compressed timeline and a live client relationship to manage, there was no room for slow iteration. I structured the work in parallel: discovery and design happening simultaneously, with client sessions feeding directly into design decisions the same day.

I ran the SME sessions solo, navigating clinical terminology, provider workflows, and health plan compliance requirements while simultaneously translating them into UX decisions. I also managed the client relationship directly, which meant earning trust fast and communicating design rationale in business terms, not design terms.

Three personas needed to share one system. The design had to feel native to each of them.

The three personas

The Physician

Needs personal value and speed. Has 8 minutes between patients. Wants to know their highest-priority patients and how they are performing versus peers.

The Care Manager

Needs shared context and the ability to act directly. Cannot afford to work in a silo from the physician. Must see the same patient data, in context.

The Administrator

Needs population-level compliance data. Tracks care gap closure rates, wellness visit completion, and immunisation performance across the entire provider network.

Persona cards illustration

Discovery to Delivery Timeline

Week 1

  • Stakeholder and SME sessions
  • Physician interview: competitive insight emerges
  • Persona definition
  • Strategic reframe
  • IA restructure

Week 2

  • Physician Dashboard designed
  • Patient List with AI summaries
  • Patient 360 longitudinal view
  • Daily Stats and Impact Index
  • Client presentation, designs approved

04 · The Design

From ignored to indispensable.

Every screen was designed to answer one question: why would a physician open this tomorrow?

Screen 01

Physician Dashboard

The entry point is personalised and AI-driven. The moment the physician logs in, the AI surfaces the 3 to 4 patients that need attention today and explains why. Not a table of 89 patients. Just what matters right now.

Physician Dashboard · Full width screenshot

AI Priority Feed

The AI explains the clinical reasoning. The physician knows exactly what to do next.

Glanceable KPIs

Active patients, high-risk flags, and care gaps front and centre with trend indicators.

Risk Stratification

Patient risk segmented by care category for at-a-glance panel understanding.

Screen 02

Patient List with AI Summaries

What previously required navigating three separate systems is now surfaced in a single row. Each patient comes with a complete AI-generated clinical picture before the physician opens a single chart.

Patient List with AI Summaries · Full width screenshot

Risk-First Architecture

High-risk patients always surfaced first, with risk percentages making clinical prioritisation immediate.

AI Clinical Summary Inline

Each patient row expands to show diagnosis, recent lab values, current medications, and history.

Screen 03

Patient 360 · Longitudinal View

The full picture of a patient, not just today's visit. A medication and diagnosis timeline going back years, vitals, AI suggestions for overdue tests, and the complete clinical history, all in one view.

Patient 360 Longitudinal View · Full width screenshot

Medication Timeline

Drugs added, continued, and discontinued plotted over years.

AI Suggestions In Context

One-click Place Order reduces the gap between insight and action.

Clinical Tabs

Patient Summary, Diagnosis History, Drug History, Procedure History.

Screen 04

Daily Stats · The Impact Index

This is where the behavioral insight becomes a product feature. The Impact Index shows each physician exactly how their care gap closure, wellness visits, and immunisation rates compare to the organisational average in real time.

Daily Stats and Impact Index · Full width screenshot

Impact Index

Your HbA1c, Wellness Visits, and Immunisation Rates versus org average, tracked over time.

Geographical Pulse

Percentile rankings by county. 'You are in the top 15 percentile in Raleigh County.'

Clinical Scorecard

Treemap of care category performance across all metrics.

05 · Key Design Decisions

Every tradeoff made deliberately.

Data-first dashboardAI-first priority feed

Lead with what matters

An AI-first entry point surfacing the 3 most critical patients with reasoning reduces cognitive load instantly.

Generic peer rankingEmpowering Impact Index

Competitive, but never punitive

The Impact Index was framed as a personal performance tool: your score, your trend, your geography.

Three separate productsOne system, three personas

Unified architecture

The navigation and permission model feels native to each persona while sharing the same underlying data.

AI as a separate moduleAI woven through every view

Intelligence as infrastructure

AI was embedded at every decision point: inline in the patient list, alongside the medication timeline, everywhere.

06 · Outcome

From POC to $2M contract.

$2M

Year-one contract value

2 weeks

Full UX strategy to delivery

3

Personas unified

Physician engagement by design

How design influenced the deal

I was in the client sessions from day one, not just delivering designs but actively shaping how the product value was communicated. The client trusted the design direction because I could articulate the reasoning in their language: physician adoption, care gap closure rates, value-based care outcomes. When the designs landed, the conversation shifted from "can you do this?" to "when can we start?"

Why the behavioral insight mattered commercially

The health plan's core problem was not aesthetic. It was a system with near-zero physician engagement. By solving the engagement problem at its root, giving physicians personal value and not just a better UI, the redesign addressed the client's most pressing business concern. A more usable but equally ignored system would not have closed a $2M deal. The strategy did.

07 · Reflection

What I learned in two weeks.

The speed of this project forced a discipline I now apply everywhere: ruthless prioritisation of the insight that changes everything. The physician competitiveness observation was not just a design decision. It was a product strategy decision. Recognising that and acting on it fast is what made the difference.

Working as the sole designer in direct client sessions also sharpened something important: design communication is as critical as design thinking. I had to translate clinical complexity into UX rationale, and UX rationale into business value, often in the same sentence. That skill, moving between languages, is what earned trust with the client and ultimately influenced the contract.

If I had more time I would have pushed further on the Care Manager and Administrator experiences. The physician view got the most depth. The other personas were designed correctly but could have been stress-tested more rigorously against real workflow edge cases.

Four lessons

🔍

The UX problem is rarely the real problem.

The client asked for a redesign. What they actually needed was a reason for physicians to engage. Getting to that distinction early and fast is the job.

🧠

Behavioral insight beats UX best practice.

Any competent designer could have made the UI cleaner. Only someone who understood physician psychology could have built the Impact Index. Domain knowledge is a design tool.

🤝

Client trust is earned in the room, not in Figma.

Being present in the sessions, asking the right questions, speaking the client's language: that is what turned a design handoff into a business partnership.

⏱️

Constraints sharpen decisions.

Two weeks is not a constraint. It is a forcing function. Every decision had to count. That clarity produced a more focused, more intentional product than unlimited time would have.