Designing the Activation Intelligence Platform
for Clinical Trials at Mayo Clinic
COMPANY
Mayo Clinic
ROLE
Lead Designer and Researcher
TIMELINE
2026 to present
STATUS
Physician Validated
MY ROLE
Led end-to-end UX research and design for RSM Phase 2
Drove alignment across engineering, AI infrastructure, and physician leadership
Defined system-level workflow strategy across 10+ disconnected tools
Built the research program from scratch — no prior user research existed
16–30
Week activation cycle I'm designing to compress
10+
Disconnected systems replaced by one activation surface
30+
Users interviewed across 3 Mayo sites
3+
User groups researched: RPS, CRC, and APM
THE SITUATION
Before a patient can join a trial,
somone has to open it.
Trials move through review, approval, and setup across multiple systems before a single patient can enroll.
At Mayo Clinic, that process takes 4 to 8 months and involves dozens of people across three sites.
When it takes too long, sponsors go to faster hospitals. Patients lose access. Funding disappears.
The people managing this were doing it manually, across ten-plus disconnected tools, with no unified picture of where anything stood.
Nobody had ever designed for them.
THE INSIGHT
The problem wasn’t the process.
It was where it started.
Improving individual tools had not solved anything.
When I talked to the people doing the work, the pain was not in the middle of the process. It was right at the start.
Every error made at setup got corrected manually, over and over, across every downstream tool.
The whole workflow was paying for mistakes made in the first hour.
“I would really love to not have to enter the same data into multiple places. That’s the single biggest thing.”
HOW I WORKED
Thirty conversations.
It was where it started.
No research existed. I built it from scratch.
Fifteen coordinators across all three Mayo sites, meeting with me regularly throughout the project. Pharmacists. Program managers. Three groups, each with their own version of what better would look like.
While the research was running, I was prototyping — putting concepts in front of people in real time so every finding shaped what I tested next.
3
Distinct user groups researched
30+
Participants across interviews, usability tests, and surveys
3
Mayo locations represented
0
Prior research to build from
🔬
Research personas — coordinator, pharmacist, program manager
Three validated personas built from 30+ research sessions across all three Mayo sites.
THE VISION
Upload once.
Everyone else follows.
The hardest data work happens once, at the very beginning.
A coordinator uploads a single document. The AI layer reads it, figures out what every connected system needs, and routes it automatically. The coordinator never re-enters data. They work from one place.
The system carries the administrative burden. The human makes the decisions that actually require judgment.
🖥
Intake wizard and Activation Control Panel
Platform screens in active development. Prototype available on request
THE HARDER WORK
Keeping the human in
a technology conversation.
Every meeting pulled toward infrastructure. Engineers, AI architects, program managers all had strong opinions about what to build first. In that room, the user disappears fast.
Early on, conversations kept returning to system integrations. I re-anchored around intake errors and user workflow repeatedly, until the research made it hard to ignore.
The central design challenge for agentic AI is not automation. It is giving people enough transparency and control that they trust the system to act on their behalf. That is what the Activation Control Panel was built to do.
🗺
Stakeholder alignment across five teams
Coordinator panel, engineering, AI infrastructure, physician leadership, and program management — all aligned around a shared vision.
THE OUTCOME
Senior physicians called it
transformational.
In March 2026 I presented the research, platform architecture, and prototype to senior physician leaders at Mayo Clinic.
The room asked specific, substantive questions — the kind that come from people who think something is worth building and want to pressure-test it.
One physician described it using Amazon as a reference: start a task, finish it, never need to know how many systems are running behind the scenes. That framing came from the room, not from the deck.
“Love the current approach. Think it will be transformational.”