TECH INC.
Technology Incubation was a digital innovation lab at Memorial Sloan Kettering Cancer Center.
When I joined in December of 2020, the 17 of us would spend the next quarter building the Remote Oncology Sensing Engine (ROSE), a product that let clinicians conduct remote physical assessments of their patients.
Except ROSE was an alpha release with limited scope, and during those 3 months, the majority of the team went underutilized.
To realize our ambitious goals, it was clear something in our process needed to change.
PROJECT INFO
Tech Inc.’s goal was to “imagine the art of the possible.”
In addition to building ROSE, we took time debating exactly what that meant.
We concluded it to mean:
GOALS
- Identify new problem spaces that tech could improve.
- Identify new tech that could better solve problems now or in 1-5 years
- Integrate these discoveries into IT’s roadmaps.
TEAM
Me – Lead Product Designer
Louis R. – Lead Product Manager
VERSION 1.0 PROBLEMS
We lived the old process, too, so some problems were immediately apparent to us. The details, though, we sussed out with the rest of the team.
SHADOW PROJECTS
Rather than sit idle, team members would work on undocumented side projects.
While indicative of a motivated team, the lack of visibility made governance hard and prevented teammates from collaborating.
DEAD-ENDS
The experimental nature of Tech Inc. made dead-end pursuits a certainty.
But it wouldn't be valuable for Tech Inc. to say “don’t pursue NFTs, VR, or laser disc” when IT wasn’t going to do that anyway.
NEVER LOOKING BACK
There are multiple healthcare-related use cases for any given technology, but our first process would have us abandon a tech entirely after our initial exploration.
Remote Oncology Sensing Engine (ROSE) - Patient View
Me - UX/UI
Rest of team - Everything else
DESIGNING PROCESS 2.0
Tech Inc. had a lot going for it. The new process should take advantage of those strengths.
TECH INC. WAS SMALL
While that meant lower capacity, it also meant easier to manage, allowing an alacrity and flexibility not afforded to larger groups.
ITS MEMBERS WERE MOTIVATED & DIVERSE
We labored to find technologists with diverse backgrounds, skill sets, and interests when hiring. The new process should benefit from those accumulated experiences.
IT WAS AUTONOMOUS
Tech Incubation existed largely outside of the usual IT hierarchy and wasn’t beholden to other teams’ roadmaps.
A core principle for the new process was simplicity.
It couldn’t be a burden to the team's creativity.
Additionally, I wanted to resist accounting for all edge-cases right away, instead opting to modify the process only in response to a real event.
ROSE - Clinician View
Divya G. - UX/UI
Me - UX/UI/Research
Rest of team - Everything else
TECH INC. PROCESS 2.0
We wanted to go broad and deep into our explorations.
Our second process would see the team's output increase 5X.
SOURCE
Tech Incubation found opportunities primarily through three means.
Org strategy and roadmap
While we weren’t strictly beholden to the organization’s roadmap, we wanted projects whose benefit was immediately obvious to other teams.
Specific requests
High-up clinicians and administrators could “commission” Tech Incubation to explore an idea on their behalf.
We wanted to be open to collaborating, and specific requests would have a built in sponsor for a project.
“External” inspiration
Taking inspiration from sci-fi, industry trends, tech trends, etc., meant we weren’t stuck following the vision of the org’s few leaders.
We could explore bold ideas whose benefits wouldn’t be realized right-away.
EXPLORE
Team members not currently assigned to a project were free to choose their next assignment from a prioritized list, with a cap on team size.
A Two-sided Approach
A project team as small as one (1) could explore an idea from two perspectives:
Technology
What can this technology do?
What is it like working with this technology?
What are its limitations?
Use Case
What problems can this technology solve?
What area of the hospital would benefit most from this technology?
How is it being used in healthcare or other industries?
IMPACT
When a project had immediate opportunity, we'd partner with a stakeholder to create a minimal viable product (MVP) to run a 2-6 month real-world pilot study.
Once complete, Tech Incubation developed a plan with IT to incorporate into future roadmaps.
SUPPORTING STRUCTURE
IDEA SUBMISSION
Tech Incubation members had a responsibility to independently submit ideas to keep the shop running.
During “Idea Submission and Pruning”, team members would put forth new project ideas using a standardized Jira form, which we’d review together and vote on which ones to place in our "Ready to be Picked" backlog.
These projects could be picked up by any team member once their current project was completed.
Regardless of who proposed the idea.
Buttercup
A journaling app that detects a patient's voice for acoustic features suggesting mental health issues.
Effie R. - Data Modeling
Me - UX/UI/Concept
Sundeep Y. - Front-End/Back-End.
PROJECT CHECKPOINTS
The first few sprints of a project were spent doing research to understand technological affordances and opportunity areas, which informed the proof-of-concept built in subsequent sprints.
At sprint end, projects were evaluated to determine whether they should continue or allotted more resources.
When a project had run its course, we packaged any assets, code, and lessons learned, marking the project as complete.
.gif)
Foxglove
A gesture-based image manipulation tool to maintain sterility in a surgical environment.
Eric S. - Dev/UI
Me - Interaction Design/UX/Concept
CONCLUSION
We 5x’d our output to about 20 projects a year
I’d personally complete 10 during my time at Tech Inc.
Eventually, we’d partner with clinicians and other IT departments who would “sponsor” explorations.
These special projects would function more analogous to traditional software development and would take priority over other kinds of work.
DENOUEMENT
Tech Inc. would shutter its doors about 2 and a half years later.
At the time, we were working with IT leadership to integrate Real-time Location Services (RTLS) and machine vision into the new Kenneth C. Griffin Pavilion, set to be completed in 2030.
As of writing, what remains of Tech Inc. is still working on it.
freeDispo (Mango)
A joint-effort between Tech Inc. and the Informatics department. Re-thinking clinical order entry with AI/ML.
Louis R. - Product Management
Me - Lead Designer
Sundeep Y. - Front-End
Eric S. - Back-End