Have a design problem to solve?
I led a cross-functional team in developing the first AI-powered digital therapeutic experience for a J&J company.
Developed as a mobile app tested in clinical trials, the AI Coach grew out of research into the needs of patients who's average blood sugar is wildly out of control.
Behavioral analysis of people with diabetes through secondary clinical research
Technical landscape analysis of the necessary data sources and J&J's AI capabilities
Observing and interviewing patients in New York, Atlanta, and Los Angeles over three months
Understanding US regulations for using patient data in therapeutic settings
Among people with diabetes whose condition was wildly out of control, we found that self-managing through tools, no mater how easy to use, didn't address the real issues.
These patients needed help with the behaviors that led to their condition, and that help needed to be empathetic, simple to use, and specific to the moments that were most challenging for the individual.
First we needed to better understand the truth of life with diabetes. What were the challenges facing people with wildly out of control glycated haemoglobin levels.
Our subject matter experts felt that the only way to really make a difference would be to have a personal chef, a nurse, a personal trainer, and a life coach follow patients around 24 hours a day.
Our team had extensive experience talking to people with type-2 diabetes and was able to map out their typical journey. This allowed us to target users who we could impact the most, and who's needs would extend to the larger population.
Once we had a handle on the ins and outs of life with diabetes for our target users, we were able to identify what we needed our service to do.
We were then able to identify the minimum viable service and which components could be digital.
We then conducted research…
and co-created prototypes with patients.
Once we had a handle on where people were struggling, we jumped into converting the science into a digital user experience.
We figured out what data would provide the greatest most important insights…
we established what content would have the greatest impact on user progress…
and we created a script for our pilot MVP.
Then we designed an interactive prototype that we tested with our users with the help of a small research agency.