Sanofi, Eddie – A virtual coach for Type 2 Diabetes patients

Sanofi approached us to help them create a unique value proposition leveraging Generative AI to deliver personalised coaching support to Type 2 Diabetes patients.

To explore deeper levels of immersion, me and the design team worked alongside researchers to test and validate how to achieve engagement and stickiness through design, interactions and technology.

At the end of Phase 2 we had consolidated Eddie's value proposition around four functional areas: hyper-personalisation through user profiling, conversational support, self-monitoring and reflection, support learning and competency building.

Additionally, we developed an evaluation framework for Gen AI applications within Sanofi. Due to the healthcare system rigid laws and regulations, creating an evaluation framework of the data input and output within the LLM used to build an hyper-personalised space for patients was critically necessary.

My role & outcome: Understanding and making sense of current LLM developments and through that i was able to create an interface that uses voice and text as primary point of contact with diabetes patients (initial level).

General Research Participation: During research me and the research team understood that doctors needed help with keeping in contact with diabetes patients for a longer period than it was possible through a normal hospital visit, the app concept was built as an extension of the Diabetes specialist visit, following patients from the initial diabetes treatments through their life.

Ethical Approach to patient data & Ai Usage: The design choices i've made took in consideration the vulnerability of those patients since most of them were just aware of the disease. These design choices can be seen through the simple interface, friendly characters and calming colours.

User Research for colour and typography: Both the app aesthetic such as typography, tone of voice and colour were presented to patients and doctors during the research sessions, those inputs helped us understand the sentiment of patients and apply those with the design.

Learning Outcomes: The research process allowed me to remain grounded to the users i was designing for, being able to understand that patients were in a very sensitive part of their treatment allowed design to accommodate and incorporate elements that otherwise wouldn't be possible. Part of the concept initial ask was to explore 3d and realistic elements (realistic renders), these elements were removed due to bad ratings and feedback in the research.