Small signals, sustained engagement
AI engagement strategy • Chronic condition management
AI-powered nudges for Teladoc's diabetes management program.
Context
Teladoc's diabetes management program serves over a million members in chronic care. The program had strong clinical infrastructure and dedicated health coaches, but engagement between touchpoints was inconsistent. Most members who disengaged did so quietly, and by the time a coach noticed, the window to intervene had often closed.
I designed the experience layer for an AI-powered nudge system: personalized notifications sent to mobile and cellular-connected devices, timed to individual behavior patterns and powered by predictive models that identified members at risk for uncontrolled outcomes before they got there.
Approach
The nudges used predictive modeling to suggest a next-best action for each member: coaching, digital activities, or self-monitoring prompts. The challenge was designing interventions that felt helpful without becoming noise. Timing, tone, and frequency all had to adapt to how each person was actually engaging with the program.
We also redesigned the weekly email members received, replacing standard newsletter content with personalized recommendations powered by the same models.
Impact
The research, conducted over nine months, was presented at the ADA's 84th Scientific Sessions in June 2024 and covered by Fierce Healthcare, the American Hospital Association, and Yahoo Finance.
Read the Fierce Healthcare coverage