Understand patterns in daily context
People need help understanding how glucose patterns relate to meals, routines, and follow up decisions.
GlucoGuide is a continuous glucose monitoring concept designed to help people understand patterns, recognize when attention may be needed, and prepare better questions for their care team without turning the product into a diagnostic authority.
AudienceB2C
The full 25-screen Figma prototype spans CGM device pairing, glucose history, AI insight cards, medication logging, and the community flow.
Product designer across the health app structure, critical screen flows, and plain language interaction patterns.
Use CGM data, severity tiers, and carefully limited guidance to reduce interpretation burden without presenting a diagnosis.
Four product states show device setup, daily context, dosage insight, and a predictive warning with explicit care boundaries.
GlucoGuide is a continuous glucose monitoring concept for people managing diabetes. I designed the product structure, critical screen flows, and plain language patterns that connect sensor setup, daily context, dosage insight, and carefully limited guidance.
People could see readings without knowing what mattered, when attention was needed, or what to discuss with a care team. The product had to reduce interpretation burden without presenting itself as a diagnostic authority.
Global diabetes prevalence provides important context for the scale of the problem, but it is not a product result. The design focus remained on understandable patterns, restrained guidance, and a clear boundary between product support and clinical care.
People need help understanding how glucose patterns relate to meals, routines, and follow up decisions.
Overconfident language, excessive alerts, and unexplained AI confidence can create anxiety or false reassurance.
Use plain language, visible evidence, restrained alert tiers, and clear escalation language while preserving the role of the care team.
The sequence connects sensor setup, daily patterns, dosage context and a predicted low.
The setup state explains why the CGM connection matters and keeps the primary action unambiguous.
Current glucose, time in range, variability, and daily context appear as a calm hierarchy rather than a wall of measurements.
The dosage view places recent insulin context beside the weekly pattern while avoiding a claim that the product determines treatment.
The warning names the expected timing, current context, and a clear next action without presenting the prediction as a diagnosis.
The monitor state is treated as a high consequence communication problem, not another notification style.
A predicted low based on the current trend and the time window shown in the product.
Lead with plain language, make timing explicit, and place the recommended next action beside the warning.
People need to understand urgency without translating a dense chart or interpreting model confidence alone.
The warning receives stronger hierarchy than routine insights, while the detailed chart remains available for context.
The product can surface a pattern and encourage an appropriate follow up. It cannot diagnose, prescribe, or replace the care team.
Use plain language by default and introduce clinical terminology only where accuracy, medication context, or care team communication requires it.
The interface separates current status, trend context, and follow-up actions so a single reading never carries more certainty than the data supports.
Use three communication tiers: Informational pattern, Attention needed, and Urgent follow up.
Keep confidence visible on insight cards and connect it to the evidence behind the observation.
Plain language reduces interpretation burden, while precise terminology remains available when accuracy and care communication require it.
I kept the interface calm and explicit about limits, even when a denser dashboard could have surfaced more measurements at once.
The product can surface observations and next questions, but it should not sound like a clinician or imply diagnosis or treatment authority.
Current glucose, range, variability, and focused snapshot cards create a progressive path into deeper data rather than exposing every metric at once.
Health design demands a different kind of restraint. The instinct is to surface everything available, but the patient’s goal is to make fewer, clearer decisions rather than process every available data point. The most important design decision in this project was choosing what not to show.
Arizona Water Chatbot makes public water guidance safer and easier to navigate.
Prakhar Jain · Product Designer
3+ years in healthcare, enterprise software and digital products.
Open to relocation anywhere.
Explore selected work
I like sharing rough thinking, hearing useful disagreement and finding out what happens when an idea is built.
