Aeri
AI-powered type 2 diabetes care and management via an AR-integrated contact lens
AI Hackathon
AR Wearable
Timeline
1 day (Feb 27-28 2026)
Role
Product Design — Interaction Design, Visual Design, Motion Graphics, System Architecture
Team
Melody Ekbatani (Product Design)
Isabella Mixton-Garcia (Researcher)
Tools
Overview
Aeri was created for the Parsons x University of Arizona 24-hour AI Hackathon, where teams were assigned. It is an AI-powered preventative care concept designed for people managing type 2 diabetes.
The project explores a more proactive alternative to today’s reactive health tools by helping users anticipate changes earlier. Aeri combines continuous sensing, predictive AI, and an augmented reality interface to deliver clearer, more actionable guidance throughout the day.
Design Challenge
What if type 2 diabetes care could become more preventative?
Problem Space
Diabetes care today is largely reactive
Where current tools fall short
Health tools provide constant data, but diabetes management still depends heavily on the individual. Users must interpret readings, track patterns, and repeatedly check devices throughout the day.
With sensors that require routine replacement, care becomes a cycle of monitoring, maintenance, and reaction. This revealed an opportunity for a more preventative and supportive experience.
1.0 Current devices show the data, but users are left to carry the cognitive load
Solution
Shifting toward proactive care by anticipating changes early
Our goal
We set out to design a system that helps people act before issues intensify. We focused on the management of type 2 diabetes approaching the problem through earlier awareness, clearer guidance, and more confident decision-making.
Our solution, Aeri, is an AI-powered preventative care concept that utilizes an AR-integrated contact lens to help users anticipate changes earlier throughout the day.
Key Features
Aeri is structured across three integrated layers
Biosensing
Continuously tracks metabolic signals
Captures glucose data in real time
Sends readings to a connected system
Smart lens
Delivers support through an AR-integrated lens
Interact with insights in real time
Adapts to your focus with three modes: widget, side panel, and full view
AR interface
Detects patterns across glucose and behavioral data
Predicts shifts before they escalate
Turns data into timely recommendations
Approach
A more supportive experience
Turning metabolic data into everyday guidance
Aeri’s core experience is built around four connected parts: live interventions, daily overviews, metric breakdowns, and user controls. Together, they make health insights more useful, timely, and actionable.
User controls also give people more choice over what is surfaced, how prominently it appears, and when they are notified.
2.0 Real-time blood glucose cards with contextual guidance and activity feedback triggered by current readings
2.1 Detailed views of cholesterol, glucose, and nutrition with range indicators and AI-generated insights
2.2 End-of-day overview combining glucose trends, LDL patterns, recommendations, and timestamped alerts
2.3 Granular settings for health metric tracking, notifications, privacy, and accessibility preferences
Context
Over 740 million adults worldwide have type 2 diabetes
A growing global health reality
Its scale reflects a growing need for tools that better support the ongoing decisions and mental effort of daily care. Managing type 2 diabetes is shaped by repeated choices around food, activity, sleep, and stress, all of which influence glucose levels throughout the day.
As a result, care often becomes a continuous process of monitoring, interpreting, and responding. This creates an opportunity to design tools that do more than report data by offering guidance that feels timely and easier to live with.
Market Research
Reactive systems create reactive behavior
This revealed a gap.
Conversations and user interviews point to a system that interrupts daily life rather than supporting it: frequent sensor failures, constant replacements, and the need to repeatedly check numbers create frustration and fatigue. Instead of feeling supported, many feel tethered to their devices, managing alerts, troubleshooting issues, and trying to make sense of inconsistent readings.
3.0 Research revealed the emotional and mental strain of reactive diabetes management
Competitive Research
Tracking diabetes data isn't easy
The limits of today’s tools
Across the category, products have improved in sensing accuracy, connectivity, and real-time display. However, reliability issues, short sensor lifecycles, and fragmented device ecosystems continue to introduce friction.
As conditions become more complex, the operational overhead increases, highlighting an opportunity to reduce system friction and create a more seamless experience that requires less active management from the user.
Iterations
Translating ideas into a working vision
Exploration
Mapped early system directions
Explored sensing, prediction, and interaction flows
Tested multiple concepts to identify viable pathways
Prompting
Generated interface behaviors and system responses
Explored interaction patterns at higher fidelity levels
Evaluated clarity, usability, and feasibility
Asset Creation
Built interface and visual system outputs
Developed scenes showing the product in context
Refined consistency across touchpoints
AI Workflow
How I worked
Building Aeri in 24 hours forced sharper decisions about where AI added speed and where it added cleanup. The biggest gains came from pairing tools to specific stages (research, ideation, production) rather than relying on one to do everything.
Perplexity
Research
Fast source-backed research on type 2 diabetes management, CGM technology, and existing preventative care tools.
Gemini
Research
Used for broader synthesis and quick comparisons. Helpful for connecting disparate findings into framing we could actually use in concept work.
ChatGPT
Ideation
Brainstorming partner for feature directions, user scenarios, and naming.
Figma Make
Ideation
Rapid prototyping of UI directions for the AR interface.
Claude
Thinking Partner & Writing
Pressure-tested concept decisions and refined pitch and case study copy throughout the sprint.
ElevenLabs
Voiceover
Generated the pitch video voiceover with natural pacing and tone.
Midjourney
Visual Production
Created mood imagery and conceptual visuals for the AR interface and pitch deck.
Reflection
My takeaways
A 24-hour sprint sharpened my time management, prioritization, and decision-making under pressure, where scope choices mattered as much as execution.
It also reinforced the value of dividing responsibilities early based on each person's strengths, which made the team faster and the final story stronger.
As my first AI hackathon, it expanded how I think about AI in design. I learned how to use prompting, rapid asset generation, and concept development more intentionally, while also considering scalability, feasibility, and product strategy.
Next Steps
Paths for further development
User Research
Expand user research with individuals managing type 2 diabetes to ground the system in lived experience.
Model Accuracy
Refine predictive modeling with clinical datasets to improve accuracy and reliability.
Biomedical Feasibility
Further evaluate the biomedical feasibility of the sensing approach and underlying hardware.
AR Validation
Test AR guidance interactions in real-world contexts to validate usability and behavior.
