From Reminder Tool to Behavioral Support: UX Research & Design with Agentic AI for a Medication Adherence Mobile App

Project Summary: Led end-to-end UX research, iterative design, and validation for a behavior-first medication adherence mobile app called BetterMed. Designed core adherence flows and later extended the MVP with an agentic AI layer to support long-term behavioral consistency and emotional safety.

Responsibilities: UX Researcher, Behavioral Insight Synthesis, AI-Assisted Prototyping, Interaction Designer / UX Designer, Usability Testing & Iterative Refinement, Agentic AI Design.

Key Outcome: Validated core adherence flows with 100% task completion, eliminated structural friction from Round 1, and introduced an agentic AI layer that delivers personalized, context-aware support to sustain long-term adherence.

PROBLEM STATEMENT

How might we move beyond reminder-based solutions to support lasting medication adherence through behavior change, adaptive support, reduced cognitive load, and increased user trust? (Explored through an Agentic AI-assisted MVP for rapid concept validation)

SNEAK PEAK: BEFORE AND AFTER

Before: Tasks were completable, but the experience lacked the ease users needed.

  • Clunky information architecture and hidden medication states created confusion and disrupted continuity.

  • Generic symptom logging felt impersonal and lacked relevance for users.

After: The experience shifted from functional flows to more intuitive, repeatable interactions with less cognitive burden.

  • Strengthened dashboard visual hierarchy by elevating primary actions and re

  • Dashboard updates strengthened visual hierarchy by elevating primary actions (add / take / skip / reschedule), reducing tappable ambiguity, and improving scannability.

  • Transformed symptom and mood logging into a more guided experience that reduced decision fatigue and supported faster, more confident check-ins.

OUTCOMES & NEXT STEPS

Key outcomes from user testing included a 100% task completion with a decreased error rate of 5%.

However, research revealed the bigger opportunity was not improving reminders, but rethinking adherence as a behavior support problem.

That insight was rapidly explored through an Agentic AI-assisted MVP.

This insight is only the beginning of how BetterMed is evolving - with the full case study coming soon.