PRODUCT DESIGN • 2025

Bridging Clinical Discovery
and Contextual Commerce

Scroll down to explore
Laptop mockup (to be added)
Overview
[i]
ValeHealth AI: Personalized Sleep Care
Transforming conversational AI into validated sleep improvement plans through a diagnostic journey that bridges user needs with contextual product discovery
Role
UX/UI Designer (AI Experience Design)
Deliverables
Research, Wireframes, High-Fidelity Prototype, Conversation Flows
Timeline
6 weeks
Tools
Figma, Miro, UX Research
Partners
Engineering, Medical Partners (Froedtert and MCW health network)
The Impact
35%
Increase in Trust
2.5x
Higher Conversion
DIAGNOSTIC DATA + USER AGENCY = PERSONALIZED PLANS = Higher Conversion
Problem / Background
[ii]
Context
Sleep issues are common, yet finding the right solution online is often confusing. Valehealth is a trusted health marketplace where users expect confidentiality and medically credible guidance. However, traditional product browsing feels overwhelming and impersonal, especially for sensitive health concerns, leading to drop-off.
Core Problem
Users cannot accurately self-diagnose their sleep challenges, which makes them hesitant to purchase products. Without a consultative bridge, the transition from “I’m tired” to “I need this specific supplement” feels sales-driven rather than health-driven.
Trust & Medical Foundation
Valehealth is built on user trust and close collaboration with medical providers and healthcare institutions. Any solution must respect privacy, reflect clinical insights, and communicate supportively rather than feeling sales-driven.
Opportunity
By leveraging AI-assisted diagnostics, we can transform the marketplace into a supportive health partner that provides “Prescriptive Commerce”, recommending products only after validating the user’s specific clinical needs.
Goals / Success Criteria
[iii]
Primary Goals
Conversational Clarity & Trust: Establish an AI-assisted intake that replaces static forms with a supportive dialogue, ensuring users feel heard before being “sold” to.
Dynamic Plan Personalization: Develop a “Sleep Improvement Plan” that allows users to curate and modify their own path to better sleep.
High-Intent Contextual Discovery: Bridge the gap between advice and action by integrating products directly into the plan via an interactive “Card Flip” mechanism, ensuring commerce is always relevant to the diagnosis.
Success Metrics
Conversation Engagement
To assess whether the AI-guided flow encourages participation.
Conversation Completion
To evaluate clarity and pacing of the AI-assisted experience.
Recommendation Interaction
To understand alignment between AI suggestions and user intent.
Research & Discovery
[iv]
1
Existing Experience Audit
I went through the existing static survey flow to identify friction points, drop-off risks, and limitations in adapting questions based on user responses.
Purpose
This helped establish why a linear, form-based experience was insufficient for collecting nuanced sleep-related information.
2
User Population & Research Context
Demographic analysis revealed that the core user base skews older, with the majority falling between 35–64 years old.
User population age distribution
3
Competitive & Market Research
Health and wellness platforms using conversational experiences were reviewed to understand how different interaction models support sensitive topics such as sleep.
This included chatbots, quiz-based flows, and guided experiences.
What I evaluated:
Gamified flows
Segmented multi-step forms
Conversational chat experiences
Focus on:
User effort
Clarity of questions
Perceived trust and credibility
4
Chat Flow Pattern Exploration
Different conversational structures were explored to identify an approach that balances information gathering, user comfort, and AI-assisted reasoning.
Patterns compared:
Gamified interactions
Quiz-style decision trees
Guided conversational chat
Key consideration
Which pattern allows AI to adapt questions while keeping user input structured and safe.
5
User Behavior & Mental Model Analysis
User behavior patterns and feedback from existing sleep-related tools were analyzed to understand how users prefer to discuss personal health concerns and how much effort they are willing to invest.
Insights I looked for:
Trust in medical or expert framing
Comfort with selecting vs typing
Preference for short, progressive questions
6
AI Opportunity & Constraints Analysis
Potential AI touchpoints were evaluated to determine where AI could meaningfully improve the experience beyond the static survey without introducing unnecessary complexity or risk.
What I evaluated:
Adaptive question sequencing
Recommendation logic
Explanation generation using medical sources
These considerations represent design-led hypotheses used to shape the experience; final AI logic and behavior were defined by the engineering team.
Decision Making & Design Principles
[v]
Conversational, not robotic
The chatbot was designed to feel supportive and human without mimicking open-ended AI chat. Language, pacing, and response structure were intentionally constrained to maintain clarity and credibility in a healthcare context, while still feeling approachable and calm.
Cognitive Load Optimization
Instead of replicating a long survey, questions were broken into small, sequential steps. This reduced cognitive load, made sensitive topics easier to approach, and allowed the system to adapt the flow based on prior responses.
Explain recommendations, not just present them
Product suggestions were always paired with clear reasoning and medical context. Explaining why a recommendation appears reinforces trust , supports informed decision-making , and avoids the perception of sales-driven automation.
Why This Approach
Most healthcare marketplaces fail because they treat medical concerns as simple retail transactions. By designing an AI-assisted diagnostic bridge I replaced passive browsing with an active clinical partnership. This model doesn't just sell a product; it secures user trust by delivering a validated improvement plan where commerce is the final, logical step of the journey.
Interaction Design & Prototypes
[vi]
From AI-Guided Conversation to Informed Product Discovery
This section demonstrates how the experience is intentionally structured into two primary stages: guided conversation and personalized planning, supported by AI-assisted reasoning, evidence-led guidance, and user-controlled interactions.
Stage 1 - Main chat interface
Stage 2
Stage 3
What I Learned / Next Steps
[vii]
Key Learnings
Conversational UX needs tight scaffolding to avoid confusion
Users respond better to explanations behind suggestions
Future Improvements
Add sleep score tracking
Integrate personalized tips over time
← Back to Selected Projects