AI Design· UX Research · UX Design · IterateUX
A mobile app that helps patients understand the severity of their symptoms, describe them accurately, and prepare for doctor visits — designed to meet users where they are, not where the healthcare system assumes they are.
Context
IterateUX AI-UX Design Challenge · Team 1 · 5-week sprint
Role
Co-lead Researcher · Lead Designer · Presenter
Methods
Secondary research · Competitive analysis · 26-person survey · 4 user interviews
Outcome
🏋 1st place overall · Best UX Design Execution award
Users are unsure of symptom severity when experiencing medical symptoms — for themselves or someone they care for. Explaining those symptoms accurately can be equally difficult.
This causes anxiety and stress, and introduces the risk of under- or overdiagnosis — which can become a serious health risk, or a waste of time, money, and resources.
We ran a mixed-methods sprint across the 5-week timeline: secondary market research, a competitive analysis of five existing platforms, a 26-person quantitative survey, and 4 qualitative user interviews.
The survey gave us breadth — enough signal to identify patterns quickly. Interviews gave depth — the reasoning, language, and emotional texture behind the numbers. We embedded an opt-in at the end of the survey, so interview participants were already contextualized by their survey responses. This enriched interviews and helped us move faster.
I co-led research with a focus on synthesizing findings into actionable design direction, using AI as a synthesis partner to move from data to insight in hours rather than days.
Research Process
Secondary Research
Competitive Analysis
Survey
Interviews
Market Context
64% say the healthcare system is complicated to navigate. Half of consumers avoided care in the past year because finding the right doctor was too difficult. 2 in 3 who searched for a doctor spent an average of 5 hours doing so.
Primary Research — 26-person Survey
69%
were unsure whether their symptoms warranted medical care
80%
didn’t know what type of specialist to see based on their needs
65%
had previously under- or overestimated the severity of a symptom
Three themes emerged from synthesis. Also notable: 53% of respondents reported English was not their native language — a constraint almost entirely ignored by existing tools.
Theme 01
AI as tool, not decision-maker
Users want speed and synthesis from AI — but a human as the final word. This became a design principle, not a limitation to design around.
“An AI can be used as a second opinion model.”
Theme 02
Being understood is as important as being helped
Language barriers and difficulty articulating vague symptoms aren’t user failures. They’re design failures on the system’s part.
“If I’d had AI to help me describe the symptom back to myself first, I would have had an easier time explaining it to the doctor.”
Theme 03
Navigating logistics is as hard as navigating health
Insurance coverage, finding specialists, knowing who to trust — users become exhausted before they even describe a symptom.
“I think that will be helpful to give me words and ideas to explain my child’s symptoms.”
Research synthesis produced two personas that anchored every design decision. Every architectural and copy choice was filtered through one question: would Bethel understand this? Would Sarah trust this?
Persona 01
Sarah Jenkins, 32
Tech-literate · Health-conscious · Articulation-anxious
Avoids appointments because she’s afraid of wasting a doctor’s time. Not sure she can describe what’s wrong clearly enough to justify going. So she waits, Googles, and ends up more confused than when she started.
Persona 02
Bethel Brown, 38
Caregiver · Doctor-trusting · AI-skeptical
A caregiver managing her son’s health. Trusts her family’s doctors and is skeptical of AI. When medical terms appear without explanation, she disengages. When a tool gives an answer without showing its reasoning, she doesn’t trust it.
Insight → Decision
Articulation anxiety is a design problem
Users struggled to put symptoms into words. The solution wasn’t to ask them to communicate better — it was to design inputs that met them where they were.
Body Map
Tap where it hurts instead of describing in writing
Symptom Descriptors
Pre-set vocabulary (sharp, dull, burning) to select rather than produce
Symptom Sliders
Pain, fatigue, mood scaled 1–10 to lower cognitive load
Image Upload
Show rather than describe — with explicit in-app privacy communication
Insight → Decision
Users need to feel ready for the doctor, not diagnosed by an app
Users don’t want AI to replace the doctor. They want to arrive at the appointment prepared.
Visit Prep Checklist
Transforms fragmented symptom inputs into organized clinical talking points to bring to the appointment
Terminology to Know
Explains every medical term the AI uses in plain language, inline — no unexplained jargon
Transparent AI Reasoning
Shows what the AI is doing and why — a direct response to AI-skeptical users like Bethel
[Hi-Fi screens, Prototype]
Outcome
🏋 1st Place Overall
IterateUX AI-UX Design Challenge — out of all competing teams across the April 2026 cohort
Award
Best UX Design Execution
Awarded across all competing teams. Case study presented to a panel of UX leadership as a recorded presentation.
CareIQ is an AI product — but AI also had a real role in building it. I used AI to accelerate iteration between wireframe stages and to stress-test hi-fi screen directions faster than any traditional process would allow. The judgment always stayed human: which direction felt right for Sarah, which one Bethel would distrust, which copy would make a skeptical user lean in rather than close the app. This mirrors how CareIQ itself works: AI accelerates and synthesizes, but the human remains in control of every final decision.
Usability Testing
The most critical next step. The symptom input flows would benefit especially from testing with users who match Bethel’s profile: AI-skeptical, less comfortable with touch interaction, and navigating in a second language.
Emergency Escalation Path
CareIQ currently has no safety net for acute crisis situations. A static, always-visible trigger that surfaces 911 and ER guidance when high-risk symptoms are detected is the most critical design problem left to solve.
Provider-Side Research
The doctor handoff feature only works if its output is useful to the provider, not just reassuring to the patient. Future research would need to bring healthcare professionals into the process.
Three themes emerged from synthesis of our 26-person survey and 4 interviews. Also notable: 53% of respondents reported English was not their native language — a constraint almost entirely ignored by existing tools.
Theme 01
AI as tool, not decision-maker
Users want speed and synthesis from AI — but a human as the final word. This became a design principle, not a limitation to design around.
“An AI can be used as a second opinion model.”
Theme 02
Being understood is as important as being helped
Language barriers and difficulty articulating vague symptoms aren’t user failures. They’re design failures on the system’s part.
“If I’d had AI to help me describe the symptom back to myself first, I would have had an easier time explaining it to the doctor.”
Theme 03
Navigating logistics is as hard as navigating health
Insurance coverage, finding specialists, knowing who to trust — users become exhausted before they even describe a symptom.
“I think that will be helpful to give me words and ideas to explain my child’s symptoms.”
Outcome
🏋 1st Place Overall
IterateUX AI-UX Design Challenge — out of all competing teams across the April 2026 cohort
Award
Best UX Design Execution
Awarded across all competing teams. Presented to a panel of UX leadership as a recorded presentation.
CareIQ is an AI product — but AI also had a real role in building it. I used AI to accelerate iteration between wireframe stages and stress-test hi-fi directions faster than any traditional process would allow. The judgment always stayed human: which direction felt right for Sarah, which one Bethel would distrust, which copy would make a skeptical user lean in rather than close the app. This mirrors how CareIQ itself works: AI accelerates and synthesizes, but the human remains in control of every final decision.
Usability Testing
The most critical next step. The symptom input flows would benefit especially from testing with users who match Bethel’s profile: AI-skeptical, less comfortable with touch interaction, and navigating in a second language.
Emergency Escalation Path
CareIQ currently has no safety net for acute crisis situations. A static, always-visible trigger that surfaces 911 and ER guidance when high-risk symptoms are detected is the most critical design problem left to solve.
Provider-Side Research
The doctor handoff feature only works if its output is useful to the provider, not just reassuring to the patient. Future research would need to bring healthcare professionals into the process.
