Role
Product Designer · Frontend Developer
Earlier healthcare AI work: an LLM-powered pre-consultation flow exploring patient data control, clinician handoff, and clinic wait-time reduction.
Artifact pass pending before public launch: replace placeholders with app screenshots, prototype frames, and testing artifacts.
Collecting relevant medical information before patient–clinician meetings lacked automation, largely due to staff shortages. In 2024, Toronto recorded the longest average wait in the province — 72 minutes.
No patient-focused system addresses medical personnel shortages, long wait times, and low patient engagement and agency in their own healthcare.
An app featuring general walk-in clinic navigation and an AI-powered chatbot that collects pre-consultation information — giving patients agency while easing the load on clinical staff.
“My pre-consultation feels rushed, and important information is often lost in the process.”
— Patient interviewEmpathy maps for patients and clinicians were created after reviewing literature, conducting user interviews, and distributing surveys — to understand emotional and practical needs during pre-consultation.
“Telemedicine utilizes chatbots, but do users understand how?”
— Competitor analysis“Users desired emotional support, such as greetings, and further treatment suggestions.”
— Literature review on chatbot interactionUsers prefer chatbot interactions with:
A 30-minute brainstorm shared team ideas effectively, surfacing requirements for future wireframes while keeping user needs front and centre.
Mind-mapping results fed into structured Crazy 8's sketching focused on the chat-summary flow — pushing creative variations within 8 minutes.
The strongest ideas became a conceptual prototype for a user study with questionnaires, semi-structured interviews, contextual inquiries, and detailed observation supported by keylogger data.
“If there's a human picture instead of the robot here, it will look more appealing.”
— User feedbackProblem: Apps with pre-consultation assessments don't let patients view the generated medical summary, limiting data control.
Solution: A summary review step was added to the chatbot conversation to enhance data transparency and patient engagement — including a generated summary with edit and submit options.
Usability testing compared two approaches:
“The patient reports chronic fatigue for one month, episodes of orthostatic dizziness, and intermittent cephalalgia.”
“The patient has had constant tiredness for a month, occasional dizziness when standing up, and occasional headaches.”
Finding: Most participants preferred summaries containing their original quotes — underscoring the importance of transparent data handling.
The editing feature was designed around technical limits. Two API approaches were tested; Request #2 (field-specific modification) proved faster and more reliable than Request #1 (full regeneration).
After early testing revealed confusion, a progress bar was added to show next steps, alongside pop-up messages to guide users and minimize cognitive load.
Wait-time impact was tracked through average wait time for a patient to be seen, and average patients seen per day — with and without the assistant. SUS was measured via a 10-question, 5-point survey.
“This felt more intentional — if I miss something I could ask again or mention it later on.”
— Participant“Summary is a lot more detailed than when I would normally talk to a nurse before seeing the doctor.”
— Participant“Reviewing the summary made me realize I forgot to mention a few things, so I added them before sending it off.”
— Participant“It's more helpful than the paper form. The chatbot is definitely more interactive and personalized.”
— ParticipantAs the sole designer on the team, I had to structure a non-linear product process around testing, technical constraints, and clinical feedback. The strongest portfolio next step is visual: add the actual prototype states, summary editing flow, and wait-time evaluation artifact so the case reads with the same evidence density as the Geotab work.
I'm here to listen, collaborate, and design solutions that make impact.
Get in touch! →