Designing trustworthy AI
for modern healthcare
Hami is an AI clinical companion by Boston Health AI. I led its design end to end, for the patients who talk to it before a visit and the physicians who rely on it during one.
Lead Product Designer
Patient and Physician Apps
Healthcare AI
50+
Clinics running Hami
22+
Specialties configured
3
Countries live
2025
Public launch in Pakistan
Role
Lead Product Designer
Scope
End to end, patient and physician
Platforms
Patient mobile and clinical web
Timeline
4 months to launch
What Hami does
Patient details ready before the visit
Vitals, medications and lab reports come together before the consultation, so the visit starts with the full picture instead of a blank chart.
Vitals report
Current medications
Lab reports
The problem: doctors don’t have enough time
Within a limited consultation window, a physician must understand symptoms, ask the right diagnostic questions, review medical history, document the encounter, and make informed clinical decisions.
Through interviews we learned physicians were not looking for another digital tool to manage. They needed a system that quietly supports their workflow, reduces repetitive tasks, and provides better clinical context before and during consultations.
How might we design an AI-assisted clinical workflow that reduces friction for physicians without disrupting the patient-care experience?
Research with doctors and patients
Physician interviews
Emergency, primary care and specialists
Patient interviews
In Urdu, English and Arabic
Usability testing
Prototypes and live builds
Market research
Heidi, Nabla, Sully AI and more
What I asked doctors
Emergency, primary care and specialist doctors, interviewed between shifts
“
How do you collect a patient’s history when you are short on time?
“
Which parts of documentation feel repetitive, and what is missing when a patient walks in?
“
Where could AI help you decide, without taking control away?
What I asked patients
Patients in Pakistan and the UAE, interviewed in their own language
“
How do you describe your symptoms, and what do you usually forget?
“
What would make an AI assistant feel supportive rather than robotic?
“
Which language do you think about your health in?
Usability testing with doctors
Every major flow went in front of physicians before it shipped. These five moments changed the product the most.
01
The physician side moved from a mobile dashboard to a clinical web workspace.
Doctors reviewed patients at a desk, not on a phone.
02
Every suggestion shows its specialty and code, and the doctor ticks it.
AI suggestions without context were skipped.
03
Intake handles mixed languages and shows the split to the doctor.
Patients mixed English and Urdu inside a single answer.
04
A choice of doctor avatars before the conversation starts.
Patients opened up more when they could choose who they talk to.
How I used AI to research faster
I connected Claude to Mobbin through MCP, so I could ask for patterns in plain language and get real screens from shipped apps in seconds, then compare them with Heidi, Nabla and Sully AI before designing.
1. Ask
Describe the problem to Claude
2. Pull
Claude searches Mobbin through MCP
3. Compare
Group patterns next to competitors
4. Design and test
Turn them into Hami variations
References pulled from Mobbin through MCP for research purposes. All rights belong to their owners.
Competitor review: Heidi, Nabla and Sully AI
They set the bar for ambient documentation, and showed us where Hami could go further.
Heidi Health
What they do well
A calm note editor and flexible templates.
What we took
Template tabs and a cleaner note header.
Nabla
What they do well
Ambient capture that stays out of the way.
What we took
A quiet, always visible recording state.
Sully AI
What they do well
Thinking beyond the scribe, across the visit.
What we took
Proof that the whole visit is the product.
Most AI scribes start when the doctor presses record. Hami starts when the patient says hello, in their own language.
How the design changed
The first release proved the idea. Testing with physicians and patients reshaped almost every screen after it.
Patient app
Physician experience
In the news
Hami launched in Pakistan in 2025. In early 2026 the UAE introduced it through Emirates Health Services as Amal, its first AI physician assistant.
Gulf News
UAE unveils Amal, first AI physician assistant to transform patient care
Feb 2026
GEO News
Pakistani American surgeon rolls out UAE first AI clinical platform
Feb 2026
Economy Middle East
WHX 2026: Minister of Health reviews key AI powered projects in healthcare
Feb 2026
Business Recorder
AI won’t create more doctors, but it can extend their reach
Oct 2025
Dawn
AI powered health platform Hami launched
Aug 2025
The final design
A patient app that listens in their language, and a clinical workspace that writes the note.
Results
92K+
Patients supported globally
98%
Physician satisfaction
30%
Increase in provider capacity
“The platform thoughtfully integrates AI into clinical practice in a way that feels intuitive, efficient, and genuinely useful.”
Umer Ahmed
Director of Physician Affairs, Altus Community Healthcare
Live in three countries
Pakistan, the UAE and the USA
National rollout
Introduced in the UAE as Amal
In the press
Dawn, Gulf News, GEO and Business Recorder
“We don’t necessarily capture the essence of what the patient is saying, and Boston Health AI does a much better job of that.”
Dr. Jackson
Emergency Medicine Physician, Neighbors ER
“Just a few clicks during registration, and the patient is ready for the provider.”
Melinna Barrera
Lead Administrative Assistant, Neighbors Emergency Center
What I learned
Healthcare AI has to feel clear, reliable and quietly supportive. The goal was never to add complexity. It was to take repetitive work off a doctor’s plate and hand them back time with the person in front of them.
It also changed how I work. Pairing Claude with Mobbin through MCP cut research from days to hours, which left more time for sitting with doctors and watching where the design got in their way.
“AI in healthcare works best when it makes care feel more human, not less.”
/ More work
More case studies
