Listen. Ask. Write it down.
- 01 Doctor asks a question
- 02 Patient shares their response
- 03 Doctor documents the conversation
Attention is shared between the patient and the documentation.
02 / SELECTED WORK HEALTHCARE · 2025
Soapee.
AI Medical Documentation System
More space for the conversation.
A little less paperwork.
- CONTEXT
- BitHealth hackathon
- MY CONTRIBUTION
- Research · Product design · Development
- BUILT WITH
- Next.js · Gemini · Supabase
01 / THE IDEA
Care starts with
a conversation.
Our starting point was the work around the consultation: documenting what was said while trying to stay present with the patient. We looked at published research to understand that tension.
We designed Soapee to help document and analyze the conversation, with the aim of giving healthcare professionals more space to focus on their patients.
THE RESEARCH BEHIND THE IDEAWhere does
Where does
the day go?
In a study of 57 U.S. physicians in outpatient practice, EHR and desk work occupied almost half of the office day.
EHR = electronic health record.SHARE OF THE OFFICE DAY
- EHR & desk work
- 49.2%
- Face-to-face with patients
- 27%
FROM FINDING TO DESIGN
Let the conversation lead.
In RAND’s study of 30 U.S. physician practices, doctors described time-consuming data entry and interference with face-to-face care as sources of dissatisfaction. RAND, 2013 (opens in a new tab)
Our design response: help turn the consultation into a structured draft, then give the doctor space to review, edit, and confirm it.
Keep the conversation going.
- 01 Doctor and patient talk
- 02 Soapee assists with transcription
- 03 AI helps document and analyze
The doctor reviews and confirms the resulting notes.
THE PRODUCT, REIMAGINEDA conversation,
A conversation,
in four chapters.
Subjective. Objective. Assessment. Plan.
SOAP, with a friendlier name: Soapee.An illustrated walkthrough of the prototype. Explore each step.
soapeeFROM CONVERSATION TO CARE
01 / 04
Start with their words.
A recorded consultation becomes a transcript. Confidence highlights help the doctor spot words that deserve another listen, then make corrections.
The patient’s story becomes the starting point.
CONSULTATION
Speech to text
THE PATIENT’S WORDS
“I’ve been feeling unwell for a few days.”
Words to review Illustrative text
02 / MY CONTRIBUTION
From understanding
to making it work.
A shared effort, from the first questions
to the final solution.
Understand the problem.
Reviewed published research on how doctors spend their time and experience documentation, helping the team define the problem our prototype would address.
Help shape the product.
Worked with the team on product design, translating our research into a proposed workflow for AI-assisted documentation.
Build it together.
Contributed to developing and implementing the final solution, bringing the research and product direction into a working prototype.
WHAT I TOOK WITH ME
This experience strengthened my ability to translate research insights into practical technology, with a focus on supporting healthcare professionals and improving clinical workflows.
INSIDE THE PROTOTYPE
The tools behind
the idea.
- Next.js
- Application framework
- Gemini
- Generative AI
- Google Cloud
- Vertex AI · Speech-to-Text · Cloud Storage
- Supabase
- Database
MOVING FORWARD / PROPOSED IMPROVEMENTSWhere we wanted
Where we wanted
to take it next.
We outlined four directions for future development.
- Self-hosted language models
- Retrieval-augmented generation (RAG)
- Real-time audio analysis
- Support for a wider range of test results
03 / RECOGNITION
THIRD PLACEAn idea built together.
An idea built together.
A milestone to share.
Our team placed third at HealthTech AI Challenge 2025, themed “Exploring AI Innovation for Better Healthcare.”
- AWARDED TO
- Hamza Pratama
- DATE & PLACE
- 21 November 2025 · Yogyakarta
