ELFIECARE · VISIT SCRIBE
Brought attention back to the patient and cut hours of documentation work for doctors and medical assistants
HealthTech
AI Productivity
B2B
Year
2026
Duration
3 weeks
Team
1 product manager
1 product designer (me)
4 developers
2 quality assurance specialists
Deliverables
UX/UI design
Design system updates

PROJECT OVERVIEW
Background
Visit is the core module of ElfieCare, a clinical workflow platform. The doctor records the consultation and adds medical notes, and the AI turns both into a structured summary.
The module shipped and now runs at two hospitals in Ivory Coast. The main users are doctors, supported by medical assistants, in high-volume clinics that see 80 to 100 patients a day.

PROJECT OVERVIEW
The problems
2 to 4 hours a day lost to documentation and admin, much of it after hours.
Split attention during the consultation, which weakens the doctor-patient interaction.
Patient history is hard to retrieved, so prior decisions are lost.
PROJECT OVERVIEW
Results and impact
Doctor can spend the visit thinking about the patient, not about the note. The paperwork finishes when the visit does.
Improved doctor attention during the consultation.
Reduced after-hours workload.
Cut repeated work at every follow-up.
Work is saved even when the connection or the hardware fails.

HEY YOU MAKE IT ALL THE WAY HERE!
Let's uncover more interesting insights!
CONTEXT
Users and constraints
Target audience:
Doctors and medical assistants in high-volume clinics, working through 40 to 80 patients a day.
The doctor runs the consultation
The assistant handles intake and documentation
Constraints:
Users are not heavy software users.
Connection, hardware and room noise are all unpredictable.
BRAINSTORMING & IDEATION
Mapping information
The Visit is the hub of the platform. It touches patient profiles, records, pre-visit reports, and Evidence chat. Before drawing any screen, I mapped every object the module handles, the information each one carries, and every action attached to it.
One diagram gave the PM and I a shared picture of the whole module, so we aligned on the main structure instead of discovering it screen by screen.

PROBLEMS - SOLUTION
Attention & Context
The doctor records and have the consultation information transcribed.
With preconfigured clinical note templates, consultation session turned into organized medical notes

A follow-up starts from what was previously happened. Past visit notes and the full patient profile, including data records, are available from the visit screen.

PROBLEMS - SOLUTION
Evidence
Evidence chat sits inside the visit, so the doctor does not leave the consultation to search. It arrives with the case context already loaded, so it can also answer questions about the patient in front of them.
The doctor can see where an answer comes from, how strong it is, and whether it applies where they practise, without leaving the room to check.

PROBLEMS - SOLUTION
Failures & Multitasking
Every failure case got a designed path to make sure the workflow is not broken.
Under an unstable connection, the visit switches to offline mode and keeps recording. Everything syncs when the connection is back.
When hardware fails, doctors can fall back to manual notes. The visit continues; only the input method changes.
When the environment is noisy, guidance is provided. If it is too loud, manual notes are suggested instead of producing a bad transcript.

When the doctor switches to another tab, picture-in-picture keeps the recording visible.

PROBLEMS - SOLUTION
Management
The dashboard carries the management job. It shows who is working, what is finished, and what is still open to help the operations team see the volume and pace of the whole team.

HEY YOU MAKE IT ALL THE WAY HERE!




