EMS Copilot Korea

EMS Copilot matches what paramedics know in the field (patient condition, location, time) with what hospitals know in real time capacity, specialist availability, treatment capability, and congestion.

EMS Copilot integrates patient condition, location, and time with real-time hospital capacity, specialist availability, treatment capability, and congestion. Starting in Chungbuk and Chungnam, it recommends the fastest appropriate hospital and supports the entire journey—from acceptance requests to transport and handoff.

Client

Team Side Project

Year

2026

Project type

Digital Healthcare · Mobile Platform

Credits

Product Designer / UI·UX Designer

Challenge

The nearest hospital was not always the one that could treat the patient

Paramedics assessed the patient, called hospitals one at a time, repeated the same clinical information on every call, and tracked who had answered, all while continuing emergency care.

Bed availability alone could not answer the question that actually matters which hospital can accept and properly treat this patient right now?

Research

Hospital selection delays were becoming a measurable regional risk

In 2024, Korea's 119 service completed about 1.79 million emergency transports. 27,218 of them took more than an hour from field departure to hospital arrival.

A 2023 Chungbuk smart emergency program

Applied to 38,832 cases, cut transport time by 3 minutes 6 seconds. As an external benchmark, it showed that shared real time resources can move the number.

Chungnam

Chungnam alone recorded 3,319 delayed transports, the second highest regional figure in the country. That is why the product starts in the Chungcheong region.

Key Market Signals

1.79M, 119 transports nationwide, 2024. And 27,218 transports exceeding one hour 3,319 delayed transports in Chungnam

Key Market Signals

1.79M
119 transports nationwide in 2024

27,218
Transports exceeding one hour

3,319
Delayed transports in Chungnam

My role

I designed the decision flow from field assessment to hospital confirmation and handoff

I translated the emergency coordination process into one structured mobile flow shared by paramedics, hospitals, and control teams, with a role based architecture so each group sees the same case at the level of detail its own decisions require.

Key Contributions

01

Emergency service research and problem framing

02

Role based information architecture

03

Triage, vitals, and GCS input

04

Explainable hospital recommendations

05

Parallel hospital request workflow

06

Live case and capacity dashboard

04

Explainable hospital recommendations

05

Parallel hospital request workflow

06

Live case and capacity dashboard

Key solutions

Three decisions that turn phone calls and guesswork into one coordinated run

EMS Copilot compresses the minutes between "patient assessed" and "hospital confirmed." Each solution removes one source of delay. Re entered information, unexplained recommendations, and one at a time calls.

  • Solution 01

    Capture patient information once, use it for the whole case

    Registration is split into four focused steps. Case information, demographics and triage, symptoms, and initial vitals.

    Large stepper controls, visible measurement units, inline validation, and a final confirmation screen reduce memory burden and block incomplete or implausible entries.

    That same record then drives the recommendation, every hospital request, and the eventual handoff, so nobody repeats the patient's story again.

    Design Principle

    Progressive disclosure: show only the information required at each moment.

  • Solution 02

    A recommendation that shows its reasoning

    A recommendation that shows its reasoning

    Seeing the "why" lets them confirm quickly when the recommendation fits, and override it when what they see in the field says otherwise.

    Design Principle

    A recommendation the paramedic cannot inspect is one they cannot act on.

  • Solution 03

    Request in parallel, stay visible to handoff

    Instead of serial phone calls, acceptance requests go to multiple suitable hospitals at once, with the structured case record attached.

    A live dashboard shows request status, capacity, and the transport in progress to paramedics, hospitals, and control teams alike.

    By the time the ambulance arrives, the receiving team has already read the case, so the handoff starts prepared instead of starting over.

    Design Principle

    Minutes saved on coordination are minutes returned to care.

Dashboard for writers

Accessibility & Usability

In emergency care, accessibility is part of operational safety

01

Large touch targets support quick, one handed input

02

Primary actions stay fixed, no scrolling to act

03

Every status is shown in both color and text

04

High contrast hierarchy supports at a glance reading

05

Inline validation flags implausible values immediately

Scalability

A reliable recommendation requires more than an AI model

Regional expansion would require integration with emergency resource systems, NEDIS,
hospital information systems, and locally defined transport protocols.

The data itself needs governance: clear ownership, refresh rules, and stale-data warnings for capacity figures.
And the recommendation needs safeguards. visible reasoning, manual overrides, role based access,
audit histories, and fallback procedures when real time information is unavailable.

The AI remains a decision support layer. Final clinical, acceptance,
and transport decisions stay with qualified professionals.

Next project

aqu