【Quick Inc.】Preventing the Collapse of Emergency Medicine with AI & Taxi Transportation. A New Medical Infrastructure for 119 Triage and Transportation Optimization Created by Medical Students

One in Two is Mildly Ill. Saving Japan's Emergency Medicine on the Verge of Collapse with the Expertise of Doctors and Medical Students

Thank you very much, everyone. My name is Atsuhiro Takeda, Representative Director of Quick Inc.

I am currently a sixth-year medical student at the University of Tsukuba and plan to obtain my medical license this fiscal year. Daily, during hospital training and field visits, I have witnessed the severe congestion in emergency medical scenes. With the simple but earnest desire to "save as many lives as possible," I founded Quick Inc. last October.

Emergency medicine in Japan is currently on the verge of collapse.

Year by year, the number of emergency dispatches continues to increase, and accordingly, the "hospital accommodation time" (the time from calling 119 to receiving treatment by a doctor at a hospital) is soaring. About 20 years ago, an ambulance arrived within 6 minutes of calling, and patients received treatment within 30 minutes. Today, however, it takes an average of 10 minutes for an ambulance to arrive, and 45 minutes to be accommodated at a hospital. If this trend continues, the accommodation time is predicted to reach 55 minutes by 2035, putting us on the brink of a crisis where the lives of critically ill patients cannot be saved in time.

The biggest factor causing this severe congestion is "emergency calls for mild symptoms."

Currently, of the ambulances running across Japan, one in two (about 50%) is carrying patients with mild symptoms who do not require hospitalization. The actual calls include reasons like "my cough hasn't stopped since yesterday" or "my contact lenses won't come out," which can hardly be considered urgent. Furthermore, running an ambulance once costs about 45,000 yen in taxes, including personnel and equipment costs. This has become a huge financial burden for local governments.

We propose a new emergency medical infrastructure that introduces screening (triage) using "medical AI" and "remote doctor consultation" at the 119 call stage, bypassing patients with low-urgency mild symptoms to taxi transportation.

Multiple AI Fine-Tuned with Specific Data. The 119 Triage System Undergoing Demonstration in Tsukuba City

Our system is an AI product integrated into the dashboard used by dispatchers when 119 calls are received.

It analyzes the conversation in real-time, such as "my father has chest pain," "since when," and "is he conscious," and recommends the next questions the dispatcher should ask on the screen. Furthermore, the AI instantly determines the "urgency and severity" of the patient based on the dialogue data and automatically generates a summary of the conversation.

This system is already patent-pending, and to handle sensitive medical information securely, we have signed a direct business entrustment contract with OpenAI regarding medical information handling.

As a technical advantage, our AI system does not rely on a single LLM. In addition to the GPT model, we have constructed a base engine by combining Gemini and Claude, and have highly customized it using actual dispatch and emergency data provided by the Tsukuba City Fire Department through fine-tuning and RAG (Retrieval-Augmented Generation). Compared to conventional triage methods presented at the Japanese Association for Acute Medicine in March 2025, we confirmed that the AUC (an index showing AI determination accuracy) improved by "over 40%," boasting extremely high accuracy at the practical level.

After being screened as "mild/non-urgent" by this AI determination, the call is transferred directly to the remote doctor call center integrated into the system, and a doctor performs the final mild symptom determination and triage.

If the doctor determines that there is no urgency, they ask the patient, "Would you prefer to be transported by taxi instead of an ambulance?" If the patient consents, the system automatically matches the patient with a vehicle from a local taxi company and the most suitable clinic nearby, making reservation arrangements immediately.

Drastically Reducing Municipal Budgets. Saving 16,000 Lives Annually and a 260 Million Yen Revenue Share Model per Prefecture

If we introduce this system and replace just "20%" of mild symptom patients transported by ambulances with taxi transportation, it will bring a dramatic breakthrough to the entire emergency medical system.

First, the hospital accommodation time for other patients who actually need ambulances can be shortened by an average of "10 minutes." Shortening this by 10 minutes is equivalent to saving "16,000 lives" newly per year nationwide.

Furthermore, in terms of dispatch costs, the tax cost for local governments, which used to be 45,000 yen per dispatch, will be drastically reduced to only "3,000 yen" by taxi transportation.

As for the business model, the main stream is to receive a part of this municipal emergency operation budget "reduction effect" as a service fee (reduction revenue share) for Quick. For example, in municipalities like Tokyo, Sendai, or Ibaraki Prefecture, if we achieve a 20% mild symptom replacement in 9 hours of operation, the service contract price per prefecture is estimated at approximately 2.6億円 (260 million yen) annually. By expanding operation to 24 hours or increasing the reduction rate to 40%, the price per prefecture will scale to 780 million yen, and further to 1.56 billion yen.

In addition to the revenue share from local governments, we have completed a hybrid monetization design from three directions: referral fees from partner taxi companies, and a portion of taxi fares from patients.

Currently, we are conducting a joint demonstration experiment of the AI system with the Tsukuba City Fire Department, and for the transportation system part, we are establishing a demonstration scheme by setting up study groups involving municipalities (fire department and medical policy division) like Sendai and Hitachi City and local taxi companies.

As Tsukuba University's medical AI startup, we will first establish solid evidence through demonstration experiments, aiming to secure a majority (50%) of municipal shares nationwide within 6 years. We invite local government officials suffering from emergency congestion and investors interested in our seed round funding (planned for next March) to join our challenge. Thank you very much.

Q&A and Feedback

Commentator (Mr. Itoh): Thank you very much, Mr. Takeda, for your excellent presentation that directly addresses the challenges of congested medical scenes with great social significance. The business model of obtaining a revenue share from the reduced costs of ambulances is also extremely logical, and I feel it is designed so that local governments have no reason to refuse.

As a question, businesses targeting local governments and medical institutions are known for having many stakeholders (fire department, medical association, local government), making decision-making and entry barriers extremely high. In this area, what is the "real strength" that Quick believes makes it superior to competitors and capable of breaking through these barriers?

Mr. Takeda: Thank you for the question.

As you pointed out, we recognize that the difficulty and barriers for a private company to enter the core of social infrastructure, such as 119 dispatches, are extremely high.

Our biggest strength in this is that our development team and management members are composed of "active doctors and medical students." For decision-making in emergency medical policy, the consent and consensus building of medical expert organizations, such as local "medical associations" and "Medical Control (MC) councils," are absolutely indispensable. Administrative officials and the fire department cannot decide alone.

Because we share the common language of doctors and deeply understand the logic and pain points of the medical side, we can directly build consensus from the doctor's side, arguing that "this system is indispensable to prevent the collapse of regional medical care." This depth of medical lobbying and domain knowledge is our biggest barrier to entry, which other companies can never copy.

Also, regarding the AI, after signing a medical information agreement with OpenAI, we receive demonstration data directly from fire departments like Tsukuba City, constructing a unique fine-tuning and RAG that combines the three major LLMs. As a result, we have established a highly accurate product package specialized for emergency triage rather than just a "general AI chatbot," which is a major technical strength.

Mr. Itoh: I see. Rather than selling to local governments, it's an approach of building consensus and alliances from the "inside of medicine," such as medical associations, and moving policy itself. I was convinced that it is an extremely strong advantage unique to a doctor and medical student team.

Another point: as a business KPI, there is a concern whether "patients who call 119 will actually consent to alternative taxi transportation." People calling are in a panic, but when a taxi is proposed, how do you prevent them from saying "No, send an ambulance," and how is the patient-side merit designed?

Mr. Takeda: That is indeed a very important point that determines the success of this service.

In fact, the disadvantage of being forcibly transported to a large hospital by ambulance is very large for patients as well. If a patient with mild symptoms is transported to a large hospital like a tertiary emergency center by ambulance, they will naturally be put on hold in the waiting room for hours because more severely ill patients are prioritized. Furthermore, if it is outside the specialty of that hospital, they may not receive appropriate treatment.

In our system, we match and reserve the "most suitable medium-sized clinic or specialized hospital to visit right now" in real-time through triage. When the doctor explains kindly and politely, "Rather than waiting for hours at a large hospital, you can receive treatment much faster at the most suitable clinic nearby with zero waiting time by the taxi we arrange now," we can gain consent with a high probability.

Also, as a social background, some local governments like Ibaraki and Mie Prefectures have started charging fees (selected medical fees, etc.) to patients with mild symptoms to commercialize ambulances. Since it costs extra money when transported by ambulance, but they can go to a suitable clinic faster by taxi, the incentive is easy to work, and we believe it matches this era's trend very well.

Mr. Itoh: I see! Resolving the pain of waiting times at large hospitals and providing value through smooth matching to optimal clinics. Furthermore, the policy movement of charging for ambulances is also pushing you forward.

I felt that the three-way benefit for patients, local governments, and the medical side is very well designed. I sincerely hope that you will gather the power of active doctors and medical students to realize this infrastructure that saves Japan's emergency medicine. I look forward to the progress of the demonstration experiment in Tsukuba City. Thank you very much.

Mr. Takeda: Thank you very much. We will do our best.