Conducted a Proof of Concept for a Voice-Assisted Hospital Room Aimed at Reducing the Burden on Medical Staff and Improving Inpatient Satisfaction
TradFit and Toda Corporation conducted a proof of concept for a voice-assisted hospital room system using AI to distinguish between medical and non-medical patient requests, aiming to reduce the burden on medical staff.
📋 Article Processing Timeline
- 📰 Published: April 14, 2026 at 19:00
- 🔍 Collected: April 14, 2026 at 10:31
- 🤖 AI Analyzed: April 19, 2026 at 19:16 (128h 44m after Collected)
TradFit Inc. (Headquarters: Chiyoda-ku, Tokyo, Representative Director: Yoshiki Toda) and Toda Corporation (Headquarters: Chuo-ku, Tokyo, President: Seisuke Otani) jointly built a system realizing a "voice-assisted hospital room aimed at improving operations for medical workers and enhancing patient comfort and safety" (hereafter "this technology"), and conducted a Proof of Concept (PoC) at Asa Medical Association Hospital (Hiroshima City, Hiroshima Prefecture) from July to October 2025.
In this technology, utilizing Alexa Smart Properties*2 provided by Amazon, in addition to voice operation of facility equipment in the hospital (TV, air conditioner*3, lighting) and TV video calls between inpatients and their families, Toda Corporation outsourced to TradFit to develop a proprietary system. This system allows AI to distinguish and respond to requests from inpatients by comparing them with hospital rules and judging whether they are medical acts or not. A patent for this technology has been jointly filed by Toda Corporation and TradFit.
*1 Researched by TradFit (As of February 2026, regarding the introduction of Alexa Smart Properties in domestic medical institutions)
*2 A service that provides Alexa for businesses, and can be used in public services such as local governments.
*3 After the demonstration period (July to October 2025) ended, voice control for the air conditioner is currently suspended.
Figure 1: Image of a hospital room where this technology is introduced
1. Background of this technology
Labor shortages are becoming severe in the medical industry, making "work style reform" and operational improvement for medical workers urgent issues. Requests from inpatients range widely from those requiring medical procedures like "IV replacement" and "complaints of pain" to requests related to daily living support like "I want to turn on the TV," "It's glaring," or "It's cold." Currently, these are all notified to the nursing staff without distinction, creating a heavy operational burden.
Therefore, we focused on AI that distinguishes whether something is a medical act or not as a proprietary technology to achieve a reduction in the workload of medical workers. Specifically, we are introducing smart speakers equipped with interactive voice operation technology into hospital rooms. As a result, if the patient's request is for "living support (such as operating home appliances)," the AI will respond automatically and complete the task without involving the nursing staff, thereby realizing a direct reduction in the staff's burden.
2. Operation flow and functions
As for the operation flow, it responds in the following order.
① The smart speaker recognizes the voice command uttered by the user, and the cloud for the smart speaker performs language processing.
② A control signal is sent from the cloud to the infrared device.
③ The infrared device converts it into an infrared command and transmits it, thereby operating the target equipment.
④ When the user instructs a TV video call by voice, the smart speaker calls the pre-set device of the family member via the internet and starts the TV video call.
⑤ The developed AI distinguishes the questions and requests uttered by the user, and the smart speaker responds.
Figure 2: Operation Flowchart
3. Purpose and contents of the PoC
With the cooperation of the Hiroshima City Hospital Organization and the Asa Medical Association Hospital, we installed smart speakers in the hospital's private rooms (31 rooms in total) and verified the effectiveness of this technology. The main implementation details and verification purposes are as follows.
① Verification of needs and analysis of usage status
We conducted questionnaire surveys and interviews targeting inpatients and medical workers. In addition, we collected usage data such as the frequency of voice operation for facility equipment to extract needs and consider additional functions aimed at improving the utilization rate of this technology.
② Verification of response accuracy by AI and improvement of utterance handling accuracy
We verified whether the developed AI could appropriately respond to requests "other than medical acts" that frequently occur in medical settings and questions regarding the hospital's unique rules. Through additional learning for the AI, we verified its usefulness and aimed to improve the response accuracy so that it leads to a reduction in the burden on the nursing staff.
4. Demonstration Results
As a result of the usage status data, the smart speaker...
In this technology, utilizing Alexa Smart Properties*2 provided by Amazon, in addition to voice operation of facility equipment in the hospital (TV, air conditioner*3, lighting) and TV video calls between inpatients and their families, Toda Corporation outsourced to TradFit to develop a proprietary system. This system allows AI to distinguish and respond to requests from inpatients by comparing them with hospital rules and judging whether they are medical acts or not. A patent for this technology has been jointly filed by Toda Corporation and TradFit.
*1 Researched by TradFit (As of February 2026, regarding the introduction of Alexa Smart Properties in domestic medical institutions)
*2 A service that provides Alexa for businesses, and can be used in public services such as local governments.
*3 After the demonstration period (July to October 2025) ended, voice control for the air conditioner is currently suspended.
Figure 1: Image of a hospital room where this technology is introduced
1. Background of this technology
Labor shortages are becoming severe in the medical industry, making "work style reform" and operational improvement for medical workers urgent issues. Requests from inpatients range widely from those requiring medical procedures like "IV replacement" and "complaints of pain" to requests related to daily living support like "I want to turn on the TV," "It's glaring," or "It's cold." Currently, these are all notified to the nursing staff without distinction, creating a heavy operational burden.
Therefore, we focused on AI that distinguishes whether something is a medical act or not as a proprietary technology to achieve a reduction in the workload of medical workers. Specifically, we are introducing smart speakers equipped with interactive voice operation technology into hospital rooms. As a result, if the patient's request is for "living support (such as operating home appliances)," the AI will respond automatically and complete the task without involving the nursing staff, thereby realizing a direct reduction in the staff's burden.
2. Operation flow and functions
As for the operation flow, it responds in the following order.
① The smart speaker recognizes the voice command uttered by the user, and the cloud for the smart speaker performs language processing.
② A control signal is sent from the cloud to the infrared device.
③ The infrared device converts it into an infrared command and transmits it, thereby operating the target equipment.
④ When the user instructs a TV video call by voice, the smart speaker calls the pre-set device of the family member via the internet and starts the TV video call.
⑤ The developed AI distinguishes the questions and requests uttered by the user, and the smart speaker responds.
Figure 2: Operation Flowchart
3. Purpose and contents of the PoC
With the cooperation of the Hiroshima City Hospital Organization and the Asa Medical Association Hospital, we installed smart speakers in the hospital's private rooms (31 rooms in total) and verified the effectiveness of this technology. The main implementation details and verification purposes are as follows.
① Verification of needs and analysis of usage status
We conducted questionnaire surveys and interviews targeting inpatients and medical workers. In addition, we collected usage data such as the frequency of voice operation for facility equipment to extract needs and consider additional functions aimed at improving the utilization rate of this technology.
② Verification of response accuracy by AI and improvement of utterance handling accuracy
We verified whether the developed AI could appropriately respond to requests "other than medical acts" that frequently occur in medical settings and questions regarding the hospital's unique rules. Through additional learning for the AI, we verified its usefulness and aimed to improve the response accuracy so that it leads to a reduction in the burden on the nursing staff.
4. Demonstration Results
As a result of the usage status data, the smart speaker...