An automated voice dialing system with interactive two-way communication capabilities.

Abstract

In medical settings and other fields requiring emergency response, rapid and accurate information dissemination and decision-making are essential. However, conventional systems often suffer from delays in communication and ambiguity in responses, leading to inefficient confirmation and re-notification processes. [Solution] The automated voice transmission system provides a system that enables information transmission through bidirectional communication between X and Y. Based on the text entered by X, it includes a transmission process (Process A) that automatically sends a notification to Y, a response process (Process B) that receives a response from Y, confirms the content, and asks further questions, and a notification process (Process C) that sends an interrupt notification or escalation when there is a delay in the response or before there is a delay in the response, thereby achieving rapid and accurate response management. Process B is the process of receiving a response from Y and analyzing its content, and if the response is inappropriate or insufficient, the system automatically generates further questions or confirmation messages to prompt an appropriate response.

Description

Technical Field

[0001] The present invention relates to an automatic voice transmission system for assisting interactive communication, and particularly to a system for realizing efficient two-way communication in a medical field or business communication.

Background Art

[0002] In recent years, with the progress of communication technology, efficient information transmission using voice or text messages has become possible. Systems as means of business communication have attracted attention as technologies that contribute to labor reduction and response speed improvement. Particularly in a medical field, it is advantageous to have efficient information transmission utilizing gaps in patient care and time-critical communication.

[0003] In addition to simple one-way voice transmission, a mechanism for supporting interactive responses between a recipient and a sender is useful, and further, by managing communication history and controlling notification priorities, time and labor can be made more efficient.

[0004] The message conversion server disclosed in Cited Document 1 is a system that converts text data received from a communication terminal into a voice message and transmits it as voice to a specified destination. This system allows a communication terminal user to send a voice message to a desired recipient and also to notify a fixed telephone or the like via voice.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0006] However, the system described in Reference 1 is intended for the automatic generation and transmission of voice messages, and voice messages are transmitted unilaterally only; two-way communication is not supported. Furthermore, because there is no recording or templating of response content, there is an inefficiency problem of repeating the same exchange, and there is also the problem of not being able to manage response history and tracking past communication content.

[0007] Furthermore, conventional technologies, which are specialized in general voice notification functions, have difficulty meeting specific business needs and lack the functionality required in situations where real-time is essential, such as in medical settings or emergency communications. In fields with specialized needs, such as medical settings, flexible notification functions that respond to the recipient's schedule and situation, as well as interactive responses for efficient communication, are crucial.

[0008] Furthermore, conventional systems have suffered from the lack of two-way communication and insufficient flexibility in notification functions. A mechanism that supports interactive communication between the sender and receiver, rather than simply one-way voice transmission, is useful, and there is a need for the ability to manage communication history and control notification priorities. In addition, in actual medical settings and business communications, it is necessary to optimize telephone and internet communications according to the situation and to implement flexible notification control according to importance and urgency in order to improve efficiency in terms of time and effort.

[0009] This invention has been made in view of these points, and aims to reduce labor, improve response speed, and achieve efficient communication, particularly in medical settings and business communications. Furthermore, it aims to provide a system that enables efficient and flexible information transmission and response by optimizing telephone and internet communications as communication means according to the situation and enabling flexible notification control according to importance and urgency. By applying this invention, it will support highly real-time information sharing and accurate communication even in medical settings and situations requiring emergency response, contributing to reducing the burden on users and improving the efficiency of operations.

[0010] In particular, by incorporating two-way communication and notification modules, we provide an automated voice communication tool (system, program, method) that enables interactive responses, improving practicality and efficiency, and contributing to reduced workload and faster response times in medical settings. [Means for solving the problem]

[0011] The present invention aims to solve these problems. The system of the present invention comprises a bidirectional communication and notification module and has the following functions.

[0012] (1) Transmission processing function The present invention is an automatic voice transmission system that enables bidirectional communication between X and Y, and includes telephone communication and internet communication as means of communication. Telephone communication involves the sending and receiving of voice messages via a telephone line by both the sender (X) and the receiver (Y). It enables two-way voice communication and supports real-time responses. Internet communication involves exchanging text messages and voice data over the internet. Both types of communication enable real-time chat messages, voice responses, and notifications. Because it does not use telephone lines, it allows for flexible information transmission while keeping communication costs down.

[0013] (2) Interactive response function It includes a template input module, an interactive question module, and a history management module to facilitate smooth bidirectional communication between X and Y. The template input module streamlines input work by registering standard messages. The interactive question module generates real-time questions to facilitate two-way communication, automatically generating questions necessary to check on the patient's condition (e.g., "How are your current symptoms?" or "Please tell us your preferred date and time for your next appointment") and receiving the answers to those questions. The history management module records and centrally manages the communication content and response history between X and Y, making it possible to track past message content, sending date and time, and response content. This helps to improve efficiency by templating the content of future communications based on the history. Furthermore, based on the predefined text created in the template input module, the interactive question module dynamically generates questions, and the questions and answers are recorded in the history management module and used for future communication, enabling efficient and smooth two-way communication between X and Y. <Interactive interaction> Interactive interactions between X and Y are preferably controlled using a generative AI or large-scale language model that satisfies the following conditions. The generative AI or large-scale language model must have at least 100 billion parameters. The training of the aforementioned model shall include specialized datasets for medical or business communication purposes. The aforementioned response is performed by combining options selected by the user from a specific template with a response dynamically generated by the model. <Examples of interactive responses> The automated voice dialing system is further equipped with a response control module. The interactive interaction between X and Y is controlled using a generative AI or a large-scale language model that includes the following processes. • Generating patient care-related questions for use in medical settings (e.g., "How are your current symptoms?") • Setting response priority for emergency situations (e.g., "What do you need to do to proceed to the next step?") • Natural language generation for notification messages (e.g., "Your scheduled appointment is for 10:00 AM today.") The following details will be added for further information. Algorithm Specification: "Generative AI or large-scale language models shall be built upon an attention mechanism (transformer)." Application Scenario Explicitation: "This system includes a setting that notifies users after the generated response has been reviewed by a healthcare professional."

[0014] (3) Schedule support function Designed specifically for use in medical settings, it detects the recipient's available time slots by referencing electronic medical records and consultation logs. Based on the detected time, it optimizes the timing of message transmission via telephone or internet communication. This enables efficient communication by utilizing gaps in consultation and work schedules. Furthermore, it's possible to use internet communication to send low-priority messages, or telephone lines to provide highly immediate notifications, allowing for a differentiated approach. Furthermore, it's possible to use both methods depending on the situation: immediately notifying important matters by phone while simultaneously sharing detailed information via the internet. For example, for urgent matters, you can contact the person by phone and simultaneously share detailed documents via email or chat.

[0015] (4) Notification control function Notifications will be sent in stages according to importance and urgency. Initial notifications will be sent via text message (chat message), followed by follow-up notifications via SMS, and finally by phone call, in stages according to the response deadline. Notification control based on urgency is as follows: for low urgency, notifications are sent via text (chat message), requesting a response at the recipient's convenience; for medium urgency, an SMS notification is sent after the text (chat message), combining multiple notification methods to prevent messages from being overlooked; and for high urgency, in addition to text (chat message) and SMS notifications, a phone call is made to contact medical staff regarding emergency patient care or to notify them of important tasks just before the deadline. Furthermore, when the deadline is exceeded, continuous phone calls are made to ensure the transmission of important matters.

[0016] (5) Support important decision-making processes in the medical field The response control module includes a confirmation message generation module for confirming with Y whether a specific treatment or action can be performed on a patient, together with information on the patient's hospitalization period for a specific ward or area, and a response option module that allows Y to select and input options of "good" or "bad" and, in the case of "bad", also input the reason. The confirmation message generation module is a function that generates a message to confirm whether appropriate treatment or actions can be performed based on information about patients in a specific ward or area (e.g., hospitalization period, medical treatment status), and conveys questions and confirmation content to Y simply and clearly. It obtains the patient's hospitalization period and medical condition from a database (response history DB, electronic medical record), identifies the necessary treatment or actions, automatically generates a message to confirm their feasibility, and sends it to Y. As a specific example, confirmation messages such as "The hospitalization period of the patient (ward Z) has exceeded one week. Should we proceed with the discharge preparation?" and "Should we perform a new examination on the patients receiving treatment in area Z?" are generated. The response option module is a function that supports Y's response to the confirmation message from X. It presents options of "good" or "bad" so that Y can easily input an answer, and provides a field for inputting the reason when Y selects "bad". By allowing additional reasons to be input in the case of "bad", specific situations are recorded and reflected in the next action. As a specific example, for the question "Should we proceed with the discharge of the patient?", the response options are "good" → "Proceed with the discharge procedures." and "bad" → reason input (e.g., "Blood pressure is not stable"), etc.

Advantages of the Invention

[0017] This invention overcomes the lack of two-way communication and insufficient flexibility in notification functions in the prior art, and achieves reduced labor, improved response speed, and efficient communication, particularly in medical settings and business communications. Furthermore, by optimizing telephone and internet communication as communication means according to the situation, and enabling flexible notification control according to importance and urgency, it enables efficient and flexible information transmission and response.

[0018] By applying this invention, it will support highly real-time information sharing and accurate communication in medical settings and situations requiring emergency response, contributing to reducing the burden on users and improving operational efficiency. [Brief explanation of the drawing]

[0019] [Figure 1] A diagram showing the overall network configuration of the automatic voice transmission system of the present invention. [Figure 2] A block diagram showing an example of the system configuration of an automated voice dialing system. [Figure 3] A schematic diagram of the processing of an automated voice dialing system. [Figure 4] A block diagram showing an example of the functional configuration of Server 20. [Figure 5] A follow-up chart showing the processing steps executed by the program. [Figure 6] A follow-up chart showing the processing steps executed by the program. [Figure 7] A follow-up chart showing the processing steps executed by the program. [Figure 8] A follow-up chart illustrating the integrated flow of this system. [Best Mode for Carrying Out the Invention]

[0020] Embodiments of the present invention will be described below with reference to the drawings.

[0021] Figure 1 shows the network configuration centered on the automated voice dialing system 2. It is connected to user terminals 1 (X, Y, etc., multiple terminals) and medical systems 3 within the hospital (e.g., electronic medical record system, accounting system, etc.) via telecommunication lines and communication networks such as the Internet / intranet / hospital LAN, enabling real-time collaboration. The network is configured so that system 2 can access medical systems 3 when it needs to refer to data within medical systems 3. User terminals 1 are not limited to mobile communication terminals; they may also be telephone and communication devices within the hospital (existing landlines, hospital PHS, etc.).

[0022] Figure 2 is a block diagram showing an example of the system configuration of the server 20 of this system 2 and other devices. This system 2 performs response control or display control with user terminals 10 (X,Y) connected by the server 20. The computer of the server 20 is equipped with a communication IF 22, an input / output IF 23, memory 25, storage 26, and a processor 29.

[0023] Communication IF22 is an interface for inputting and outputting signals so that the server 20 can communicate with external devices. Input / Output IF23 functions as an interface to an input device for receiving user input operations and an output device for outputting information to the user. Memory 25 is for temporarily storing programs and data processed by programs, and is a volatile memory such as DRAM. Storage 26 is a storage device for saving data, such as flash memory or an HDD. The processor 29 is hardware for executing the instruction set written in the program and consists of an arithmetic unit, registers, peripheral circuits, etc.

[0024] The user terminal 10 is communicably connected to the server 20 via the network 80, and has display means for visualizing the transmitted content transmitted by the server 20, and display means or notification means by which the transmitted content is made by a visible screen or voice / notification sound. The user terminal 10 is connected to the network 80 by communicating with communication devices such as a radio base station 81 compatible with various communication standards such as LTE, and other wireless LAN routers compatible with IEEE and wireless LAN standards. The user terminal 10 is a general-purpose terminal including a communication IF 12, an input device 13, an output device 14, a memory 15, a storage unit 16, and a processor 19. The user terminal 10 is a fixed terminal including a desktop type or laptop type PC, or a portable terminal such as a tablet or a smartphone.

[0025] The communication IF 12 is an interface for the user terminal 10 to communicate with the server 20 and input / output signals. The input device 13 is an input device (such as a microphone, keyboard, touch panel, touch pad, pointing device such as a mouse, etc.) for receiving an input operation from the user. The output device 14 is an output device (such as a display, speaker, etc.) for presenting information to the user. The memory 15 is for temporarily storing programs, data processed by programs, etc., and is, for example, a volatile memory such as DRAM. The storage unit 16 is a storage device for storing data, and is, for example, a flash memory or an HDD. The processor 19 is hardware for executing an instruction set described in a program, and is composed of an arithmetic unit, registers, peripheral circuits, etc.

[0026] FIG. 3 is a schematic diagram of the processing of the automatic voice transmission system, and shows the processing outlines of transmission processing A, response processing B including interactive response, and notification processing C performed by the server 20.

[0027] <A. Transmission Processing> The mechanism of the transmission processing A manages the process of transmitting the information input by the sender X to the receiver Y. Input Reception: The message input module inputs the information transmitted by X in text. Or it can be input by voice. Voice conversion: The voice / text conversion module converts text information into voice data. It is primarily assumed that the recipient will receive it as voice. When X conveys a voice message input by voice to Y, there is no need to convert the voice information into text information. Transmission: The automatic transmission module makes a phone call to transmit the voice message to Y based on the converted voice data. The transmission from X to Y is carried out via server 20, and efficient transmission can be expected particularly by using the following functions of notification management and response available time setting. Notification management: The notification module manages the notification process regarding the transmission status including the transmission status and receipt confirmation. Response available time setting: The function of setting the response available time of Y by the response time setting module is performed automatically or manually, and the automatic transmission module attempts to automatically transmit to Y according to the specified response available time. Automatic setting is useful for saving time at a busy site and refers to the information output from the electronic medical record or business application. Manual setting is flexible and reliable, but considering that it takes time to operate, it is preferably used as an auxiliary. The response available time setting may be performed in real time or may be automatically set each time according to the schedule and situation of Y on the same day.

[0028] <B. Response processing> The mechanism of response processing B is a process that manages the communication and response between X and Y by the response control module. That is, it manages the transmission by two-way communication between X and Y, including not only the transmission from X to Y, but also the transmission (reply) from Y to X and the transmission (re - reply) from X to Y. The response control module has a function of controlling the interactive response between X and Y using a generation AI or large - language model with the text message input by X as a prompt, and includes a program that converts the response into voice and outputs it with the information obtained by converting the message content by X and the reply content by Y into text as input. By utilizing the text message input by X as a prompt and having the large - language model (LLM) generate the response and instruction text of Y based on the prompt, an efficient interactive response between X and Y can be realized. Two-way telephone connection: If necessary during automatic dialing, X and Y will be directly connected. For example, if Y's scheduled time slot coincides with X's receiving call, or if X is available to answer the phone while Y is responding, the system will connect the calls based on the work status and timing of X and Y. Interactive Response: The response control module configures the agent (generates prompts for automated responses), generates questions and response requests based on the purpose and intent of the response, and retrieves responses (voice and text responses). Response processing: Receive Y's response and transmit it to X. The basic process for handling the interaction involves recording Y's response to the voice message as audio, and the voice / text conversion module converting the recorded audio into text data and providing Y's response to X as text. For X, receiving a text response to a text message from Y helps prevent misunderstandings and misinterpretations compared to audio communication. In particular, in medical settings, clear instructions are crucial, and text reduces the risk of misunderstanding. Furthermore, compared to audio, text can be viewed without being affected by ambient noise, making it easier for X to access information in any situation. Asynchronous communication between X and Y is also possible, as even if X cannot immediately check the response, it remains in text form and can be reviewed at a convenient time. To facilitate visual confirmation, visual effects that allow for quick grasp of key points at a glance are added, and particularly important parts are highlighted (color coding, bold text, etc.) to enable rapid comprehension. In addition, when a response is returned by voice, there are advantages such as smooth two-way communication, saving the time for reading text, and easily attracting X's attention on the spot. However, once voice is missed, it needs to be confirmed again. Voice cannot be confirmed at a glance like text, is difficult to hear in noisy places or when the communication state is poor, is prone to unclear pronunciation or misunderstanding by the listener, and information may become scattered in voice, making important parts difficult to understand. Also, in the case of voice messages, it is cumbersome to enter a response, and it is more convenient to record the response as it is, convert it to text, and return it to X. Whether to use voice for response or text for response, it is desirable to appropriately distinguish between the use of voice and text according to the environment and application. Or, transmit the message in text in a format that matches the desired types of X and Y, or make a phone call and communicate by voice.

[0029] <C, Notification Processing> Notification processing manages notifications related to transmission and response by the notification schedule module. Transmission Monitoring: Check whether the transmission to Y has been successful or whether a response from Y→X has been received. The notification module also performs a similar function. Count the remaining time until the deadline and send notifications according to the remaining time until the deadline. Response Schedule Management: Manage the deadline for responses and escalate notifications and remind when there is no response. Notification Based on Deadline: Gradually escalate using multiple means such as chat messages, SMS, voice notifications, etc. sequentially before the deadline. The deadline is set by the sender of the transmission / reply. When Deadline is Exceeded: Implement an emergency notification (such as continuous phone calls).

[0030] <Functional Configuration of Server 20> Figure 4 is a block diagram showing an example of the functional configuration of server 20. Server 20 is electrically connected by a bus to a communication means 220, an input device 230, a display device 240, a storage means 280, a control means 290, an input means 291, a display control means 292, a notification processing means 293, a user IF 294, and a response control means 295.

[0031] The communication means 220 performs modulation and demodulation processing for the server 20 to communicate with the user terminal 10, processes the signal calculated by the control means 290 for transmission, and transmits it to external devices and equipment. The communication means 220 processes the signal received from the outside and outputs it to the control means 290. In this way, the communication means 220 interprets commands or input content and provides them to each means, and also functions as an interface that interprets various display commands issued from the storage means 280 and performs output control.

[0032] The input device 230 is a device used by an administrator to input instructions or information to operate the server 20 as needed, and may be a keyboard, mouse, reader, or touch-sensitive device. The input device 230 also converts the instructions input by the administrator into electrical signals and outputs the electrical signals to the control means 290. The input device 230 also includes a receiving port that accepts electrical signals input from external input devices. The display device 240 is a display device such as an LCD or organic EL display 241 for presenting information to an administrator who operates the server 20 as needed. The display 241 can display data according to the control content of the control means 290 and can check the communication status between the server 20 and other external devices 10.

[0033] The storage means 280 is implemented by memory (RAM) 25 and storage 26 such as a disk device (floppy disk, hard disk, or magneto-optical disk, etc.) and stores data, programs, etc. used by the server 20. The storage means 280 stores the application program 282 of this system, as well as data for the work area 281, data storage area 283, and screen definition storage area 284.

[0034] The work area 281 is a region that is secured when the system is started and where various data input and output by the system are temporarily stored. The data storage area 283 is a region where data temporarily stored in the work area 281 is semi-permanently stored through write control when a save request is made. The screen definition area 284 is a region where screen definition information for various screens to be output and displayed on the in-hospital terminal 10 is stored in advance, and includes format information for screen settings to be displayed by the display control means 292.

[0035] The data storage area 283 includes DB283A for text-to-speech conversion, DB283B for interaction history, DB283C for templates, and DB293D for other purposes. <DB283A for text-to-speech conversion> This module stores data for text-to-speech and speech-to-text conversion. The stored data includes text-to-speech correspondence information (speech samples and dictionaries for conversion algorithms), phoneme data necessary for speech synthesis, speech recognition training data, and language models for natural language processing (NLP) and speech processing. It is related to the "Speech / Text Conversion Module." <Call History DB283B> It stores a history of sent and received messages and responses. The stored data includes the message content, X and Y ID information, the date and time the message was sent and received, the response content (voice and text), and its type (automated response, manual response). In relation to the "History Management Module," it is used for history referencing to support decision-making based on the content of the responses. <Template DB283C> It stores standardized responses and questions, enabling consistent and efficient responses. The stored data includes standardized messages used by X and Y (e.g., instructions, requests for confirmation), standardized phrases and expressions used in Y's responses, and template data for generating interactive responses. It is related to the "Interactive Question Module," "Answer Acquisition Module," and "Standardized Message Templates." <Other DB283D> This section stores system configuration information (e.g., response time, notification schedule), user information and access rights data, log information (e.g., electronic medical records and patient records), and data for analysis. It is related to the "Response Time Setting Module" and is used for a function that detects gaps between patient consultations based on electronic medical record log information. Each database plays the following roles and functions as a foundation supporting specific functions. Text-to-speech conversion database: for speech conversion and recognition processing. Response History Database: Manages the recording and referencing of response history. Template DB: Responses based on predefined templates. Other DBs: Used to store auxiliary data that supports system operation.

[0036] The control means 290 is configured such that various processes of the input means 291, display control means 292, notification processing means 293, user IF 294, and response control means 295 are implemented by a processor. The processor is one or more processors. At least one processor is typically a microprocessor such as a CPU (Central Processing Unit), but may be other types of processors such as a GPU (Graphics Processing Unit). At least one processor may be single-core or multi-core. Furthermore, at least one processor may be a broad-sense processor such as a hardware circuit that performs some or all of the processing (e.g., FPGA (Field-Programmable Gate Array) or ASIC (Application Specific Integrated Circuit)).

[0037] The display control means 292 visualizes the text-based communication content of this system and controls its display on the user terminal 10. When controlling the display on the user terminal 10, highlighting is performed with consideration for ease of understanding the communication content. Highlighting includes color coding, assigning different colors to each importance level or category (red for urgent information, blue for confirmation items), bold text to highlight particularly noteworthy parts or keywords (highlighting deadlines and important instructions), and font size adjustment to display headings and high-priority information in a larger font. The AI ​​analyzes the communication content and automatically sets appropriate highlighting according to priority and content.

[0038] The notification processing means 293 notifies the user (X or Y) of important information and the status of responses by transmitting information at an appropriate time and by appropriate means. It is responsible for the function of providing sequential notifications according to deadlines and continuous notifications after deadlines have been exceeded.

[0039] User interface 294 performs processing to implement functions such as receiving message input, operations, and commands from user terminals 10 and 30 via keyboard input or voice input, and creating a layout that is easy for the user to view based on the screen format information in the screen definition area 284.

[0040] The response control means 295 executes the following processing based on the various data in the data storage area 283. <Voice / Text Conversion Unit 295A> If Y responds verbally, it is converted into text data. The converted data is stored in the response history DB283B. The text entered by X is converted into audio data. The generated audio data is automatically sent to Y as a message. The speech / text conversion algorithm can be based on speech recognition (ASR) and text-to-speech (TTS) algorithms. <Transmission Control Unit 295B> It controls the entire transmission process from X to Y, ensuring that the input message is transmitted in the appropriate format (voice or text). Furthermore, if a response is received from Y, it is relayed to X (re-transmitted). The response history DB283B and template DB283C are referenced. <Interactive response generation unit 295C> Based on template DB283C, questions for Y are automatically generated. The system can process responses appropriately regardless of whether they are voice or text. To encourage a response, provide a prompt (e.g., "Please choose your next action based on these results"). Based on questions posed by X, Y is provided with interactive answer selection options. This assists healthcare professionals in making decisions based on patient information. <Answer acquisition section 295D> Retrieve a voice or text response from Y. Retrieve and record Y's response (e.g., "I understand," "It's been arranged," etc.). Adjust the settings so that Y can choose to respond via voice or text. The acquired response will be saved to the response history DB283B. <Template Generation and Management Department 295E> Create frequently used messages as affiliate templates. New pre-written phrases are added and updated, and past templates are optimized. Based on the input of X or Y, the system refers to template DB283C and selects and uses the appropriate template. For example, messages such as "Please check your test results" can be automatically generated depending on the situation. <History Management Department 295F> Records message sending and response content, as well as timestamps. Search past interaction history and present the necessary information. We analyze response history and use it to improve the system and support decision-making.

[0041] Each component of the present invention uses the following block mechanisms and memory areas, but other block mechanisms and memory areas may also be used; in general, it is implemented by a processor and memory. The "message input module" is a module for inputting information that X wants to convey in text format, and corresponds to input means 291. The "Voice / Text Conversion Module" is a module that converts input text into speech data, or speech data into text, and corresponds to the Voice / Text Conversion Unit 295A. The "automatic transmission module" is a module that automatically transmits voice messages based on the converted voice data, and corresponds to the transmission control unit 295B. The "notification module" is a module for X to confirm that the voice message has been successfully transmitted to Y, and corresponds to the notification processing means 293. The "Response Time Setting Module" is a module for automatically or manually setting the time period during which Y can receive (respond), and corresponds to the notification processing means 293 and other DB283D. It has the function of detecting Y's available response time by referring to information output from the electronic medical record or business application. The "function to reference" information from electronic medical records and business applications corresponds to DB283D and transmission control unit 295B. The "interconnection" of X and Y corresponds to the transmission control unit 295B and the communication means 220. • "Selection of response format" corresponds to the response acquisition unit 295D. The "Notification Schedule Module" is a module that provides sequential notifications using multiple notification methods based on a specified deadline, and if the deadline is exceeded, it provides notifications by making a series of telephone calls. This corresponds to the notification processing means 293. The "response control module" is a module that performs interactive responses between X and Y, and corresponds to the interactive response generation unit 295C. The interactive response generation unit 295C includes a template input module that templates the accumulated response content, an interactive question module that automatically generates interactive questions, and a response acquisition module for obtaining answers from X and Y. The "Template Input Module" corresponds to the template generation and management unit 295E. The "interactive question module" corresponds to the interactive response generation unit 295C. The "response acquisition module" corresponds to the response acquisition unit 295D. The "standard message template" is a standard message template used to unify the message content entered by X and Y, and the voice conversion module generates voice data based on this template. This corresponds to template DB283C and template generation / management unit 295E. The "history management module" is a module that records the content and date and time of transmission of voice messages sent to Y, and corresponds to the history management unit 295F and the response history DB 283B. The "Confirmation Message Generation Module" corresponds to the interactive response generation unit 295C. It generates a message to confirm with Y whether a specific treatment or action can be performed on the patient, along with information on the patient's length of hospital stay for a specific ward or area. Y can select and input either "good" or "bad," and if the answer is "bad," it includes a response option module for inputting the reason. The "response option module" corresponds to the response acquisition unit 295D. ·

[0042] In other words, the input and response acquisition functions are realized by the input means 291 and the response acquisition unit 295D, the notification and transmission functions are realized by the notification processing means 293 and the transmission control unit 295B, the voice / text conversion function is realized by the voice / text conversion unit 295A, the history management function is realized by the history management unit 295F, the template generation function is realized by the template generation and management unit 295E, and the interactive response generation function is realized by the interactive response generation unit 295C.

[0043] Figures 5 to 7 are flowcharts showing the processing procedures executed by the program of an automated voice call system according to an embodiment of the present invention. Figure 5 relates to the transmission process A, Figure 6 to the notification process C, and Figure 7 to the response process B subroutines.

[0044] In Figure 5, first, the system receives the message entered by X (step S501). User X enters the message via the input device 13 (keyboard or voice input), and if it is text, the voice / text conversion unit 295A converts it to speech after input. The voice / text conversion unit 295A uses a text-to-speech conversion algorithm (e.g., Tacotron, WaveNet). In step S502, it is determined whether the reception is complete. If it is complete, the transmission control unit 295B manages the received message and hands over processing to the appropriate module. Whether the message is complete can be determined by the passage of time, and if it is not completed even after the time limit has been exceeded, the process is terminated due to a timeout.

[0045] Next, the system automatically or manually sets Y's available time slots (step S503). Automatic setting involves the system automatically setting available time slots by detecting gaps between consultations and other tasks, based on Y's schedule and electronic medical record logs (such as consultation start / end times and entry / exit records). Manual setting allows Y to directly specify available time slots in advance using the system, or an administrator to set user times in bulk. The template generation and management unit 295E can also notify Y of a confirmation message when manual setting is performed (step S504).

[0046] Next, the system automatically dials Y based on the converted voice data (step S504). The notification processing means 293 adjusts the dialing timing based on the electronic medical record log and the response time setting, and the transmission control unit 295B monitors the dialing status and retries if an error occurs. With the above processing, this subroutine related to transmission process A is terminated.

[0047] In Figure 6, notification process C monitors for responses (checks for responses) by notification processing means 293 (step S601) and determines whether or not a response has been received (step S602). Notification processing means 293 supports writing the presence or absence of a response to the response history DB 283B and saves it as history. Notification processing means 293 then continues or temporarily suspends the response waiting period based on the response availability time setting.

[0048] While monitoring for responses (step S603) continues, it is determined whether the response was completed within the deadline (step S604), an alert is issued if it is close to or past the deadline (step S605), and the system continues waiting for a response (step S602). The notification processing means 293 proceeds in stages according to the remaining time until the deadline: notification by chat message → sending an SMS message → notification by phone call, and if the deadline is exceeded, it makes continuous alert notifications by phone calls.

[0049] The response monitoring status is stored in the history management unit 295F and other DB283D, along with deadline information, for use in schedule management. Furthermore, if there are no issues with the deadline and the response is completed, a completion record is made (step S606), and this subroutine terminates.

[0050] In Figure 7, response process B receives the message from X, and if it contains a request for a response, it receives the request for a response and records it as a trigger to start response process B (step S701). The response acquisition unit 295D analyzes whether or not there is a request for a response and sets the trigger.

[0051] Next, based on the content and status of communication from X, this system sets up an agent (interactive response support function) and generates prompts (question sentences and confirmation messages) as needed (step S702). The interactive response generation unit 295C automatically generates questions and confirmation messages using the response history DB 283B and template DB 283C, or generates prompts using pre-configured standard phrases using the template generation and management unit 295E.

[0052] Next, the response acquisition unit 295D selects and executes a response method according to the type of response (voice or text) specified by X (step S703). If the response types of X and Y do not match, the voice / text conversion unit 295A converts them to the selected response type.

[0053] Next, the response received from Y is recorded as part of the history (step S704). The response content (voice, text) and response timestamp are recorded in the response history DB283B and the history management unit 295F.

[0054] Next, a response is transmitted from Y to X (step S705). The transmission of the response by the transmission control unit 295B transmits the recorded response to X or Y.

[0055] Next, response history management is performed (step S706). All history related to the response (sent and received content, date and time, etc.) is recorded in the history DB under the direction of the history management unit 295F.

[0056] Next, a termination decision is made (step S707). If the necessary response processing is completed, the system determines that the process is finished and terminates this subroutine. If the system does not determine that the process is finished, for example, if Y's response is insufficient to the question or if part of the answer is missing, the interactive response generation unit 295C refers to the response history DB 283B and requests additional questions or reconfirmation for the incomplete response. In this case, processing resumes and the system returns to S702 (agent settings) or S703 (response method selection). If Y's response clearly contains a misunderstanding of the context or the question, the response acquisition unit 295D analyzes the response content and detects that it does not match the question content, and the interactive response generation unit 295C notifies that the response does not match the question and requests a revised response.

[0057] Specific examples of interactive responses are as follows: <X→Y> X uses this system to request Y's approval for a specific treatment or test. Example: Message to Y: "Please approve the CT scan for patient Z." "Approval is required to perform a CT scan on patient Z. Do you approve?" <Y→X> Y receives the notification and selects "Approve" or "Deny," and enters the reason for denial if necessary. Example: "Approval" → Proceed with CT scan. Enter "Denial" → "Low priority for testing". <X→Y> If Y denies it, X will request further confirmation or additional information based on the reasons given. Example: "A CT scan is needed for the diagnosis of patient Z. Please reconsider." <Y→X> Y will review the matter and submit their final decision. Examples: "I approve." "I deny. I am awaiting the results of further tests."

[0058] Figure 8 is a schematic flowchart that divides the integrated system behavior into three blocks: Process A (Transmission Process), Process B (Response Process), and Process C (Notification Process).

[0059] First, the transmission process of process A is the process that manages the sending of messages and requests from X to Y (step S801). When X requests "Please approve a CT scan for patient Z," this system checks Y's schedule and transmits the request at the appropriate time and in the appropriate manner. Specifically, the response time setting module obtains Y's available time from the electronic medical record or schedule management system, and if X's request is in text format, the speech / text conversion unit converts it into speech data.

[0060] Next, the response process in process B is a process that obtains a response from Y, confirms its content, and then performs additional processes as necessary (step S802). If the response from Y is not appropriate, the response acquisition unit determines in step S803 that the response is incomplete, returns the process to step S802, and the system asks further questions or clarifies the situation. If Y responds with "approval" or "denial," and denies, the system generates an additional question such as "Please tell me the reason." The response content is recorded in the response history DB, and if there is an insufficient response, the interactive response generation unit generates a further question by referring to past history and the template DB. Examples of further questions include "Please enter a specific reason" or "Please tell me the alternative examination if a CT scan is not necessary."

[0061] The notification process in process C performs interrupt processing depending on the situation, and executes re-notification or escalation (step S804). If the response is delayed or the situation changes, a notification is executed as interrupt processing before returning to process A or process B. For example, if the response deadline is approaching, the notification scheduling module automatically executes an emergency notification (call), and if there is still no response, an escalation notification is sent to the administrator or another attending physician. If the response is unclear, a follow-up question is sent to facilitate a response appropriate to the situation. Interrupt processing plays a role in supporting rapid decision-making.

[0062] Upon completion of all processes, a history update (new step S805) is performed, recording the entire process and terminating. During the history update, all notification content, response content, follow-up questions, and response history are saved in the response history DB. This allows for reference of past interactions and can be used in subsequent processes. The template generation and management department uses this history information to create new templates and improve existing ones.

[0063] <Specific example 1 related to "Generative AI or large-scale language models"> The mechanism by which generative AI and large-scale language models operate based on specific input data (messages entered by X) is defined by the following configuration of ad. a) Text input and prompt generation The system processes (prompts) the messages input by X into a format suitable for generative AI and large-scale language models. For example, it automatically supplements historical data and contextual information related to the input message, creating a state where the AI ​​model can easily generate the optimal response. By incorporating past patient history and additional information into the prompt, more accurate responses become possible. Furthermore, clearly stating "what you want to do (e.g., answer a question, suggest a specific action)" in the prompt improves the accuracy of the response. b) Prompt structure The prompt doesn't simply pass the input X to the AI, but structures it as follows: • Format: "You are a medical assistant. Please answer the following questions." X's input: "The patient's blood pressure is unstable. How should I respond?" • Supplementary information on patient history: "The patient's history data is as follows: Unstable blood pressure (last 3 measurements)..." Task description: "Please suggest appropriate actions based on the question." c) Response generation process The AI ​​model receives a prompt and generates a response appropriate to the task. The response is converted into text, audio, or other output formats. d) Optimizing responses Further review and revise the response (e.g., simplify technical terms, convert to concise expressions) and communicate it to Y. [Industrial applicability]

[0064] This invention is useful as an automated voice communication system that enables efficient and accurate decision-making support in medical settings and work environments requiring emergency response. It is also useful in supporting business efficiency in approval processes and task management between teams by providing accurate information transmission and flexible response management. In particular, in work environments where speed and accuracy are required, it contributes to improved efficiency and reliability through features such as recording response history and automatically generating follow-up questions. Furthermore, its industrial value lies in its ability to enable flexible responses tailored to different situations by utilizing diverse and wide-ranging communication methods. [Explanation of Symbols]

[0065] 1: User terminal 2: Automated voice dialing system 3: Medical system 10: User terminal (13: Input device, 14: Output device, 15: Memory, 16: Storage unit, 19: Processor) 20: Server (25: Memory, 26: Storage, 29: Processor) 80: Internet network, 81: Wireless base station

[0066] The program of the present invention can be installed in computer terminals having computer functions such as CPU, memory, and storage, as well as mobile devices such as smartphones, tablets, and wearable devices, digital home appliances such as smart TVs, smart speakers, and smart home appliances, recording media such as USB memory, SD cards, hard disk drives (HDDs) / solid state drives (SSDs), dedicated equipment and terminals such as POS terminals, vending machines, ATMs, and medical equipment, and game consoles (home and portable). When this program is installed in medical equipment, it may be linked with the hospital's electronic health record (EHR / EMR) system.

[0067] [Note B1] A program that runs on a computer having a processor and memory, It enables bidirectional communication between X and Y, A message input step for X to enter the information they want to convey as text, A speech / text conversion step that converts the input text into audio data, An automatic sending step that automatically sends a voice message based on the converted voice data, A notification step for X to confirm that the voice message has been delivered to Y, A program that includes and implements a process that automatically sends a text message entered by X as a voice message to Y. [Note B2] Includes a response time setting step to automatically or manually set the time period during which Y can receive (respond to), The program described in [Appendix B1] is characterized in that the automatic dialing step implements a process to automatically dial during the time period. [Note B3] The program described in [Appendix B2] is characterized in that the response time setting step detects Y's available response time by referring to information output from an electronic medical record or business application, and the automatic dialing step implements a process to automatically dial during the available response time. [Note B4] The program described in [Appendix B3] is characterized in that the automatic dialing step also dials X during the automatic dialing process, thereby enabling the interconnection of telephone communications between X and Y. [Note B5] Record Y's response to the voice message, The program described in [Appendix B1] is characterized in that the voice / text conversion step implements the process of converting the recorded voice into text data. [Note B6] The program described in [Appendix B1] is characterized by recording Y's response to a voice message as voice recording or as text input, and then executing a process that selects either a voice or text response as specified by X. [Note B7] It has a notification scheduling step for sequentially sending notifications using multiple notification methods based on a specified deadline, The program according to [Appendix B1], wherein the notification scheduling step implements a process that provides at least one of the following depending on the remaining time until the deadline: notification via chat message, sending an SMS message, or making a phone call. [Note B8] The program described in [Appendix B7] is characterized in that the notification scheduling step implements a process of making notifications by making a series of telephone calls when the deadline is exceeded. [Note B9] [Appendix B1] The program described further includes a response control step, The program is characterized by implementing a process in which the interactive interaction between X and Y is controlled by either a generative AI or a large-scale language model that uses the text message input by X as a prompt. [Note B10] The aforementioned response control step is: A template input step that stores template data, which is created by accumulating and templated response content, An interactive question step that automatically generates interactive questions based on the aforementioned template data, The program described in [Appendix B9] comprises a response acquisition step for obtaining a response from X and Y, and is characterized by implementing the processing of each step. [Note B11] The program described in [Appendix B9] is characterized in that the response control step uses a standard message template to unify the message content entered by X and Y, and the voice conversion step implements a process to generate voice data based on the template. [Note B12] The program described in [Appendix B9] is characterized in that the response control step implements the processing of a history management step that records the content and date and time of transmission of the voice message sent to Y. [Note B13] The program described in [Appendix B9] is characterized in that the response control step includes a confirmation message generation step to ask Y whether a specific treatment or action can be performed on the patient, along with information on the patient's length of hospital stay for a specific ward or area, and a response option step in which Y can select and input the option "good" or "bad," and if "bad," input the reason, and the processing of each step is realized.

[0068] [Note C1] A method obtained by operating a computer comprising a processor and memory, It enables bidirectional communication between X and Y, A message input step for X to enter the information they want to convey as text, A speech / text conversion step that converts the input text into audio data, An automatic sending step that automatically sends a voice message based on the converted voice data, A notification step for X to confirm that the voice message has been delivered to Y, A method that includes and automatically transmits a text message entered by X as a voice message to Y. [Note C2] Includes a response time setting step to automatically or manually set the time period during which Y can receive (respond to), The method according to [Appendix C1], wherein the automatic dialing step is characterized by dialing automatically during the time period. [Note C3] The method described in [Appendix C2], characterized in that the response time setting step detects Y's available response time by referring to information output from an electronic medical record or business application, and the automatic dialing step automatically dials during the available response time. [Note C4] The method described in [Appendix C3], characterized in that the automatic dialing step also dials X during the automatic dialing process, thereby enabling the interconnection of telephone communications between X and Y. [Note C5] Record Y's response to the voice message, The method described in [Appendix C1], wherein the voice / text conversion step is characterized by converting the recorded voice into text data. [Appendix C6] The method described in [Appendix C1], characterized by recording Y's response to a voice message as an audio recording or text input, and selecting either an audio or text response type specified by X. [Note C7] It has a notification scheduling step for sequentially sending notifications using multiple notification methods based on a specified deadline, The notification scheduling step is characterized by providing at least one of the following depending on the remaining time until the deadline: notification by chat message, sending an SMS message, or making a phone call. [Appendix C1] [Note C8] The notification scheduling step is characterized by making notifications by making a series of telephone calls when the deadline is exceeded, as described in [Appendix C7]. [Note C9] [Appendix C1] The method described further includes a response control step, The method is characterized in that the interactive interaction between X and Y is controlled by one or more of either a generative AI or a large-scale language model that uses the text message input by X as a prompt. [Note C10] The aforementioned response control step is: A template input step that stores template data, which is created by accumulating and templated response content, An interactive question step that automatically generates interactive questions based on the aforementioned template data, The method described in [Appendix C9], characterized by comprising a step for obtaining answers from X and Y. [Note C11] The method described in [Appendix C9], characterized in that the response control step uses a standard message template to unify the message content entered by X and Y, and the voice conversion step generates voice data based on the template. [Note C12] The method according to [Appendix C9], characterized in that the response control step includes a history management step that records the content and date and time of transmission of the voice message sent to Y. [Note C13] The method according to [Appendix C9], wherein the response control step comprises a confirmation message generation step to ask Y whether a specific treatment or action can be performed on the patient, along with information on the patient's length of hospital stay for a specific ward or area, and a response option step in which Y can select and input the options "good" or "bad," and if "bad," input the reason.

Claims

1. An automatic voice dialing system comprising a processor and memory, which dials voice to a recipient terminal, The aforementioned processor, A function to detect the recipient's available response time based on information about the recipient's medical treatment or work activities output from electronic medical records or business applications, A function that automatically makes a voice call during the aforementioned response time, An automated voice dialing system characterized by achieving this.

2. The automatic voice dialing system according to claim 1, characterized in that the function for detecting the responseable time detects the interval between medical treatment or work from a log including records of the start and end of medical treatment and entry and exit.

3. The automatic voice dialing system according to claim 1, characterized in that the recipient terminal includes a landline telephone or a mobile terminal.

4. The automatic voice transmission system according to any one of claims 1 to 3, characterized in that the voice is voice generated from text.

5. A method for transmitting voice to a recipient terminal, A detection step that detects the recipient's available response time by referring to information about the recipient's medical treatment or work activities output from an electronic medical record or business application, A voice transmission process that automatically transmits a voice message during the responseable time, A method for transmitting sound, characterized by including [a specific feature].

6. The voice transmission method according to claim 5, characterized in that the detection step detects the responseable time using a medical log or an entry / exit log.

7. The voice transmission method according to claim 5, characterized in that it does not transmit during times other than the aforementioned responseable time.

8. The voice transmission method according to any one of claims 5 to 7, characterized in that the transmission is performed as a voice message.

9. The voice transmission method according to any one of claims 5 to 7, characterized in that a call connection is established between the sender and the receiver when transmitting the aforementioned signal.

10. The voice transmission method according to any one of claims 5 to 7, further comprising a process that enables the sender to obtain information regarding the transmission result or whether or not a response was received.

Images