Office assistance system

Abstract

According to the present invention, a computer system executes: a countermeasure information generation step for inputting an image obtained by imaging a managed person and instructing AI to generate a health countermeasure to be output to the managed person in order to manage the health or improve the morale of the managed person; and a health countermeasure output step for outputting the health countermeasure generated by the AI to the managed person. In addition, the computer system performs a facial expression / voice acquisition step for acquiring a facial expression or voice of the managed person. In the countermeasure information generation step, the facial expression or the voice is acquired as a reaction of the managed person, which is obtained as a result of outputting a prescribed facial expression or voice so as to be seen or heard by the managed person, in the facial expression / voice acquisition step, any one of the facial expression or the voice is analyzed to identify the type of emotion from a plurality of types of emotions including any one of stress, fatigue, joy, sadness, anger, surprise, fear, or the like, and the AI is instructed to generate an appropriate health countermeasure for the managed person on the basis of the type of emotion.

Description

Business Support System 【0006】 【0001】 The present invention relates to a business support system. The present invention claims the priority of Japanese Patent Application No. 2024-210178 filed on December 3, 2024, and Japanese Patent Application No. 2025-094744 filed on June 6, 2025. For designated countries where incorporation by reference is permitted, the contents described in those applications are incorporated herein by reference. 【0002】 Patent Document 1 describes an information processing apparatus including an acquisition unit that acquires a machine learning model learned to estimate health information indicating a health state from biological information, and a mapping unit that maps the health information of a user to the latent space of the machine learning model, and an estimation unit that estimates ideal biological information, which is biological information corresponding to desired health information indicating a desired health state of the user, based on first health feature information corresponding to the health information of the user mapped to the latent space. 【0003】 Japanese Unexamined Patent Application Publication No. 2023-180124 【0004】 In the above technology, the health level of a user is mapped to the latent space of a pre-learned machine learning model M1. Thereby, based on the first health feature information corresponding to the health level of the user mapped to the latent space, by gradually changing the value of the latent variable in the latent space from the first health feature information, it becomes possible to generate the first health feature information corresponding to the health level of the user from the health level of the user. However, since it does not perform a specific action on the user for health management or morale improvement, a practical effect cannot be directly obtained. 【0005】 An object of the present invention is to provide a technology that performs a specific action on a user for health management or morale improvement. 【0006】The present invention includes several means for solving at least some of the above problems, but an example is as follows. One aspect of the present invention is a business support system executed by a computer system for supporting a business office where an administrator and an administrator who is managed by the administrator coexist, and includes a countermeasure information generation step in which a generating AI is input with video images of an administrator, and the AI ​​analyzes the administrator's walking pattern, body movements, and posture from observation points in the video image diagnosis, and uses the results of the analysis to generate health measures to be output to the administrator for health management or morale improvement, and a health measure output step in which the health measures generated by the generating AI are output to the administrator. 【0007】 Furthermore, a business support method implemented by a computer system for supporting a business establishment where an administrator and an administrator-managed person coexist, comprising: a countermeasure information generation step instructing a generating AI to input an image of an administrator, analyze the administrator's walking pattern, body movements, and posture, and use the results of the analysis to generate health measures to be output to the administrator for the purpose of health management or morale improvement; and a health measure output step in outputting the health measures generated by the generating AI to the administrator. 【0008】 Furthermore, a business support program that operates a computer system to support a business establishment where an administrator and an administrator-managed person coexist, includes a countermeasure information generation step instructing a generating AI to input an image of the person-managed person, analyze the person-managed person's walking pattern, body movements, and posture, and use the results of the analysis to generate health measures to be output to the person-managed person for health management or morale improvement; and a health measure output step instructing the generating AI to output the health measures generated by the generating AI to the person-managed person. 【0009】According to the present invention, it is possible to provide technology that enables specific interventions for users in order to manage their health or improve their morale. 【0010】 Other issues, configurations, and effects not mentioned above will be clarified by the following description of the embodiments. 【0011】 This is a diagram illustrating the configuration of a business support system according to this embodiment. This diagram shows other configuration examples of the business support system according to this embodiment. This is a diagram illustrating an example of the configuration of the server device. This is a diagram illustrating an example of the data structure of user information. This is a diagram illustrating an example of the data structure of attendance record information. This is a diagram illustrating an example of the data structure of detection information. This is a diagram illustrating an example of the data structure of countermeasure information. This is a diagram illustrating an example of the data structure of report information. This is a diagram illustrating an example of the hardware configuration of the server device. This is a diagram illustrating an example of the hardware configuration of the attendance record device. This is a diagram illustrating an example of the time stamping process flow. This is a diagram illustrating an example of the report creation process flow. This is a diagram illustrating an example of the countermeasure information display screen. This is a diagram illustrating an example of the report information display screen. 【0012】 Below, a business support system 1 to which an embodiment according to one aspect of the present invention is applied will be described with reference to the drawings. In the following embodiments, when necessary for convenience, the description will be divided into multiple sections or embodiments. Unless otherwise specified, these are not unrelated, and one is a modification, detail, or supplementary explanation of part or all of the other. 【0013】 Furthermore, in the following embodiments, when referring to the number of elements, etc. (including number, numerical value, quantity, range, etc.), unless specifically stated or clearly limited in principle to a particular number, it is not limited to that particular number, and may be greater than or less than that specific number. 【0014】 Furthermore, it goes without saying that in the following embodiments, the constituent elements (including elemental steps, etc.) are not necessarily essential, except in cases where they are specifically indicated or where they are clearly essential in principle. 【0015】Similarly, in the following embodiments, when referring to the shape, positional relationship, etc., of components, unless otherwise specifically stated or when it is clearly not the case in principle, it shall include those that substantially approximate or resemble such shapes, etc. The same applies to the numerical values ​​and ranges mentioned above. 【0016】 Furthermore, in all the drawings used to illustrate the embodiments, the same reference numerals are generally used for identical components, and repeated explanations of them are omitted. 【0017】 Figure 1A is a diagram illustrating the configuration of the business support system according to this embodiment. The business support system 1 is a system that supports business establishments where managers and those managed by the manager coexist. For example, a manager is a person in a position to manage employees, such as the head of the business establishment or the management team. Those managed are, for example, employees, part-time workers, and sometimes employees or contractors of business partners or subcontractors. The business support system 1 includes a server device 100, an attendance recording device 200, and a generation AI service 300. Those managed, such as employees, use the attendance recording device 200 to perform actions for managing working hours when they arrive at or leave work, such as clocking in or clocking out. The server device 100 communicates with the attendance recording device 200 and controls the operation of the attendance recording device 200. 【0018】 The Generative AI Service 300 is a service that provides the functions of so-called Generative AI, such as GPT and Gemini, via an API (Application Programming Interface) or the like. The Generative AI Service 300 obtains desired results such as text, images, audio, and video by giving commands (prompts) in natural language to the Generative AI. The Generative AI Service 300 may utilize publicly known services such as ChatGPT, which publish services on the internet. Furthermore, the Generative AI Service 300 may also be AGI (Artificial General Intelligence) or ASI (Artificial Superintelligence). 【0019】In this embodiment, when the generation AI service 300 receives instructions, for example via an API, it causes the generation AI to generate information and transmits the result to the source of the instructions as a return value of the API. At that time, the generation AI receives various detection information of the person being managed (mainly image information such as videos and still images, and audio information such as speech) detected by the attendance recording device 200, and outputs health measure information that should be output to the person being managed for health management or morale improvement, in accordance with instructions from the source (for example, health measures including providing positive messages or fun jokes to the person being managed). Alternatively, regardless of instructions from the source, the generation AI outputs health measure information according to predetermined settings. The generation AI has been trained through deep learning and machine learning to output health measure information that should be output to the person being managed for health management or morale improvement, using various detection information of the person being managed detected by the attendance recording device 200 as input information. 【0020】 Furthermore, the generation AI service 300, upon receiving instructions for report creation along with past health status and warning history, generates a report indicating the health status of the person being managed. The generation AI has been trained through deep learning and machine learning to output a report indicating the health status of the person being managed, using the various health statuses and warning history of the person being managed detected by the attendance record device 200 as input information. However, it is not limited to this, and the generation AI service 300 may also obtain health information by searching for it from other devices via the internet or other means at runtime. 【0021】The server device 100 responds to various requests from the attendance recording device 200 via a data transmission protocol. For example, when the server device 100 receives a request from the attendance recording device 200 to record attendance times, it identifies the person to be recorded and stores the information in attendance record information 122. The server device 100 also receives imaging data and sound data from the attendance recording device 200, identifies the person to be recorded, and stores the information in detection information 123. After storing the detection information 123, the server device 100 passes the past detection information of the person to the generating AI and instructs it to create countermeasure information. The server device then displays the created countermeasure information on the attendance recording device 200. Figure 1B shows another example of the configuration of the business support system according to this embodiment, where the business support system 1A includes the server device 100A, the attendance recording device 200A, the generating AI service 300A, and the user terminal 400A. Employees and other managed personnel use the attendance recording device 200A to perform tasks such as clocking in and out or entering the gate for work time management purposes. The server device 100A communicates with the attendance recording device 200A and controls its operation. The attendance recording device 200A exchanges information with each device, such as the imaging unit 210A, sound collection unit 220A, human body recognition unit 230A, and clocking unit 240A, either directly or through communication. The attendance recording device 200A may also be equipped with a backup system of the server device 100A's processing unit 110A and storage unit 120A, for example, to enable simplified operation in case communication with other devices via the data communication network 50 is not possible. Furthermore, the user terminal 400A communicates with other devices via the data communication network 50 and can connect to the attendance recording device 200A as appropriate even if communication with other devices via the data communication network 50 is not possible. This enables emergency responses in the event of system failures or other emergencies to be carried out from the user terminal 400A. Furthermore, the health measures output unit 260A can also be provided on the user terminal 400A. Additionally, the generated AI service 300A may encompass the functions and configuration of the server device 100, and may also encompass the attendance recording device 200A, or it may control the processing performed by each processing unit of the attendance recording device 200A.Furthermore, the attendance recording device 200A may be equipped with an AI processing circuit equivalent to that of the generation AI service 300A. In addition, the attendance recording device 200A can connect to expert AIs (e.g., legal consultation AI) or other business AIs (e.g., employee health management AI) via the data communication network 50. Alternatively, the generation AI service 300 may be included within the attendance recording device 200A. Also, the user terminal 400A and the attendance recording device 200A may be connected to communicate without going through a data communication network 50 such as the Internet. The data communication network 50 may be configured as an in-house LAN. Furthermore, the business support system 1A may be made operational even in the event of problems such as errors in the internet connection. In addition, it is possible to operate using only the in-house system's AI without using an external generation AI service, ignoring or excluding other expert AIs. 【0022】 Figure 2 shows an example of the configuration of a server device. The server device 100 comprises a processing unit 110, a storage unit 120, and a communication unit 130. The storage unit 120 includes user information 121, attendance record information 122, detection information 123, countermeasure information 124, and report information 125. The processing unit 110 includes a countermeasure information generation unit 111, an output generation unit 112, an attendance record unit 113, and a report creation unit 114. 【0023】 Figure 3 shows an example of the data structure of user information. User information 121 includes user ID 121a and identification feature information 121b. User ID 121a is information that identifies the user (employee, etc.) from other managed persons. Identification feature information 121b is a feature that distinguishes the managed person from other managed persons when identifying them. For example, it could be various types of information such as a facial image, appearance, or voice. Note that user ID 121a may be coded information such as a QR code (registered trademark) or a URL. 【0024】Figure 4 shows an example of the data structure for attendance record information. Attendance record information 122 includes user ID 122a, attendance date and time 122b, and departure date and time 122c. User ID 122a is information that identifies the user (employee, etc.) from other managed persons. Attendance date and time 122b is information that identifies the date and time when the managed person arrived at work. Departure date and time 122c is information that identifies the date and time when the managed person left work. In addition, although not shown in this example, attendance records may be recorded multiple times on the same day, such as for business trips within working hours from the start to the end of work. 【0025】 Figure 5 shows an example of the data structure of detection information. Detection information 123 includes user ID 123a, date and time 123b, and detection information 123c. User ID 123a is information that identifies the user (employee, etc.) from other managed persons. Date and time 123b is information that specifies the date and time when the managed person was detected. Detection information 123c is various detection information of the managed person (mainly image information such as videos and still images, and audio information such as speech) obtained when the managed person was detected. Although not shown in this example, it may be possible to record detection information multiple times on the same day. The detection information may also include images of the external eye part of the managed person's eye (white of the eye, eyelid, iris, etc.). 【0026】 Figure 6 shows an example of the data structure of the countermeasure information. The countermeasure information 124 includes a user ID 124a, a date and time 124b, and countermeasure information 124c. The user ID 124a is information that identifies the user, who is a person under management (employee, etc.), from other persons under management. The date and time 124b is information that identifies the date and time when the server device 100 obtained the person's health countermeasure information from the generation AI service 300. The countermeasure information 124c is the person's health countermeasure information obtained by the server device 100 from the generation AI service 300. Although not shown here, it is also possible to record the output of countermeasure information multiple times on the same day. 【0027】Figure 7 shows an example of the data structure of report information. The report information 125 includes a user ID 125a, an analysis period 125b, and report data 125c. The user ID 125a is information that identifies the user, who is a managed person (employee, etc.), from other managed people. The analysis period 125b is information that specifies the period to be analyzed in the report that analyzes the health information of the managed person. The report data 125c is report data about the health of the managed person obtained by the server device 100 from the generated AI service 300. 【0028】 The countermeasure information generation unit 111 is a processing unit that instructs the generation AI service 300 to input images of the person being managed (especially their face, gait, body movements, facial expressions, etc.) and to generate health measures to be output to the person being managed for the purpose of health management or morale improvement. The countermeasure information generation unit 111 acquires images of the person being managed from a camera installed in the attendance recorder 200. Although not shown here, it is also possible to generate countermeasure information multiple times on the same day. 【0029】 Furthermore, the countermeasure information generation unit 111 instructs the generation AI service 300 to analyze the facial expressions of the person being managed using an image of the person being managed, identify the type of facial expression from a plurality of types including stress, fatigue, joy, and sadness, and then determine the health or mental state of the person being managed based on the type of facial expression to generate health countermeasures. In addition, other emotions such as anger, surprise, fear, disgust, contempt, interest, confusion, relief, and expectation may also be added to health countermeasures if ethical issues are resolved. 【0030】In this embodiment, the emotional circle model (also known as Russell's circle model) is adopted as the type of emotion or facial expression. The emotional circle model is a dimension on which various emotions are arranged, with emotions represented by mutually independent two-dimensional axes (Arousal: degree of arousal and Valence: emotional valence (pleasant / unpleasant)). Various emotions such as happy, excited, tense, stressed, sad, calm, and relaxed are arranged on the circle of the emotional circle model. Therefore, in this embodiment, the generating AI service 300 identifies the type of emotion or facial expression based on Russell's circle model described above. However, it is not limited to this, and other types of emotions or facial expressions may be adopted. 【0031】 Furthermore, the countermeasure information generation unit 111 instructs the generation AI service 300 to analyze at least one of the voice characteristics (voice tone, volume, pitch, speed) of the voice spoken by the person being managed (e.g., greetings or introductions) to evaluate stress and emotional changes, and to generate health countermeasures using the results of the evaluation. The countermeasure information generation unit 111 acquires the voice spoken by the person being managed from a microphone installed in the attendance recording device 200. In addition, the countermeasure information generation unit 111 may instruct the generation AI service 300 to determine if the person is suffering from a physical ailment, such as back or knee pain, based on the image of the person being managed taken, including their gait and behavior leading up to the imaging device, for example, the movement of their left and right feet. Not limited to physical ailments, the countermeasure information generation unit 111 may also instruct the generation AI service 300 to determine if the person is experiencing psychological stress, such as worries, if they are moving their eyes around restlessly. To this end, for example, it is preferable to provide space so that the person being monitored walks a certain distance (for example, 5 meters, preferably around 8 meters) in front of the camera. Furthermore, it is even more desirable to image the person being monitored from the side, diagonally in front, and behind. 【0032】Furthermore, the countermeasure information generation unit 111 instructs the generation AI service 300 to analyze the person being managed by video image diagnosis based on walking patterns, movements, posture, etc., using the values ​​from the human body recognition sensor or captured images (video), and to generate health countermeasures using the results of the analysis. The countermeasure information generation unit 111 acquires the person being managed's body information from the human body recognition sensor installed in the attendance recording device 200. 【0033】 Furthermore, the countermeasure information generation unit 111 instructs the generation AI service 300 to analyze the posture of the person being managed using data from the motion capture system, and to generate health countermeasures, including messages that warn of prolonged poor posture and suggest improvements, using the results of the analysis. The countermeasure information generation unit 111 acquires the motion capture data of the person being managed from a camera installed in the attendance recording device 200. 【0034】 In video image diagnosis, the countermeasure information generation unit 111 can instruct the generation AI service 300 to diagnose at least one item described in the "Video Diagnosis Observation Point Summary" below, for example, and output an "Observation Point and Symptom / Discomfort Summary," and obtain the output results. By obtaining these observation points, specific examples, examples of operating phenomena, and predicted symptoms / discomforts, the generation AI service can be made to perform diagnoses with greater accuracy. 【0035】 Alternatively, when performing video image diagnosis, the countermeasure information generation unit 111 can input at least one item from the "Overview of Video Diagnosis Observation Points" and at least one item from the "Overview of Observation Points and Symptoms / Discomforts" into the generation AI service 300 to perform the diagnosis. By inputting these observation points, specific examples, examples of behavioral phenomena, and suspected symptoms / discomforts, the generation AI service can be made to perform the diagnosis with greater accuracy. 【0036】 【0037】 【0038】Here, for each point that can be observed in videos / images from walking and movements, examples of the inferred phenomena (behavioral characteristics) and possible symptoms and physical discomfort are summarized. However, it is also possible to add body parts, symptoms, and specific diseases from viewpoints such as "orthopedics, neurology, gastroenterology" and "respiratory medicine, cardiology". 【0039】 Further, the countermeasure information generation unit 111 instructs the generation AI service 300 to generate health countermeasures including providing positive messages or fun jokes to the managed person. 【0040】 Further, the countermeasure information generation unit 111 inputs the past health status and health countermeasure history of the managed person to the generation AI service 300, and for the managed person who has been the target of a warning in the past, instructs to generate health countermeasures including messages for a gentle greeting with careful consideration. 【0041】 Further, the countermeasure information generation unit 111 instructs the generation AI service 300 to display an image of the managed person with a negative expression changed to a bright expression for the managed person with a negative expression, output a positive message voice to stop them, and generate health countermeasures including information for attempting to implant an expression in the managed person. 【0042】 Further, the countermeasure information generation unit 111 analyzes changes in the expressions and postures of employees in response to greetings, predicts the possibility of business or financial irregularities, and instructs to generate countermeasures. 【0043】 The output generation unit 112 is a processing unit that outputs the health countermeasures generated by the generation AI service 300 to the managed person. 【0044】 When the attendance / leaving work recording unit 113 receives an instruction for attendance recording or leaving work recording from the attendance / leaving work recording device 200, it records the attendance date and time and leaving work date and time of the managed person. 【0045】 The report creation unit 114 instructs the generation AI service 300 to generate a report indicating the fluctuations in the health status of the managed person over a predetermined period. 【0046】The communication unit 130 communicates with the commuting record device 200 and the generation AI service 300 via a data communication network and the Internet or the like. 【0047】 Returning to the description of FIG. 1A, the commuting record device 200 includes an imaging unit 210, a sound collection unit 220, a human body recognition unit 230, a time-stamping unit 240, a communication unit 250, and a health measure output unit 260. The imaging unit 210 captures a moving image or a still image at a predetermined angle of view. For example, the imaging unit 210 captures the posture of the person being managed who operates the commuting record device 200 so that the expression and walking posture of the person can be captured. The imaging unit 210 may be able to capture the outer eye part (such as the white of the eye, eyelids, iris, etc.) of the person being managed and obtain an enlarged photograph. The sound collection unit 220 collects sound within a predetermined range. For example, the sound collection unit 220 collects the speech of the person being managed who operates the commuting record device 200. The human body recognition unit 230 captures a moving image or a still image at a predetermined angle of view. For example, the human body recognition unit 230 obtains the human body information of the person being managed using a human body recognition sensor. 【0048】 The time-stamping unit 240 identifies the date and time when the person being managed arrives at work and the date and time when the person being managed leaves work, and requests the server device 100 to record the information. Also, at the time of arrival at work or departure from work, the time-stamping unit 240 detects the information of the person being managed using the imaging unit 210, the sound collection unit 220, and the human body recognition unit 230. 【0049】 The communication unit 250 communicates with the server device 100 via a data communication network and the Internet or the like. 【0050】The health measures output unit 260 receives instructions from the server device 100, receives health measures and report data, and outputs it. If the output content includes voice output (greetings, verbal encouragement, etc.), the health measures output unit 260 outputs the voice; if it includes image display, the health measures output unit 260 displays the image (including still images and videos). For example, the health measures output unit 260 outputs voice messages such as greetings and everyday conversations with the person being managed (e.g., "It's a nice day today," "You worked hard until late yesterday"). It may not be limited to simple conversations, but may also include questions about the progress of work, or, after a holiday, questions about places visited during the holiday and whether they enjoyed themselves. In this case, a robot that walks and speaks, or a robot speaker that travels on rails, may accompany and output the conversation, with the health measures output unit 260 controlling its operation. 【0051】 Furthermore, instead of a walking path in front of the camera, the system may be equipped with moving walkways, escalators, elevators, or other mobility devices, and a robot may run alongside the user at a speed appropriate to their movement speed and engage in conversation. In addition, to make it easier for the managed individual to speak, privacy may be respected, and guidance may be provided to maintain a distance between the managed individual and those in front of them on moving walkways and escalators, and elevators may be limited to one person at a time. Furthermore, a system may be provided in which the managed individual is transported in a cart. These mobility devices may be equipped with imaging devices and conversation devices, and may also be equipped with biometric information acquisition devices that measure body temperature, pulse, blood pressure, blood alcohol concentration, and facial color. In addition, during activities within the company, the user may be encouraged to wear a wearable terminal that can measure, for example, body temperature, pulse, and blood pressure. This is particularly effective in high-temperature work, work at height, and hazardous work environments. The measurement results may also be collected in a timely manner on the server device 100. 【0052】Figure 8 shows an example of the hardware configuration of the server device 100. The server device 100 has a hardware configuration that can be realized using a so-called server device, workstation, personal computer, smartphone, or tablet terminal enclosure. The server device 100 includes a processor 101, memory 102, storage 103, communication device 104, and a bus connecting each device. 【0053】 The processor 101 is a computing device such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit). 【0054】 Memory 102 is a memory device such as RAM (Random Access Memory). 【0055】 The storage device 103 is a non-volatile storage device capable of storing digital information, such as a hard disk drive, solid state drive (SSD), or flash memory. 【0056】 The communication device 104 is a network interface card (NIC) or the like that communicates with other devices via the data communication network 50. 【0057】 The countermeasure information generation unit 111, output generation unit 112, attendance record unit 113, and report creation unit 114 of the server device 100 described above are implemented by a program that causes the processor 101 to perform processing. This program is stored in memory 102, storage 103, or a ROM device (not shown), and is loaded into memory 102 for execution and executed by the processor 101. 【0058】 Furthermore, the storage unit 120 of the server device 100, user information 121, attendance record information 122, detection information 123, countermeasure information 124, and report information 125 are realized by the memory 102 and storage 103. The above is an example of the hardware configuration of the server device 100. 【0059】The configuration of the server device 100 can be further classified into many components depending on the processing content. Alternatively, it can be classified so that each component performs even more processing. 【0060】 Furthermore, each processing unit (countermeasure information generation unit 111, output generation unit 112, attendance record unit 113, report creation unit 114) may be constructed using dedicated hardware (ASIC, GPU, etc.) to realize its respective function. Also, the processing of each processing unit may be executed on a single piece of hardware or on multiple pieces of hardware. The information processing device that realizes the generation AI service 300 is envisioned as a cloud service, but is not limited to this, and may have a hardware configuration that is basically the same as that of the server device 100. 【0061】 Figure 9 shows an example of the hardware configuration of the attendance recording device. The attendance recording device 200 also has basically the same hardware configuration as the server device 100. The attendance recording device 200 includes a processor 101, memory 102, storage 103, communication device 104, input / output device 205, camera 206, microphone 207, and a bus connecting each device. 【0062】 The input / output device 205 is a device that combines a display device such as a liquid crystal display or an organic EL display with various input devices such as a touch panel, hardware buttons, a keyboard, and a mouse. The time stamping unit 240 and the health management output unit 260 are realized by the input / output device 205, the processor 101, the memory 102, the storage 103, and the communication device 104. 【0063】 Camera 206 captures video or still images at a predetermined angle of view. The imaging unit 210 is implemented by camera 206, processor 101, memory 102, storage 103, and communication device 104. The human body recognition unit 230 is implemented by camera 206, processor 101, memory 102, storage 103, and communication device 104. 【0064】The microphone 207 collects sound within a predetermined range. The sound collection unit 220 is realized by the microphone 207, processor 101, memory 102, storage 103, and communication device 104. 【0065】 Furthermore, the processing of each component of the attendance recording device 200 may be performed on one piece of hardware or on multiple pieces of hardware. Also, the processing of each component of the attendance recording device 200 may be implemented by one program or by multiple programs. 【0066】 The above is an example of the hardware configuration of the attendance recording device 200. 【0067】 Each component of the attendance recording device 200 can be further classified into many more components depending on the processing content. Alternatively, each component can be classified to perform even more processing. 【0068】 Furthermore, each processing unit (imaging unit 210, sound collection unit 220, human body recognition unit 230, time stamping unit 240, communication unit 250, health countermeasure output unit 260) may be constructed using dedicated hardware (ASIC, GPU, etc.) to realize its respective function. Also, the processing of each processing unit may be executed on a single piece of hardware or on multiple pieces of hardware. 【0069】 Next, the operation of the business support system 1 in this embodiment will be described. 【0070】 Figure 10 shows an example of the time clocking process flow. When the attendance recording device 200 is activated, it enters a standby state and starts when the person being managed makes an operation to the attendance recording device 200. 【0071】 First, the time stamping unit 240 of the attendance recording device 200 receives user ID information read from the managed person's ID card, such as a magnetic card or IC card, and an attendance / departure classification indicating whether the managed person is recording their attendance or departure (step S001). 【0072】Then, the attendance recording device 200 performs imaging and sound collection (step S002). Specifically, the imaging unit 210, sound collection unit 220, and human body recognition unit 230 acquire video or still images including facial expressions, audio data of speech, and human body information of the person being managed, respectively. However, it is not limited to this, and the imaging unit 210 and human body recognition unit 230 may also treat the walking and standing postures of the person being managed as they approach the attendance recording device 200 as human body information by taking pictures of them in advance. 【0073】 Then, the time stamping unit 240 of the attendance recording device 200 transmits the attendance record to the server device 100 (step S003). Specifically, the time stamping unit 240 transmits the user ID and the attendance category to the server device 100. 【0074】 The attendance record unit 113 of the server device 100 authenticates and stores the information in the attendance record (step S004). Specifically, when the attendance record unit 113 receives a user ID, it authenticates it and stores the attendance date and time or departure date and time in the attendance record information 122 according to the attendance category. 【0075】 Then, the time stamping unit 240 of the attendance recording device 200 transmits the detection information to the server device 100 (step S005). Specifically, the time stamping unit 240 transmits the user ID and the detection information detected in step S002 to the server device 100. 【0076】 The attendance record unit 113 of the server device 100 authenticates and stores the detection information in the detection information 123 (step S006). Specifically, when the attendance record unit 113 receives a user ID, it authenticates and stores the detection information in the detection information 123 for each person managed. 【0077】Then, the countermeasure information generation unit 111 instructs the generation of countermeasure information to be displayed (step S007). Specifically, the countermeasure information generation unit 111 extracts detection information of the person under management from the detection information 123 for a predetermined period in the past (for example, the most recent month), transmits the detection information of the person under management and the past detection information of the person under management to the generation AI service 300, and instructs the generation of countermeasure information to be displayed. The instruction to generate countermeasure information is based on the instruction to create health countermeasure information to be output to the person under management, but may also include specifying a type of facial expression from a plurality of types of facial expressions that include stress, fatigue, joy, or sadness, and instructing the system to determine the health or mental state of the person under management based on the type of facial expression and generate health countermeasures. 【0078】 Furthermore, instructions for generating countermeasures information may also involve analyzing at least one of the following aspects of the managed person's speech: voice condition (tone, volume, pitch, speed) and footstep patterns, to evaluate stress and emotional changes, and then using the results of this evaluation to generate health countermeasures. Alternatively, instructions for generating countermeasures information may involve presenting the managed person with facial expressions or voices generated from images or videos, analyzing the results to obtain either the managed person's facial expression or the facial expression of their response, identifying the type of facial expression from a range of expressions including stress, fatigue, joy, sadness, anger, surprise, fear, disgust, contempt, interest, confusion, relief, and anticipation, evaluating emotional changes, and then using the results of this evaluation to generate the aforementioned health countermeasures. In reality, facial expressions, gait, body movements, and voice production are easily influenced by individual differences, regional variations, and culture, making it difficult to judge them using the same criteria as others. Therefore, regularly collecting information on an individual managed person basis and analyzing their emotional expressions will improve accuracy. Therefore, it is desirable that the countermeasure information generation unit 111 analyzes and accumulates emotional fluctuations as data. Alternatively, the generation AI service 300 may be entrusted with the analysis and accumulation of this data. 【0079】Furthermore, instructions for creating countermeasures information may also include an instruction to analyze the gait pattern or posture of the person being managed using values ​​from a human body recognition sensor, and to generate health measures using the results of that analysis. Additionally, instructions for creating countermeasures information may also include an instruction to analyze the physical and mental health status of the person being managed using photographs of the external eye area (white of the eye, eyelid, iris, etc.), and to generate health measures using the results of that analysis. 【0080】 Furthermore, instructions for creating countermeasures information may also include instructions to analyze the posture of the person being managed using data from the motion capture system, and to generate health measures that include messages warning of prolonged poor posture and suggestions for improvement using the results of that analysis. 【0081】 Furthermore, instructions for creating countermeasures information may also include instructions for generating health measures that include providing positive messages or lighthearted jokes to those being managed. 【0082】 Furthermore, instructions for creating countermeasure information may also include instructions to generate health measures that include thoughtful and considerate messages for managers who have previously been subject to warnings, based on their past health status and history of health measures. 【0083】 Furthermore, instructions for creating countermeasures information may also include instructions to generate health measures that include information to attempt to instill positive expressions in the person being managed, such as displaying an image of the person with a brighter expression to those with negative expressions, outputting positive audio messages that speak to the person, or making them stop and stand still. 【0084】 Furthermore, instructions to create countermeasures information may also include instructing employees to analyze changes in their facial expressions and posture in response to being approached, predict the possibility of work-related or financial misconduct, and generate countermeasures based on that analysis. 【0085】Furthermore, instructions for creating countermeasure information can also involve instructing the conversation to focus on topics such as work or household finances. Additionally, instructions for creating countermeasure information can include replacing negative expressions in those conversations with positive ones and encouraging confirmation, thereby easing the mood or making the conversation more cheerful. Moreover, instead of forcing a cheerful tone on negative topics, consciously following the flow of "empathy → positive perspective → encouragement / acknowledgment → future-oriented / positive conclusion" will result in a more natural and uplifting response.Therefore, instructions for creating countermeasure information may include the following points, advice, specific replacement examples, and phrase points: 1. Start with empathy. Before jumping straight into positive transformation, acknowledge the other person's feelings with phrases like "That must have been tough" or "That must have been difficult," to build trust. 2. Subtly focus on the good points and progress. Even with negative facts, pick out "efforts," "lessons learned," and "positive elements." 3. Add words that specifically acknowledge the other person's presence and efforts. For example, "Thanks to you," "It wouldn't have been possible without you," or "I appreciate your help." 4. Think about and support the future together. Conclude with a phrase that leaves a positive impression of looking ahead, such as "We're just one step away from here," "Today will surely be the day," or "Let's take it easy." 【0086】Specific Replacement Examples Example 1: When work wasn't finished Employee: "I couldn't finish my work yesterday, and it took me until late, so I'm exhausted." Manager (Example): "You really did a great job. I think it's amazing that you were able to tackle such a difficult task! Are you almost finished? Please don't push yourself too hard today. If you have any problems, please don't hesitate to let me know." Example 2: When injured Employee: "I fell down the stairs at the station yesterday and scraped my right leg, so it hurts." Manager (Example): "That must have been terrible. But I'm really glad it wasn't a serious injury. Please walk carefully today. It will serve as a reminder to everyone! Thank you for letting me know." Example 3: When lacking motivation / feeling tired Employee: "Lately, I just haven't been feeling motivated..." Manager (Example): "Everyone has days like that. But I think it's important to proceed at your own pace without pushing yourself too hard. Let's start with what you can do at your own pace. Thank you for always working so hard!" The above, when work is finished Employee: "I somehow managed to finish my work yesterday." Manager (Example): "Great job! Your ability to see things through to the end is definitely one of your strengths, [Name]." "This is the result of your hard work, even late into the night. You must feel relieved now that you've reached a milestone." "I'm so glad it's all finished successfully. Take some time to relax today before moving on to the next step." "You did a great job completing such a demanding task. It's very encouraging to see the results!" Key points of the phrases: Specific praise for "what was accomplished" Appreciation for "hard work" and "the process" Expressing expectations for the future and suggesting time to refresh 【0087】Also, in order to make it possible to naturally internalize the words spoken as one's own, an instruction may be given to the generative AI to rephrase them as if the employee himself / herself were speaking. That is, by making the person being managed have the feeling as if they had spoken positive words (rewriting of self-talk), the effect of fixing in memory and changing the mood can be enhanced. Therefore, the instruction for creating countermeasure information for the generative AI may include the following content. Ingenuity / Technique: Express as "your own words" → When the AI returns, rephrase it in the first person as if the employee himself / herself were speaking Keep it short and concise → Not redundant, making the person think "I could say this myself" End with a confirmation or encouragement that naturally induces the feeling of "Surely I can do it" or "Okay, let's do our best" Actual conversation examples Employee (negative comment): "I have to finish this work by the end of this week. I'm in trouble." AI (self-talk rewritten version): "Upon reflection, there are still three days left this week, and if I consult, it will surely be okay. Let's start with what I can do!" Employee (negative comment): "I'm having trouble getting motivated recently..." AI (self-talk rewritten version): "There are times like this. Without forcing myself, let's do what I can at my own pace today. Let's move forward step by step without blaming myself." Employee (negative comment): "I couldn't finish my work yesterday. I'm so tired." AI (self-talk rewritten version): "I really worked hard yesterday. Since there's just a little more left, let's also move forward today without forcing myself. I'm not bad for having come this far!" Implementation image The AI not only addresses the employee in the second person but also reconstructs positive words as first-person self-talk to smoothly promote "self-talk" → The employee naturally incorporates a positive way of speaking into themselves Continuing this will also have a positive effect on the employee's own "thinking habits" and self-image Additional advice Example of guidance to employees: Tell them something like "Please read the words returned by the AI again in your head. It's okay to read them out softly in a low voice." This will increase the fixing effect! 1. Self-reflection style (self-introspection style) arrangement Self-reflection is a method of looking back on one's own actions, emotions, and events and leading them to "awareness" and "positive interpretations". Specific example Employee (negative comment) "I have to finish this work by the end of this week.: "Oh no." AI (Self-Reflection Transformation) "I just realized I'm panicking. But I still have three days, and if I can find a better way to do it, I might be able to get through it. Experiences like this are what help me grow." Employee (Negative Statement) "I'm having trouble finding motivation lately..." AI (Self-Reflection Transformation) "There are times when you just don't feel motivated. There's no need to blame yourself too much for that. Just start with something small, and it'll be okay if you gradually get back into the swing of things." Employee (Negative Statement) "I couldn't finish my work yesterday. I'm so tired." AI (Self-Reflection Transformation) "I really did my best yesterday. I didn't finish everything, but being able to do this much is proof of my effort. I'll try to make a little more progress today." In addition, the instructions for creating countermeasures information for the generated AI may also include the following: Other "self-transforming" techniques and affirmations (positive self-declarations) By voicing encouragement and affirmation to yourself, such as "I can do it" and "I've overcome challenges before," you can form a positive self-image. Example: "I've overcome every difficulty in the past. I'm sure I can overcome this one too." ・Reframing (changing the framework) Change your perspective on things and reframe the same facts in a positive light. Example: "I'm behind on my work" → "There's still an opportunity to find areas for improvement." ・Solution-focused thinking Instead of focusing on the problem, focus on "what should happen next" and "what can be done." Example: "How can I make this situation even a little bit better?" "What should I start with, the first step?" ・Gratitude & Mindfulness Express gratitude for even the small things that exist right now, saying "thank you" and "that was good," and acknowledge your feelings and efforts in the "here and now." Example: "Thank you to myself for being able to come to work on time today." "I'm grateful that my colleague helped me." 【0088】Furthermore, instructions for generating countermeasures information for the AI ​​may include instructions to guide the user toward correct posture and walking by showing correct posture when posture or walking is poor, and overlaying images of correct posture and walking onto the user's image. For example, instructions may include slowly moving the image from the current posture to the correct walking style, or indicating areas for improvement with lines or arrows. In addition, instructions may include showing examples of how posture and walking will eventually worsen if they continue as they are. 【0089】 The generation AI service 300 creates countermeasure information in response to the generation instruction and transmits the countermeasure information to the server device 100 (step S008). 【0090】 Then, the countermeasure information generation unit 111 stores the countermeasure information transmitted from the generation AI service 300 in the countermeasure information 124 (step S009). 【0091】 Then, the output generation unit 112 transmits the countermeasure information to the attendance record device 200 (step S010). 【0092】 The health countermeasure output unit 260 displays the countermeasure information (step S011). 【0093】 The above is an example of a time clock processing flow. The time clock processing allows for the display of health measures that should be output to the employee in question for health management or morale improvement, enabling specific actions to be taken. 【0094】 Figure 11 shows an example of the report creation process flow. The report creation process starts when the server device 100 receives an operation from an administrator (such as the head of the business or management, or any person in a position to manage employees). 【0095】First, the report creation unit 114 accepts the specification of the reporting period and the target individuals (step S401). Then, the report creation unit 114 acquires attendance records, detection information, and countermeasure information for each target individual for the reporting period (step S402). Specifically, the report creation unit 114 extracts attendance records for the reporting period from the attendance record information 122 in the storage unit 120, detection information from the detection information 123 in the storage unit 120, and countermeasure information from the countermeasure information 124 in the storage unit 120, as RAG information. 【0096】 Then, the report creation unit 114 prompts the generation AI service 300 to create a report (step S403). Specifically, for each subject, the report creation unit 114 analyzes the attendance records, detection information, and countermeasure information for the reporting period extracted as RAG information, and uses the results of the analysis to create a prompt instructing the generation AI service 300 to create a report, and passes it to the generation AI service 300. Then, the report creation unit 114 receives the report data from the generation AI service 300 (step S404). Then, the report creation unit 114 stores it in the report information 125 (step S405). 【0097】 The above is an example of the report generation process flow. This report generation process allows for an objective analysis of the health and mental state of the person being managed, based on monitoring data over a specified period. 【0098】 Figure 12 shows an example of a countermeasure information display screen. The countermeasure information display screen 700 is the screen on which countermeasure information is displayed in step S011 of the time clock process. The countermeasure information display screen 700 displays an image 701 of the person under management whose negative expression in the captured image has been changed to a brighter expression. In addition, the countermeasure information display screen 700 displays a positive message 702 for the person under management. 【0099】 Figure 13 shows an example of a report information display screen. The report information display screen 800 is an example of a screen used to display the stored report information in step S405 of the report creation process. 【0100】 The report information display screen 800 includes a user ID display area 801, an image of the person being managed 802, an analysis graph 810, and a message display 820. The analysis graph 810 uses time on the horizontal axis 811 and a measured quantity on the vertical axis 812, plotting, for example, an index of mental health 813, the number of abnormal movements 814, an index of emotional information 815, and the number of poor postures 816, visualizing the trends of change. The message display 820 describes warning information and the analysis results of health trends in natural language. 【0101】 The above describes the Business Support System 1. As demonstrated in the above embodiment, the Business Support System 1 allows for specific actions to be taken with users for health management or morale improvement. However, it is important to ensure transparency and consent regarding the means, timing, and content of acquiring information on those under management, to protect privacy, to conduct regular health checkups, to establish a mental health support desk (creating an environment where employees can consult if they are experiencing stress or mental health problems), and to formulate ethical guidelines (implemented with transparency for employees). Furthermore, it is considered important to raise awareness throughout the organization by implementing education and training (education and training on ethics and compliance for employees). In light of these points, it is desirable to adopt programs, support systems, and operating methods that combine these functions. In addition, it may be beneficial to connect with other expert AIs and exchange information. 【0102】Furthermore, emotions, states, and facial expressions at that time can be associated as follows. Therefore, the countermeasure information generation unit 111 can instruct the generation AI service 300 to identify emotions from the facial expressions contained in the captured images according to the following correspondence. Joy: Expresses happiness or satisfaction: smile, sparkling eyes Sadness: Shows disappointment or loss: corners of the mouth turn down, eyes well up with tears Anger: Expresses irritation or hostility: frown lines, pursed lips Surprise: Reaction to an unexpected event: eyes wide open, mouth open Fear: Reaction to danger or threat: eyes wide open, eyebrows raised Disgust: Shows discomfort or rejection: nose frown, mouth twisted Contempt: Expresses a feeling of looking down on others: one corner of the mouth turned up Interest / Attention: Shows attention: eyes wide open, eyebrows slightly raised Confusion: A reaction to an incomprehensible situation: furrowed brows, slightly open mouth. Relief: A state of relaxation, with relaxed facial muscles and a calm expression. Anticipation: Looking forward to something good: sparkling eyes and a slight smile. Fatigue: A state of exhaustion, with half-closed eyes and downturned corners of the mouth. Doubt: Showing suspicion about something, with one eyebrow raised and lips pursed. Satisfaction: A feeling of contentment, with a slight smile and relaxed eyes. Tension: A state of feeling stressed or pressured, with rigid facial muscles and sharp eyes. 【0103】 Furthermore, emotions and voices are associated as follows. Therefore, the countermeasure information generation unit 111 can instruct the generation AI service 300 to identify emotions from the acquired voice according to the following correspondence. Emotion: Voice characteristics Joy: Bright, high-pitched tone, rhythmic way of speaking Sadness: Low, slow tone, weaker voice Anger: Strong voice, fast way of speaking, voice may tremble Surprise: Voice suddenly rises in pitch, words break Fear: Voice trembles, fast way of speaking Nervousness: Voice becomes stiff, words get stuck, irregular speaking speed Excitement: High tone, fast way of speaking, louder volume Boredom: Monotonous tone, slow way of speaking, little intonation Confidence: Clear pronunciation, stable tone and rhythm Anxiety: Voice trembles, words tend to break, fast way of speaking Contempt: Sarcastic tone, cold-sounding voice 【0104】Furthermore, emotions and walking patterns can be correlated as follows. Therefore, the countermeasure information generation unit 111 can instruct the generation AI service 300 to identify emotions from the acquired walking patterns according to the following correspondence: Joy: light and rhythmic walking, skipping movements, wide strides. Sadness: slow walking, drooping shoulders. Anger: A heavy, downcast gaze. Anger: A strong, fast walk, loud footsteps, and large arm swings. Surprise: A momentary pause in walking, irregular movements, and sudden stops. Fear: Small strides, cautious movements, and occasional glances. Tension: Awkward walk, stiff movements, raised shoulders, and limited arm swing. Excitement: A fast walk, lively movements, energetic, and long strides. Boredom: A slow walk, lifeless movements, unfocused gaze, and glances around. Confidence: A straight back, firm walk, steady strides, and forward-looking gaze. Anxiety: Irregular walk, restless movements, frequent stops, and glances around. Contempt: A slow walk, condescending movements, and occasional stops to look down at the surroundings. 【0105】 Furthermore, emotions and "body movements" are associated as follows. Therefore, the countermeasure information generation unit 111 can instruct the generation AI service 300 to identify emotions from the "body movements" of the person being managed, which are identified from the acquired images, according to the following correspondence. Joy: Smiling, jumping, clapping hands. Light and cheerful movements, energetic impression Sadness: Leaning forward, slumping shoulders. Heavy and depressed impression, slow movements Anger: Waving hands forcefully, frowning. Intense movements, tense impression Surprise: Eyes wide open, mouth open. Sudden movements, instantaneous reaction Fear: Tensing up, looking around cautiously. Defensive posture, cautious movements Tension: Rubbing hands together, shaking feet. Awkward movements, unstable impression Excitement: Waving hands widely, moving quickly. Lively and energetic movements Boredom: Crossing legs, averting gaze. Indifferent impression, monotonous movements Confidence: Straightening back, firm gait. A dignified impression, steady movements. Anxiety: frequently looking around, rubbing hands together. Restless movements, a wary impression. Contempt: narrowing eyes, raising one eyebrow. A cold impression, slow movements. 【0106】Furthermore, "physical condition" and "characteristics" are associated as follows. Therefore, the countermeasure information generation unit 111 can instruct the generation AI service 300 to identify the physical condition from the "characteristics" of the person being managed, according to the following correspondence. Physical condition: Characteristics Fatigue: No facial expression = eyes half-closed Posture = rounded back, drooping shoulders Movement = slow movements, shorter strides Voice tone = weak voice, increased sighing Energy = decreased vitality, easily fatigued High stress: Facial expression = furrowed brow, narrowed eyes Posture = raised shoulders, clenched jaw Movement = shaking feet, snapping knuckles Voice tone = shallow, rapid breathing, increased sighing Energy = restless movements, wary impression Joy: Facial expression = smiling, bright eyes Posture = upright, relaxed posture Movement = light, rhythmic movements, lively movements Voice tone = bright, clear voice, frequent laughter Energy = energetic, actively acting Sadness: Facial expression = corners of mouth drooping, watery eyes Posture = hunched over, drooping shoulders Movement = slow movements, decreased activity Voice tone = weak, low voice, reduced speaking frequency Energy = decreased vitality, easily fatigued 【0107】 Furthermore, instructions for creating countermeasures information for the generating AI may include instructions to convert the responses into a style that "sounds like a personal tweet" or "sounds like a diary or memo." Instructions for creating countermeasures information for the generating AI may also include instructions to output questions that encourage deeper thinking, such as "How did you reconsider this time?" In addition, the above-mentioned conversations may be generated even during activities within the company, and server terminals may be installed in the company to communicate about the progress of work, opinions, complaints, and grievances through conversation or text, or communication may be conducted using APIs etc. on the managed person's own mobile phone, PC, or wearable device. 【0108】Furthermore, regarding the Business Support System 1, we will describe an embodiment that focuses on the walking patterns and body movements of the person being managed. The Business Support System 1 acquires image and voice data of the person being managed (for example, an employee), and based on this data, it analyzes various physical and emotional characteristics such as walking patterns, body movements, posture, arm swing, balance of both hands, facial expressions, tone of voice, and even footsteps produced while walking, thereby comprehensively evaluating the health status of the person being managed. This makes it possible to capture early signs that tend to be overlooked in conventional health management and to take early preventive measures. 【0109】 For example, if analyzing an employee's posture from images reveals a hunched posture, rounded shoulders, or shoulder asymmetry, it may indicate modern occupational ailments such as stiff shoulders, lower back pain, straight neck, or VDT syndrome (Visual Display Terminal syndrome) caused by prolonged computer work. These signs can also serve as indicators of excessive strain on the musculoskeletal system. 【0110】 Furthermore, if imbalances in left-right balance, limping, or sluggishness are observed in walking patterns acquired from images and audio, it may indicate a decline in motor function, physical fatigue, or an early sign of neuromuscular disease. In particular, the Business Support System 1 acquires footsteps (walking sounds) as audio and analyzes characteristics such as intervals, rhythm, variations in intensity, and left-right differences in the footsteps, making it possible to detect lower limb motor disorders, joint pain, and gait unsteadiness that are difficult to judge visually. This makes it applicable to assessing the risk of falls and supporting rehabilitation for elderly employees and those in the recovery phase. 【0111】 Furthermore, by analyzing the speech-related components of the audio data, abnormalities in voice tone, speech speed, and pauses during speech can be used as mental health indicators, such as mental stress, sleep deprivation, and decreased concentration. This makes it possible to detect depressive symptoms and burnout early. 【0112】Thus, multifaceted analysis using image and audio data enables more objective and continuous monitoring of health management, which previously relied on regular checkups and self-reporting. In particular, this invention is extremely useful in that it can detect symptoms such as cervical spondylosis, straight neck, VDT syndrome, and chronic shoulder and lower back strain caused by PC work, which tend to become apparent in modern workplace environments, at an early stage, and propose and implement appropriate health measures. 【0113】 The generating AI used in the business support system 1 is configured to process multimodal inputs and comprehensively analyzes image data, audio data, and associated metadata (e.g., identification information of the person being managed, time of day, type of work, etc.). The generating AI receives input such as the following: Image data of the person being managed while walking or working (video or still images including the whole body), audio data of the person being managed (speech sounds, footsteps, coughing, etc.), metadata such as the date and time, location and duration of acquisition of the image and audio, and past analysis history for the same person being managed. This information is used for learning and inference in the AI ​​model for health status assessment and generation of countermeasures. 【0114】Furthermore, in this embodiment, prompts (instructions) are used to clearly specify the analysis purpose and output format in order to guide and optimize the analysis by the generating AI. The following is an example. <Prompt Example 1: Evaluation of Walking> "The following video shows employee A walking. Analyze the walking pattern, balance between the left and right feet, arm swing, and postural stability. If any physical ailments or risks are suspected, generate health measures including suggestions for improvement." <Prompt Example 2: Posture Evaluation and Detection of PC Work-Related Illnesses> "This image shows the person under management working at their desk. Evaluate whether there is any forward head jutting, hunchback, or rounded shoulders. If there is a possibility of straight neck or VDT syndrome, qualitatively evaluate the degree and suggest appropriate stretches and guidelines for break frequency as countermeasures." <Prompt Example 3: Evaluation of Voice Tone and Speech Content> "The following audio is the speech of the person under management while giving a work report. Analyze the voice tone, intonation, speaking speed, and presence or absence of hesitation. Evaluate whether there are any signs of mental stress or fatigue, and suggest refreshing methods and mental health measures as needed." <Prompt Example 4: Physical Function Evaluation through Footstep Analysis> "The following audio data is a recording of footsteps made while the person being monitored was walking. Pay attention to the interval between the left and right footsteps, the volume of the sounds, and the rhythm to evaluate whether there is any decline in motor function or risk of falling. If there is a risk, please indicate the reason and recommend exercises or rehabilitation." 【0115】 Based on the prompts described above, the generating AI assesses the health status and generates customized countermeasures based on the input multimodal data. The resulting output is provided as health measures appropriate for each person being managed, contributing to health management and improved workplace performance. 【0116】 Furthermore, the business support system 1 according to the present invention may be configured to capture an image of the hand of the person being managed and estimate their health status from the hand image. The appearance of the hand, the condition of the wrinkles (number of wrinkles, depth, branching, fineness, overlap, and prominence), distribution, skin tone, visible blood vessels, and nail condition are effective indicators that reflect the overall health status, aging, circulatory system condition, and lifestyle habits. 【0117】 Specifically, if the number and depth of wrinkles on the back of the hands are remarkably prominent, it may indicate dry skin, dehydration, or malnutrition. Pale skin may also be a sign of anemia or peripheral circulatory insufficiency. Furthermore, the color, luster, thickness, brittleness, and presence or absence of a lunula (half-moon) of the fingernails suggest the state of blood flow quality, liver function, and metabolic function related to nail formation. 【0118】 Furthermore, changes in the overall shape of the hand, particularly swelling or asymmetry in the joints, and finger deformities may indicate early signs of conditions such as rheumatic diseases, tenosynovitis, or osteoarthritis. To obtain this image information with high accuracy, it is recommended to acquire high-resolution images under appropriate lighting conditions. 【0119】 Based on this visual information, the generating AI will formulate health measures using prompts such as the following: <Prompt Example 5: Health Estimation Based on Hand Image> "The following is an image of the person being managed. Based on the condition of wrinkles on the back of the hand, skin tone, nail shape, color, thickness, and degree of breakage, visibility of blood vessels, joint swelling, etc., list the estimated health conditions and suggest recommended health measures for each." 【0120】 For example, the following health measures may be planned by the generated AI: If there is a strong tendency towards dryness: Promote hydration, provide guidance on moisturizing care, and recommend vitamin A and E intake. If anemia is suspected: Suggest a diet containing iron and folic acid, and recommend blood tests at a medical institution. If the nails are dull and brittle: Recommend protein and zinc intake, and provide guidance on nail care. If joint swelling is observed: Recommend seeking an orthopedic diagnosis, and suggest reducing the burden of manual work at work. If there is a rapid increase in wrinkles: Since this may be due to a decrease in metabolism associated with aging, advise promoting blood circulation through exercise and bathing. 【0121】 Thus, health status estimated from hand image information is a powerful and supplementary means of detecting early signs of overall physical ailments. It is also highly compatible with image AI analysis and is superior in that it is non-invasive and easily obtainable. Therefore, by combining it with other body movement information and voice information, more multifaceted and highly accurate health monitoring can be realized. 【0122】In the business support system 1 according to the present invention, the system may be configured to estimate health status by focusing on changes over time (temporal changes) by comparing multiple data acquired at different times for the same person under management, in addition to estimating static image and audio data at a single point in time. 【0123】 Specifically, for example, image and audio data is acquired from the person being managed a total of three times (multiple times): before the start of work (morning), after the morning work (noon), and at the end of work (evening). At each time, posture, gait, facial expressions, voice condition (voice tone, speech clarity, balance of footsteps, and even hand position) are analyzed, and the temporal changes are comprehensively evaluated by a generative AI. 【0124】 By paying attention to these changes over time, it is possible to estimate the following health conditions: If posture becomes more hunched over and facial expression becomes less expressive in the evening compared to the morning: physical fatigue or accumulation of mental stress If voice becomes quieter and loses its intonation compared to the morning: decreased concentration, drowsiness, or depressive tendencies If the legs lift less easily when walking or footsteps become more inconsistent in the afternoon: lower limb fatigue or temporary decline in balance If wrinkles and complexion of the hands become more noticeable in the afternoon: poor circulation, dehydration, or environmental stress 【0125】 To perform such analysis based on temporal changes, the following prompt is used for the generating AI: <Prompt Example 6: Health Change Assessment by Time Comparison> "The following is data on the posture, facial expressions, voice, footsteps, etc., of the same employee in the morning, noon, and evening. Based on these temporal changes, analyze and estimate changes in physical and mental state, and propose appropriate health measures as needed." 【0126】 By analyzing these temporal changes in the generating AI, it is possible to analyze not only short-term fluctuations (e.g., daily fatigue levels) but also long-term trends (e.g., chronic fatigue, changes in motivation, accumulation of health risks). Therefore, this embodiment realizes more advanced health support based on the visualization and prediction of changes, rather than merely recording the state. 【0127】In the business support system 1 according to the present invention, the system may be configured to collect and store image and audio data relating to multiple managers and to analyze them across the entire business. That is, the generating AI can not only individually estimate the health status of each manager, but also perform a cross-sectional trend analysis on the data of multiple managers to extract health issues common to the entire business. 【0128】 This structure suggests that, for example, if many managers in a particular workplace exhibit a tendency towards hunching, it may be due to the work environment (desk and chair height, screen placement, etc.), and that implementing comprehensive posture correction measures would be effective. Furthermore, if there is a general tendency for walking balance to worsen towards the evening, a review of organizational health management policies, such as reallocating work content or optimizing break times, becomes necessary. In addition, if the time of day or work content in which signs of mental stress (such as a decrease in voice tone or rigid facial expression) are concentrated among multiple individuals can be identified, measures such as reducing the workload of those tasks or improving management methods can be derived. 【0129】 This configuration allows for the visualization of trends across the entire workplace, rather than simply addressing individual issues, and enables the proposal and implementation of comprehensive health measures for common challenges. This makes the present invention extremely useful in realizing organizational health management, including improving the work environment, preventing occupational accidents, and increasing productivity. 【0130】In the business support system 1 according to the present invention, the generating AI derives collective trends using image data and voice data acquired for multiple managed persons. At this time, the information input to the generating AI is provided as multidimensional and time-series structured data, as exemplified below. (1) Image information (for all managed persons) Date and time of acquisition (e.g., Month Day, AM / PM / Evening) ID or anonymized identifier of the managed person Posture data (angle of leaning forward, left-right difference in shoulders, head position, degree of limb spread, etc.) Estimated skeletal parameters during walking (stride length, left-right balance, speed variation) Facial features (corners of mouth, eyebrow position, degree of eye opening, etc.) (2) Audio information Time of audio recording, managed person ID Voice tone, intonation, speech rate, sound pressure level Left-right intensity, rhythm difference, landing sound waveform of footsteps Frequency of occurrence of non-verbal sounds such as coughs and sighs (3) Temporal and environmental metadata Date and time of data acquisition (morning / noon / evening) Job title and type of work assigned to the managed person Time elapsed since the last acquisition, content of the work done immediately before the acquisition Environmental sensor data such as temperature, humidity, and illuminance 【0131】 These input data are supplied to the generating AI in a chronologically ordered state for each managed individual, and are converted into a format that allows for comparison among managed individuals (normalized or feature extracted). Cross-sectional analysis and clustering processes are then performed to analyze the following: "What percentage of all employees have worsening posture in the afternoon?" "Is there a statistically significant difference in the degree of facial rigidity between employees in Department A and employees in Department B?" "Is the voice tone uniformly lower in the morning and afternoon among employees performing the same tasks?" Thus, the input information is not merely images or audio itself, but includes extracted data on physical and emotional characteristics, as well as associated work and environmental metadata. The generating AI utilizes this to enable the understanding of health trends and the development of improvement guidelines at the workplace level. 【0132】Embodiments of the present invention include the following aspects: <1> A business support system executed by a computer system for supporting a business office where an administrator and a person managed by the administrator coexist, comprising: a countermeasure information generation step instructing a generating AI to input a video of a person managed, as well as observation points in the video image diagnosis (examples of actions / phenomena, suspected symptoms / illnesses), analyze the person, and use the results of the analysis to generate health measures to be output to the person for health management or morale improvement; and a health measure output step in outputting the health measures generated by the generating AI to the person. <2> The business support system according to <1>, wherein the instruction to the generating AI in the countermeasure information generation step includes analyzing audio acquired from the vicinity of the person managed and analyzing at least one of the speech content, voice condition (voice tone, speaking speed, pauses), and footstep patterns. <3> A business support system according to <2>, wherein the instructions to the generating AI in the countermeasure information generation step include cross-sectional analysis of negative statements based on the speech content including text obtained about the person being managed, and proposing health measures to improve morale through positive speech including text. <4> A business support system according to <1>, wherein the instructions to the generating AI in the countermeasure information generation step include analyzing at least one of the left-right balance, intensity, rhythm, and interval of the walking sound of the person being managed, and evaluating the state of motor function. <5> A business support system according to <1>, wherein the instructions to the generating AI in the countermeasure information generation step include receiving an image of the person being managed's hand as input, and analyzing at least one of the state of wrinkles on the hand, the depth of wrinkles, the skin tone, the state of the nails, and the swelling of the joints from the image.<6> A business support system according to <5>, wherein the instructions to the generating AI in the countermeasure information generation step include analyzing the health status, including at least one of dryness / dehydration, malnutrition, peripheral circulatory insufficiency, nail formation abnormalities, and signs of joint disease, based on information analyzed from an image of the person being managed. <7> A business support system according to <1>, wherein the instructions to the generating AI in the countermeasure information generation step include acquiring an image or voice of the person being managed at multiple time periods (a time period around the start of work, a time period around noon, and a time period around the end of work), and evaluating the health status based on the changes over time. <8> A business support system according to <7>, wherein the instructions to the generating AI in the countermeasure information generation step include estimating at least one of the degree of fatigue accumulation, decreased concentration, changes in mental stress, and changes in physical balance based on the changes over time, and generating health measures based on the results. <9> A business support system according to <1>, wherein the instructions to the generating AI in the countermeasure information generation step include generating individually optimized health measures, taking into account past health analysis results associated with the person being managed. <10> A business support system according to <1>, wherein the instructions to the generating AI in the countermeasure information generation step include cross-sectionally analyzing images and audio obtained for multiple persons being managed, extracting health risks common to the entire group, and proposing a health guidance policy for the entire business.<11> A method for supporting a business establishment in which an administrator and an administrator who is managed by the administrator coexist, the method being implemented by a computer system, comprising: a step of generating countermeasures information instructing a generating AI to input an image of an administrator, analyze the administrator's walking pattern, body movements, and posture, and use the results of the analysis to generate health countermeasures to be output to the administrator for the purpose of health management or morale improvement of the administrator; and a step of outputting health countermeasures to the administrator, outputting the health countermeasures generated by the generating AI. <12> A business support program that operates a computer system for supporting a business establishment where an administrator and an administrator-managed person coexist, comprising: a countermeasure information generation step instructing a generating AI to input an image of an administrator, analyze the administrator's walking pattern, body movements, and posture, and use the results of the analysis to generate health measures to be output to the administrator for the purpose of health management or morale improvement of the administrator; and a health measure output step in which the health measures generated by the generating AI are output to the administrator. 【0133】 The present invention is not limited to the embodiments described above. The embodiments described above can be modified in various ways within the scope of the technical idea of ​​the present invention. For example, the functions of the server device 100 may be realized by a system consisting of one or more computers. 【0134】 Furthermore, the technical elements of the embodiments described above may be applied individually, or they may be divided into multiple parts, such as program components and hardware components, and applied accordingly. 【0135】 The present invention has been described above, focusing on its embodiments. 【0136】1... Business support system, 50... Data communication network, 100... Server device, 110... Processing unit, 111... Countermeasure information generation unit, 112... Output generation unit, 113... Attendance record unit, 114... Report creation unit, 120... Storage unit, 121... User information, 122... Attendance record information, 123... Detection information, 124... Countermeasure information, 125... Report information, 130... Communication unit, 200... Attendance record device, 210... Imaging unit, 220... Sound collection unit, 230... Human body recognition unit, 240... Time stamping unit, 250... Communication unit, 260... Health countermeasure output unit, 300... Generation AI service.

Claims

1. A business support system for a business office where an administrator and an administrator-managed person coexist, the system being operated by a computer system, wherein the computer system inputs an image of the administrator-managed person and instructs the AI ​​to generate health measures to be output to the administrator-managed person for the purpose of health management or morale improvement of the administrator-managed person; and performs a health measures output step to output the health measures generated by the AI ​​to the administrator-managed person; further, the computer system performs an expression and voice acquisition step to acquire the expression and voice of the administrator-managed person; in the measure information generation step, the system acquires the expression and voice as the administrator-managed person's reaction obtained as a result of outputting predetermined expressions and voices to be seen and heard by the administrator-managed person; analyzes any of the expressions or voices to identify the type of emotion from a plurality of emotion types including stress, fatigue, joy, sadness, anger, surprise, fear, disgust, contempt, interest, confusion, relief, and expectation; and instructs the AI ​​to generate appropriate health measures for the administrator-managed person based on the type of emotion.

2. The office support system according to claim 1, wherein the computer system has a storage unit that stores the past health status, health measures history, and warning history of the person being managed, and in the countermeasure information generation step, the past health status and health measures history of the person being managed are input, and for persons who have been the target of warnings in the past, the AI ​​is instructed to generate a message that includes suggestions for specific health measures such as meal menus appropriate to the person's condition, based on the person's past health status and warning history.

3. An office support system according to claim 1, wherein the computer system has a storage unit that stores the past health status of the person being managed, the history of health measures, and the history of warnings, and in the countermeasure information generation step, the past health status of the person being managed, the history of health measures, and the history of warnings are input, and the AI ​​is instructed to generate health measures, including careful consideration and verbal communication, for the person being managed who has been the target of a warning in the past.

4. The office support system according to claim 1, wherein the computer system, in the countermeasure information generation step, (A1) generates an image of the person being managed who has a negative expression, by image processing, and (A2) instructs the AI ​​to display the generated image of the person being managed who has a positive expression.

5. A business support system according to claim 1, wherein the countermeasure information generation step includes analyzing changes in the facial expression and posture of the person being managed in response to being spoken to, and based on the results of the analysis, predicting the possibility of misconduct in any of the following cases: (B1) when the person being managed shows an unnatural change in facial expression in response to being spoken to, or (B2) when the person being managed shows an unnatural posture in response to being spoken to, and instructing the AI ​​to generate health measures, including a warning message to the person being managed or a notification to the manager, based on the prediction results.

6. A business support system according to claim 1, wherein the instructions given to the generating AI in the countermeasure information generation step include: (1) acquiring images of the person being managed at multiple times, including the start and end of work, and evaluating their health status based on temporal changes; (2) estimating the degree of fatigue accumulation, decrease in concentration, progression of mental stress, or change in physical balance based on temporal changes, and generating health countermeasures based on the results; or (3) extracting health risks based on statistical information obtained by cross-sectionally analyzing videos or audio obtained for multiple persons being managed as health risks that may be common to the entire group, and proposing a health guidance policy for the entire business.

7. A business support system, executed by a computer system, for supporting a business establishment where an administrator and a person managed by the administrator coexist, wherein the computer system performs a countermeasure information generation step of inputting an image of the person managed and instructing the AI ​​to generate health measures to be output to the person for the purpose of health management or morale improvement of the person, and a health measure output step of outputting the health measures generated by the AI ​​to the person, and further, the computer system has a storage unit that stores the person's past health status, history of health measures, and history of warnings, and in the countermeasure information generation step, (1) inputs the person's past health status and history of health measures and instructs the AI ​​to generate a message, including a suggestion of a specific health measure meal menu, etc., according to the person's condition, based on the person's past health status and warning history, for the person who was previously the target of a warning, or (2) An office support system that inputs the past health status of the person being managed, the history of the health measures taken, and the history of the warnings, and instructs the AI ​​to generate health measures, including thoughtful communication, for the person being managed who has been the target of a warning in the past.

8. A business support system, executed by a computer system, for supporting a business establishment where an administrator and an administrator-managed person coexist, wherein the computer system performs a countermeasure information generation step in which it inputs an image of an administrator and instructs the AI ​​to generate health measures to be output to the administrator for the purpose of health management or morale improvement of the administrator, and a health measure output step in which it outputs the health measures generated by the AI ​​to the administrator, and further, in the countermeasure information generation step, for an administrator with a negative expression, (A1) generates an image in which the administrator's expression is converted to a bright expression by image processing, (A2) instructs the AI ​​to display the generated image of the bright expression, or, in the countermeasure information generation step, analyzes the changes in the administrator's expression and posture in response to being spoken to, and based on the results of the analysis, predicts the possibility of misconduct in any of the following cases: (B1) when the administrator shows an unnatural change in expression in response to being spoken to, or (B2) A business support system that, if the person being managed exhibits an unnatural posture in response to the aforementioned verbal instruction, instructs the AI ​​to generate health measures, including a warning message for the person being managed or a notification to the manager, based on the prediction result.

9. A business support system, executed by a computer system, for supporting a business establishment where an administrator and an employee managed by the administrator coexist, comprising: a countermeasure information generation step instructing a generating AI to input video footage of the employee, analyze the employee by video image diagnosis of predetermined observation points including the employee's walking and movements, and use the results of the analysis to generate health measures to be output to the employee for health management or morale improvement; and a health measure output step in which the health measures generated by the generating AI are output to the employee, wherein the instructions to the generating AI in the countermeasure information generation step include: (1) acquiring images of the employee at multiple times, including the start and end of work, and evaluating the health status based on temporal changes; (2) estimating the degree of fatigue accumulation, decrease in concentration, progression of mental stress, or change in physical balance based on temporal changes, and generating health measures based on the results, or (3) A workplace support system that includes extracting health risks based on statistical information obtained by cross-sectionally analyzing videos or audio recordings of multiple persons under management, as health risks that may be common to the entire group, and proposing a health guidance policy for the entire workplace.

Images