Ai agent system and program

Abstract

The present invention comprises a facility agent that mediates communication between a user and a facility, wherein the facility agent comprises: a target facility information reception unit that receives information pertaining to a target facility for which the facility agent is responsible; an instruction information reception unit that receives an instruction pertaining to the target facility; a scenario generation unit that generates a response scenario on the basis of the information pertaining to the target facility and the content of the instruction pertaining to the target facility; an analysis processing execution unit that analyzes a state pertaining to the target facility on the basis of the response scenario; and an analysis result output unit that outputs an analysis result.

Description

AI agent system and program CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application is based on Japanese Application No. 2024-147986 filed on August 29, 2024, the contents of which are incorporated herein by reference.

[0002] The present invention relates to an AI agent system and program.

[0003] To maintain the operation of facilities such as factory automation (FA) equipment, maintenance management, such as daily maintenance inspections and trouble prediction, is essential. Traditionally, maintenance management work has often relied on manual labor and the know-how of the workers. Meanwhile, as described in Patent Literature 1, for example, a method has been proposed in which the status of factory facilities is monitored by a system and anomalies in the facilities are detected using big data and machine learning.

[0004] Patent No. 7465008

[0005] However, conventional systems such as that disclosed in Patent Document 1 did not implement a mechanism in which an AI agent autonomously manages the status of each piece of equipment in response to requests from people.

[0006] One of the objects of the present invention is to enable autonomous maintenance management of facilities using AI agents.

[0007] The AI ​​agent system of the present invention includes an equipment agent that mediates communication between a user and equipment, and the equipment agent is equipped with a target equipment information receiving unit that receives information related to the target equipment for which the equipment agent is responsible, an instruction information receiving unit that receives instructions related to the target equipment, a scenario generation unit that creates a response scenario based on the information related to the target equipment and the content of the instructions related to the target equipment, an analysis processing execution unit that analyzes the state of the target equipment based on the response scenario, and an analysis result output unit that outputs the results of the analysis.

[0008] According to the above configuration, the user can grasp the state of the target equipment while interacting with the equipment agent. Furthermore, the equipment agent can autonomously perform maintenance management of the equipment.

[0009] The equipment agent may further include a scenario storage unit that stores past analysis results and previously generated response scenarios related to the target equipment, and the scenario generation unit may create the response scenarios using the past analysis results and the previously generated response scenarios. With the above configuration, it is possible to create more appropriate response scenarios by using past analysis results and response scenarios.

[0010] The equipment agent also includes a processing means storage unit that stores information to be used in analyzing the status of the target equipment, and the scenario generation unit can create the response scenario using the information stored in the processing means storage unit, thereby enabling efficient analysis using accumulated information from tools, etc.

[0011] The equipment agent may also include a visualization data output unit that generates and outputs visualization data that visualizes the analysis results of the status of the target equipment, thereby making it easier for a user to understand the analysis results.

[0012] The system may also include an equipment agent that mediates communication between the user and the equipment, the equipment agent including: a target equipment information receiving unit that receives information related to the target equipment for which the equipment agent is responsible; an instruction information receiving unit that receives instructions related to the target equipment; a scenario generation unit that creates a response scenario based on information about the target equipment and the content of the instructions related to the target equipment; and a control unit that controls the target equipment based on the response scenario. This allows the user to appropriately control the target equipment while interacting with the equipment agent. Even an operator who is not familiar with handling the target equipment can automatically perform appropriate control by the equipment agent by simply providing instructions and guidelines.

[0013] The instruction information receiving unit may also receive the results of the analysis of the target equipment as instructions for the target equipment, thereby enabling further analysis by feeding back the results of the analysis itself, thereby enabling more accurate and realistic analysis.

[0014] The system may further include a control agent that mediates communication between the user and the equipment agents, the control agent including a user communication unit that receives inquiries from the user in natural language and outputs responses to the user in natural language, and an agent communication unit that outputs processing requests to the equipment agents and receives processing results from the equipment agents, and the control agent outputs processing requests to the equipment agents in response to the inquiries received from the user. This allows the user to comprehensively grasp the status of each piece of target equipment via the control agent, even if they do not have detailed knowledge of each piece of target equipment.

[0015] The equipment agent may also include an inter-agent communication unit that outputs processing requests to other equipment agents and receives processing results from the other equipment agents. This allows the equipment agents to communicate with each other while performing overall maintenance management, which is particularly effective in the case of a line where each piece of equipment is hierarchically organized.

[0016] The target equipment information receiving unit may receive operation log data of the target equipment as information related to the target equipment, and the analysis processing execution unit may analyze the status of the target equipment using the operation log data. In this way, the equipment agent analyzes a huge amount of log data and outputs analysis results of malfunctions, etc., allowing the user to efficiently know the status of the target equipment.

[0017] The analysis processing execution unit may also extract data from the operation log data for a time period in which an abnormality was detected in the target equipment, and perform analysis of the target equipment using the extracted data for that time period. This makes it possible to narrow down the points to be analyzed from a large amount of log data, thereby enabling efficient and accurate analysis.

[0018] The scenario generator may also select a method to be used for analysis based on the contents of the operation log data, thereby enabling selection of a more appropriate analysis method suited to the actual operation of the target equipment.

[0019] The equipment agent may further include an order processing execution unit that executes an order processing for parts necessary for repairing the target equipment when the state of the target equipment satisfies a first condition based on the result of the analysis. This allows the necessary parts to be ordered quickly and at an appropriate time, and reduces the workload of the user.

[0020] The equipment agent may further include a maintenance request execution unit that issues a maintenance request for the target equipment when the state of the target equipment satisfies a second condition based on the results of the analysis. This allows the necessary maintenance request to be made promptly and at an appropriate time, and reduces the workload of the user.

[0021] The program of the present invention causes a computer to function as a target equipment information receiving unit that receives information related to the target equipment for which the computer is responsible, an instruction information receiving unit that receives instructions related to the target equipment, a scenario generation unit that creates a response scenario based on the information related to the target equipment and the content of the instructions related to the target equipment, an analysis processing execution unit that analyzes the status of the target equipment based on the response scenario, and an analysis result output unit that outputs the results of the analysis, thereby causing the computer to function as an equipment agent that mediates communication between users and equipment.

[0022] According to the above configuration, the user can grasp the state of the target equipment while interacting with the equipment agent. Furthermore, the equipment agent can autonomously perform maintenance management of the equipment.

[0023] According to the present invention, it is possible to autonomously perform maintenance management of facilities using AI agents.

[0024] A diagram illustrating an overview of an AI agent system 1 according to an embodiment of the invention. A diagram showing an example of the hardware configuration of an equipment agent 10 and a control agent 20 according to an embodiment of the invention. A block diagram showing an example of a functional module executed by a processor 11 of the equipment agent 10 according to an embodiment of the invention. A block diagram showing an example of a functional module executed by a processor 11 of the control agent 20 according to an embodiment of the invention.

[0025] (Overview of AI Agent System 1) Fig. 1 is a diagram illustrating an overview of an AI agent system 1 according to an embodiment of the present invention. As shown in Fig. 1, the AI ​​agent system 1 includes equipment agents 10A and 10B and a control agent 20. The equipment agents 10A and 10B are AI agents that mediate communication between a user 50 and devices 30A and 30B (target equipment) installed at a factory work site or the like. The equipment agent 10A is responsible for the device 30A, and the equipment agent 10B is responsible for the device 30B, respectively, and can autonomously perform maintenance management of the equipment they are responsible for in response to requests from the user 50.

[0026] The control agent 20 is an AI agent that mediates communication between the user 50 and the facility agents 10A and 10B. The control agent 20 converses with the user 50 in natural language, outputs necessary processing requests to the facility agents 10A and 10B, outputs responses to the user 50 based on the processing results received from the facility agents 10A and 10B, and executes processing to meet the requests of the user 50.

[0027] For example, when a user 50 asks the control agent 20, "How's the production line going?", the control agent 20 outputs an instruction to the equipment agents 10A and 10B to analyze the operating status of the devices 30A and 30B. The equipment agents 10A and 10B each create a response scenario (work process) for analyzing the operating status of the devices 30A and 30B, and perform analysis processing in accordance with the created response scenario. Furthermore, the equipment agents 10A and 10B output the results of the analysis processing to the control agent 20. The control agent 20 consolidates the analysis results received from the equipment agents 10A and 10B, creates and outputs a response to the user 50. For example, if the analysis results by the equipment agent 10A indicate that an abnormality has occurred in the operation of the device 30A, the control agent 20 responds to the user 50 that the operation of the device 30A is unstable. The user 50 may instruct the control agent 20 to perform necessary processing (such as ordering parts for repairing the device 30A) based on the response from the control agent 20. The control agent 20 may perform processing to order parts from an external vendor in accordance with the instructions from the user 50.

[0028] It is also possible for the user 50 and the equipment agents 10A and 10B to directly interact with each other without going through the control agent 20. For example, when the user 50 asks the equipment agent 10A about the condition of the device 30A, the equipment agent 10A may create and execute a response scenario for analysis, and respond to the user 50 directly with the analysis results. Furthermore, the equipment agents 10A and 10B may output requests to the user 50 (such as a request to order necessary repair parts) based on the analysis results.

[0029] Alternatively, direct communication may occur between the equipment agent 10A and the equipment agent 10B. For example, if the equipment 30B is a lower-level equipment from the viewpoint of the equipment 30A, the equipment agent 10A may request the equipment agent 10B to analyze the status of the equipment 30B and receive the analysis results from the equipment agent 10B. Furthermore, the equipment agent 10A may determine the operating status of its own (the equipment 30A) including the status of the lower-level equipment 30B, and output the analysis results to the user 50 or the control agent 20.

[0030] In addition, the equipment agents 10A and 10B may create a response scenario for controlling the devices 30A and 30B based on instructions from the user 50 or the control agent 20, and control the devices 30A and 30B based on the created response scenario.

[0031] In this embodiment, the AI ​​agent may be software that uses generative AI. The AI ​​agent is interactive software that, when a user inputs instructions or requests in natural language, identifies and automatically executes the processing required to respond to the input instructions or requests. The AI ​​agent can independently assemble and execute the work steps required to achieve the instructions or requests, even without detailed instructions from the user. The AI ​​agent can also autonomously improve the work content by memorizing the results of past work, accumulating them as knowledge, and learning from them.

[0032] (Hardware Configuration) FIG. 2 is a diagram illustrating an example of the hardware configuration of the equipment agent 10 (10A, 10B) and the control agent 20 according to this embodiment. As illustrated in FIG. 2, the equipment agent 10 and the control agent 20 are computers including, as hardware resources, a processor 11, a main memory 12, an input / output interface 13, a communication interface 14, and a storage device 15. The storage device 15 is a computer-readable storage medium such as a semiconductor memory (e.g., a volatile memory or a nonvolatile memory) or a disk medium (e.g., a magnetic storage medium or a magneto-optical storage medium). The storage device 15 may store a program for causing the processor 11 to execute the AI ​​agent function, accumulated past analysis results, previously created response scenarios, information on an analysis method used by the equipment agent 10 to analyze the target equipment, and information on a control method for the target equipment. The program is loaded from the storage device 15 into the main memory 12 and interpreted and executed by the processor 11 to perform various functions. The facility agent 10 (10A, 10B) and the control agent 20 may be implemented in separate computers, or two or more agents may be implemented in one computer.

[0033] (Functional Configuration of Equipment Agent 10) Fig. 3 is a block diagram showing an example of functional modules executed by the processor 11 of the equipment agent 10. As shown in Fig. 3, the functional modules executed by the processor 11 of the equipment agent 10 include a target equipment information receiving unit 101, an instruction information receiving unit 102, a scenario generating unit 103, an analysis processing executing unit 104, an analysis result output unit 105, a control unit 106, a scenario storing unit 107, a visualization data output unit 108, an equipment agent-to-agent communication unit 109, an order processing executing unit 110, and a maintenance request executing unit 111. The storage device 15 also includes a processing means storage unit 151 that stores analysis methods used by the equipment agent 10 to analyze the target equipment and external processing programs used to analyze and control the target equipment. As will be described later, when analyzing and controlling the target equipment, the equipment agent 10 can incorporate external processing into a response scenario by referencing the information stored in the processing means storage unit 151.

[0034] The target equipment information receiving unit 101 receives information related to the target equipment such as the device 30 managed by the equipment agent 10. Information related to the target equipment includes, for example: (1) Information related to the operation and function of the target equipment itself, such as the manual and specifications of the target equipment; (2) Information related to the operation of the target equipment that can be set in advance by the user, such as the operating parameters set for the target equipment and the threshold for alarm generation; (3) Dynamic information indicating the operating status of the target equipment, such as operation information such as the log data of the target equipment and operation information detected by sensors installed in the target equipment.

[0035] The instruction information receiving unit 102 receives instructions related to the target equipment. Instructions related to the target equipment can be received from a terminal operated by the user 50, a control agent, or other equipment agents. The terminal operated by the user 50, the control agent, and other equipment agents are connected via a communication network. The content of the instructions can also be received in natural language.

[0036] The scenario generation unit 103 creates a response scenario based on information about the target equipment received by the target equipment information receiving unit 101 and the content of the instruction related to the target equipment received by the instruction information receiving unit 102. The response scenario is a measure for responding to the received instruction, and if the instruction instructs analysis of the operating status, the response scenario selects an analysis method for analyzing the operating status of the target equipment. If the instruction instructs control of the target equipment, the response scenario determines a specific control policy. The response scenario may be created using an LLM (large-scale language model). Furthermore, when creating a response scenario, the scenario generation unit 103 may cooperate with a knowledge management tool to select an appropriate analysis method. Furthermore, the scenario generation unit 103 may select an analysis method based on the content of operation log data of the target equipment. Furthermore, the scenario generation unit 103 can incorporate external processing used for analysis and control of the target equipment into the response scenario by referring to the processing means storage unit 151.

[0037] The analysis processing execution unit 104 analyzes the target equipment based on the response scenario created by the scenario generation unit 103. The analysis method selected in the response scenario may be, for example, time series analysis, image diagnosis, bottleneck analysis, or vibration analysis, and the analysis processing execution unit 104 analyzes the target equipment using the selected analysis method. The analysis processing execution unit 104 also analyzes the operation log data of the target equipment. In this case, the analysis processing execution unit 104 may analyze the time period during which an abnormality occurred in the target equipment from the operation log data, extract data from the time period during which the abnormality was detected, and perform analysis. If the response scenario incorporates external processing, the analysis processing execution unit 104 retrieves the target program from the processing means storage unit 151, sets necessary parameters such as arguments, and executes the program. For example, if the response scenario incorporates a program for log pattern analysis, the analysis processing execution unit 104 retrieves the program script from the processing means storage unit 151, and executes the script by specifying the path to the log file and the analysis organization as arguments. The analysis processing execution unit 104 may retrieve text summary information for each input log pattern as the execution result.

[0038] The analysis result output unit 105 outputs the results of the analysis performed by the analysis processing execution unit 104. For example, in the case of a response to an instruction received from the user 50, the output may be in natural language to a terminal operated by the user 50. In addition, in response to an instruction from a control agent or another facility agent, the output may be directed to the respective agent. Furthermore, the instruction information receiving unit 102 may receive the analysis results as instruction information, and may further repeat the creation of a response scenario and the implementation of the analysis.

[0039] The control unit 106 controls the target equipment based on the response scenario created by the scenario generation unit 103. Specifically, the control unit 106 may receive information from the user 50, such as changes in the operating environment or changes in the production policy of the line, and once the scenario generation unit 103 creates a response scenario for control corresponding to those changes, the control unit 106 may perform control, such as changing the operating parameters of the target equipment, based on the response scenario. Specifically, the response scenario may include information such as what values ​​to set for which setting parameters when performing the specified control, and what format of data should be provided to the interface of the target equipment. The control unit 106 may also control the target equipment based on the results of analysis by the analysis processing execution unit 104. Furthermore, if an external process is incorporated into the response scenario, the control unit 106 may obtain the corresponding program from the processing means storage unit 151, set parameters such as necessary arguments, and execute the program.

[0040] The scenario accumulation unit 107 accumulates and manages past analysis results for the target equipment output from the analysis result output unit 105 and past response scenarios created by the scenario generation unit 103. The accumulated analysis results and response scenarios may be used by the scenario generation unit 103 when creating a response scenario. The scenario generation unit 103 may also create a response scenario using a machine learning model obtained by performing machine learning using past response scenarios and analysis results executed based on the response scenarios as learning data. This allows the model to evolve through learning so that it can create more appropriate response scenarios.

[0041] The visualization data output unit 108 visualizes the analysis results of the target equipment output from the analysis result output unit 105. Specifically, the visualization data output unit 108 converts the analysis results into a format that is easy for the user 50 to understand, such as a graph or chart, and displays the results on a terminal or the like.

[0042] The equipment inter-agent communication unit 109 outputs processing requests to other equipment agents 10 and receives processing results from other equipment agents 10. When responding to inquiries from the user 50 or the control agent 20, the equipment inter-agent communication unit 109 may issue instructions to the equipment agent 10 in charge of the lower-level target equipment as necessary, or may perform analysis while having a dialogue with the equipment agent 10.

[0043] The order processing execution unit 110 executes an order processing for parts necessary for repairing the target equipment when the condition of the target equipment satisfies a predetermined condition (first condition) based on the result of the analysis by the analysis processing execution unit 104. The notification destination can be a system of an external parts manufacturer, etc. The first condition can be a preset threshold value (frequency of abnormality, degree of deviation from normal value, etc.) for determining that part replacement is necessary.

[0044] The maintenance request execution unit 111 issues a maintenance request for the target equipment when the state of the target equipment satisfies a predetermined condition (second condition) based on the results of analysis by the analysis processing execution unit 104. The notification destination can be a system of an external maintenance contractor or the like. The second condition can be a preset threshold value for determining that maintenance is necessary (frequency of abnormality, degree of deviation from normal values, etc.). If external processing is incorporated into the response scenario, the maintenance request execution unit 111 may obtain a maintenance request processing program from the processing means storage unit 151, set parameters such as necessary arguments, and execute the program.

[0045] (Functional Configuration of the Control Agent 20) Fig. 4 is a block diagram showing an example of functional modules executed by the processor 11 of the control agent 20. As shown in Fig. 3, the functional modules executed by the processor 11 of the equipment agent 10 include a user communication unit 201 and an agent communication unit 202.

[0046] The user communication unit 201 accepts inquiries from the user 50 in natural language and outputs responses to the user 50 in natural language.

[0047] The agent communication unit 202 outputs a processing request to the equipment agent 10 in response to an inquiry received from the user 50, and receives the processing result from the equipment agent 10. The user communication unit 201 determines the content of the response to the user 50 based on the processing result from the equipment agent 10.

[0048] Next, a specific example of the operation of the AI ​​agent system 1 will be described. (Example 1: User-Equipment Agent) In the AI ​​agent system 1, a user 50 can directly instruct the equipment agent 10 via a terminal, interact directly with the equipment agent 10, request an analysis of the target equipment, and obtain the analysis results. For example, when performing maintenance and inspection of a specific device 30A, the user 50 can pose a question such as, "Is there any abnormality in the operation of device 30A?" to the equipment agent 10A in charge of device 30A. The equipment agent 10A can then analyze, for example, operation log data based on a created response scenario. In this case, the response scenario may first be used to detect the time when the abnormality occurred in the log data, and then to analyze the log data for that time period in detail. If the analysis identifies that deterioration of part C is the cause of the abnormal operation, the user may be told, for example, "Part C needs to be replaced." Furthermore, if the analysis of the log data identifies a problem with lower-level device B, the equipment agent 10A may be instructed to perform analysis on the equipment agent 10B in charge of device B. When the equipment agent 10A receives from the equipment agent 10B an analysis result indicating that maintenance of part D is required, the equipment agent 10A may respond to the user 50 by saying, "Maintenance of part D of device B is required."

[0049] (Example 2: User-Control Agent-Equipment Agent) In the AI ​​agent system 1, the user 50 gives instructions to the control agent 20 via a terminal, and the control agent 20 can give instructions to each equipment agent 10 based on the user 50's instructions. For example, when the user 50 asks the control agent 20, "How is line E doing?", the control agent 20 instructs all equipment agents 10 responsible for devices on line E to analyze their operating status. The equipment agents 10 that receive the instructions analyze the devices 30 they are responsible for. The control agent 20 acquires the analysis results from each equipment agent 10 and identifies devices 30 that require repair or part replacement. If necessary, the control agent 20 may instruct the target equipment agents 10 to perform additional analysis. The control agent 20 responds to the user 50 based on the analysis results from the equipment agents 10. For example, if part F of device 30B needs to be replaced, the control agent 20 may respond, "Part F of device 30B needs to be replaced." Alternatively, the control agent 20 may directly order part F from the parts manufacturer's system. Alternatively, the equipment agent 10B in charge of the device 30B may be instructed to place an order for the part F, and the equipment agent 10B may process the order for the part F to the part manufacturer's system. The analysis results may also be sent directly to the user 50 from the equipment agent 10B without going through the control agent 20.

[0050] (Example 3: Instructions to oneself) In the AI ​​agent system 1, the equipment agent 10 can receive the analysis results of the target equipment that it has output as instructions to itself. For example, it may be possible to modify the response scenario based on the analysis results and perform the analysis again.

[0051] As described above, according to this embodiment, the equipment agent 10, which is an AI agent, is set for the target equipment, and the equipment agent 10, while interacting with the user 50, creates a response scenario for analyzing the status of the target equipment based on the content of instructions and information related to the target equipment, performs analysis in accordance with the response scenario, and outputs the analysis results. This allows the user 50 to understand the status of the target equipment while interacting with the AI ​​agent. Furthermore, the AI ​​agent can autonomously perform maintenance management of the target equipment. Furthermore, by having the equipment agent 10 autonomously perform maintenance management of the target equipment, it becomes possible to predict problems, and by taking appropriate measures, equipment problems can be prevented before they occur.

[0052] In addition, a control agent 20 is provided that controls multiple equipment agents 10 and acts as an intermediary between the user 50 and the equipment agents 20, so that the user 50 can carry out overall maintenance management of a production line, etc. through the AI ​​agent without having to specify individual target equipment.

[0053] Furthermore, since the equipment agents 10 can communicate with each other while analyzing the status, the equipment agent 10 in charge of higher-level equipment can grasp the status of the lower-level equipment through the equipment agent 10 in charge of the lower-level equipment. This allows the AI ​​agent to efficiently manage the maintenance of a line or the like where multiple pieces of equipment are hierarchically related.

[0054] Furthermore, the equipment agent 10 performs re-analysis based on the analysis results it has output, so that a more accurate analysis can be performed while feeding back the analysis results.

[0055] Note that some or all of the above embodiments can also be described as in the following supplementary notes, but are not limited to them. (Supplementary Note 1) An AI agent system including an equipment agent that mediates communication between a user and equipment, the equipment agent comprising: a target equipment information receiving unit that receives information related to target equipment for which the equipment agent is responsible; an instruction information receiving unit that receives instructions related to the target equipment; a scenario generation unit that creates a response scenario based on the information related to the target equipment and the content of the instructions related to the target equipment; an analysis processing execution unit that analyzes the state of the target equipment based on the response scenario; and an analysis result output unit that outputs the results of the analysis. (Supplementary Note 2) The AI ​​agent system according to Supplementary Note 1, wherein the equipment agent comprises: a scenario storage unit that stores past analysis results and previously generated response scenarios related to the target equipment, and the scenario generation unit creates the response scenario using the past analysis results and the previously generated response scenario. (Supplementary Note 3) The AI ​​agent system according to Supplementary Note 1 or 2, wherein the equipment agent comprises: a processing means storage unit that stores information to be used in analyzing the state of the target equipment, and the scenario generation unit creates the response scenario using the information stored in the processing means storage unit. (Supplementary Note 4) The AI ​​agent system according to Supplementary Note 1 or 2, wherein the equipment agent comprises: a visualization data output unit that generates and outputs visualization data that visualizes analysis results of the state of the target equipment. (Supplementary Note 5) An AI agent system including an equipment agent that mediates communication between a user and equipment, wherein the equipment agent comprises: a target equipment information receiving unit that receives information related to the target equipment that the equipment agent is responsible for, an instruction information receiving unit that receives instructions related to the target equipment, a scenario generation unit that creates a response scenario based on information related to the target equipment and the content of the instructions related to the target equipment, and a control unit that controls the target equipment based on the response scenario.(Supplementary Note 6) The AI ​​agent system according to any one of Supplements 1 to 6, wherein the instruction information receiving unit receives a result of an analysis on the target equipment as an instruction on the target equipment. (Supplementary Note 7) The AI ​​agent system according to any one of Supplements 1 to 6, further comprising: a control agent that mediates communication between a user and the equipment agents, the control agent comprising: a user communication unit that receives an inquiry from the user in natural language and outputs a response to the user in natural language, and an agent communication unit that outputs a processing request to the equipment agents and receives processing results from the equipment agents, and the control agent outputs a processing request to the equipment agents in response to the inquiry received from the user. (Supplementary Note 8) The AI ​​agent system according to any one of Supplements 1 to 7, wherein the equipment agents comprise: an inter-agent communication unit that outputs processing requests to other equipment agents and receives processing results from the other equipment agents. (Supplementary Note 9) The AI ​​agent system according to any one of Supplements 1 to 4, wherein the target equipment information receiving unit receives operation log data of the target equipment as information related to the target equipment, and the analysis processing execution unit analyzes the state of the target equipment using the operation log data. (Supplementary Note 10) The AI ​​agent system according to Supplementary Note 9, wherein the analysis processing execution unit extracts, from the operation log data, data for a time period in which an abnormality was detected in the target equipment, and performs analysis of the target equipment using the extracted data for that time period. (Supplementary Note 11) The AI ​​agent system according to Supplementary Note 10, wherein the scenario generation unit selects a method to be used for analysis based on the content of the operation log data. (Supplementary Note 12) The AI ​​agent system according to any one of Supplements 1 to 4, wherein the equipment agent comprises: an order processing execution unit that executes an order processing for parts necessary for repairing the target equipment when the state of the target equipment satisfies a first condition based on the result of the analysis.(Supplementary Note 13) The AI ​​agent system according to any one of Supplementary Notes 1 to 4, wherein the equipment agent comprises: a maintenance request execution unit that issues a maintenance request for the target equipment when, based on the result of the analysis, the state of the target equipment satisfies a second condition. (Supplementary Note 14) A program for causing a computer to function as an equipment agent that mediates communication between a user and equipment by causing a computer to function as: a target equipment information receiving unit that receives information related to the target equipment for which the computer is responsible; an instruction information receiving unit that receives instructions related to the target equipment; a scenario generation unit that creates a response scenario based on the information related to the target equipment and the content of the instructions related to the target equipment; an analysis processing execution unit that analyzes the state of the target equipment based on the response scenario; and an analysis result output unit that outputs the results of the analysis.

[0056] 1...AI agent system, 10A, 10B...equipment agent, 11...processor, 12...main memory, 13...input / output interface, 14...communication interface, 15...storage device, 20...control agent, 30A, 30B...device, 50...user, 101...target equipment information receiving unit, 102...instruction information receiving unit, 103...scenario generation unit, 104...analysis processing execution unit, 105...analysis result output unit, 106...control unit, 107...scenario storage unit, 108...visualization data output unit, 109...equipment agent inter-communication unit, 110...order processing execution unit, 111...maintenance request execution unit, 151...processing means memory unit, 201...user communication unit, 202...agent communication unit

Claims

1. An AI agent system including an equipment agent that mediates communication between a user and equipment, the equipment agent comprising: a target equipment information receiving unit that receives information related to the target equipment for which the equipment agent is responsible; an instruction information receiving unit that receives instructions related to the target equipment; a scenario generation unit that creates a response scenario based on the information related to the target equipment and the content of the instructions related to the target equipment; an analysis processing execution unit that analyzes the status of the target equipment based on the response scenario; and an analysis result output unit that outputs the results of the analysis.

2. The AI ​​agent system of claim 1, wherein the equipment agent comprises a scenario storage unit that stores past analysis results and previously generated response scenarios related to the target equipment, and the scenario generation unit creates the response scenarios using the past analysis results and the previously generated response scenarios.

3. An AI agent system as described in claim 1 or 2, wherein the equipment agent comprises a processing means memory unit that stores information to be used in analyzing the status of the target equipment, and the scenario generation unit creates the response scenario using the information stored in the processing means memory unit.

4. The AI ​​agent system according to claim 1 or 2, wherein the equipment agent comprises: a visualization data output unit that generates and outputs visualization data that visualizes the status analysis results for the target equipment.

5. An AI agent system including an equipment agent that mediates communication between a user and equipment, the equipment agent comprising: a target equipment information receiving unit that receives information related to the target equipment for which the equipment agent is responsible; an instruction information receiving unit that receives instructions related to the target equipment; a scenario generation unit that creates a response scenario based on information related to the target equipment and the content of the instructions related to the target equipment; and a control unit that controls the target equipment based on the response scenario.

6. The AI ​​agent system according to claim 1 or 2, wherein the instruction information receiving unit receives the results of the analysis regarding the target equipment as instructions regarding the target equipment.

7. An AI agent system as described in claim 1 or 2, comprising a control agent that mediates communication between the user and the equipment agents, wherein the control agent comprises: a user communication unit that receives inquiries from the user in natural language and outputs responses to the user in natural language; and an agent communication unit that outputs processing requests to the equipment agents and receives processing results from the equipment agents, and the control agent outputs processing requests to the equipment agents in response to the inquiries received from the user.

8. An AI agent system according to claim 1 or 2, wherein the equipment agent is provided with an inter-equipment agent communication unit that outputs processing requests to other equipment agents and receives processing results from the other equipment agents.

9. An AI agent system as described in claim 1 or 2, wherein the target equipment information receiving unit receives operation log data of the target equipment as information related to the target equipment, and the analysis processing execution unit uses the operation log data to analyze the status of the target equipment.

10. The AI ​​agent system according to claim 9, wherein the analysis processing execution unit extracts data from the operation log data for a time period in which an abnormality was detected in the target equipment, and performs analysis of the target equipment using the extracted data for the time period.

11. The AI ​​agent system according to claim 10, wherein the scenario generation unit selects a method to be used for analysis based on the contents of the operation log data.

12. The AI ​​agent system of claim 1 or 2, wherein the equipment agent comprises an order processing execution unit that, when the condition of the target equipment satisfies a first condition based on the results of the analysis, executes an order processing for parts necessary to repair the target equipment.

13. The AI ​​agent system of claim 1 or 2, wherein the equipment agent is provided with a maintenance request execution unit that issues a maintenance request for the target equipment when the state of the target equipment satisfies a second condition based on the results of the analysis.

14. A program for causing a computer to function as an equipment agent that mediates communication between users and equipment by causing the computer to function as: a target equipment information receiving unit that receives information related to the target equipment for which the computer is responsible; an instruction information receiving unit that receives instructions related to the target equipment; a scenario generation unit that creates a response scenario based on the information related to the target equipment and the content of the instructions related to the target equipment; an analysis processing execution unit that analyzes the status of the target equipment based on the response scenario; and an analysis result output unit that outputs the results of the analysis.

Images