Information processing method, program, and information processing system

Abstract

An information processing method, a program, and an information processing system are provided that further improve the accuracy of output results. [Solution] In an information processing system 10 having a server 12 as an information processing device, a user terminal 14, and an agent 16, which are connected to each other so that they can communicate with each other via a network N, the information processing method executed by the server includes an input information acquisition step for acquiring input information, a processing subject setting step for setting a processing subject that performs processing for each of multiple processing steps set to process the input information, and a processing result acquisition step for acquiring and outputting the processing results in the processing steps by the processing subject.

Description

[Technical Field]

[0001] The present invention relates to an information processing method, a program, and an information processing system. [Background technology]

[0002] Cited Document 1 discloses a configuration in which a prompt including an idea for an invention or device and a method for drafting a claim is input into a large-scale language model to generate a first draft claim, and after obtaining an evaluation result, an improved second draft claim is generated. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Application No. 2023-139311 Summary of the Invention [Problem to be solved by the invention]

[0004] In the above prior art, the generation and evaluation of claims are automated using a large-scale language model. However, in highly specialized tasks such as patent specification drafting, the accuracy of the output results may not reach the required level, and there is room for improvement in this regard.

[0005] In consideration of the above, an object of the present invention is to further improve the accuracy of the output results. [Means for solving the problem]

[0006] According to one embodiment of an information processing method, the information processing method is executed by an information processing device and includes an input information acquisition step of acquiring input information, a processing subject setting step of setting a processing subject that performs processing for each of a plurality of processing steps set to process the input information, and a processing result acquisition step of acquiring and outputting processing results in the processing steps by the processing subject.

[0007] According to one embodiment of the program, an information processing device is caused to execute an information processing method including an information acquisition step of acquiring information about a target entity, a query step of generating query information for estimating a trust index of the target entity from the information and presenting the query information to the target entity, and an estimation step of acquiring response information from the target entity in response to the query information and estimating the trust index of the target entity from the response information.

[0008] According to one embodiment of an information processing method, the information processing method is executed by an information processing device, and includes an input unit display step of displaying an input unit that accepts input information on a user interface, a processing process information display step of displaying multiple processing process information set for processing the input information on the user interface, and a processing subject information display step of displaying processing subject information representing a processing subject of the processing process on the user interface.

[0009] According to one embodiment of the program, an information processing device is caused to execute an information processing method including an input information acquisition step of acquiring input information, a processing subject setting step of setting a processing subject that performs processing for each of a plurality of processing steps set to process the input information, and a processing result acquisition step of acquiring and outputting the processing results of the processing steps by the processing subject.

[0010] According to one embodiment of the program, an information processing device is caused to execute an information processing method including an input unit display step of displaying an input unit that accepts input information on a user interface, a processing process information display step of displaying multiple pieces of processing process information set for processing the input information on the user interface, and a processing subject information display step of displaying processing subject information representing a processing subject of the processing process on the user interface.

[0011] According to one embodiment, the information processing system is an information processing system equipped with a plurality of information processing devices, and includes an input information acquisition unit that acquires input information, a processing subject setting unit that sets a processing subject that performs processing for each of a plurality of processing steps set to process the input information, and a processing result acquisition unit that acquires and outputs the processing results of the processing steps by the processing subject.

[0012] According to one embodiment, the information processing system is equipped with a plurality of information processing devices, and includes an input unit display unit that displays an input unit that accepts input of input information on a user interface, a processing process information display unit that displays a plurality of processing process information set for processing the input information on the user interface, and a processing subject information display unit that displays processing subject information representing the processing subject of the processing process on the user interface. [Effects of the Invention]

[0013] According to one embodiment, the accuracy of the output results can be improved. [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a diagram illustrating an example of a configuration of an information processing system according to an embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of a hardware configuration of a server according to an embodiment. [Figure 3] FIG. 2 is a diagram illustrating an example of a functional configuration of a server according to an embodiment. [Figure 4] 10 is a flowchart illustrating an example of processing executed by an information processing system according to an embodiment. [Figure 5] FIG. 1 is a conceptual diagram illustrating an example of access control according to an embodiment. [Figure 6] FIG. 10 is a diagram illustrating an example of a screen display in an information processing system according to an embodiment. [Figure 7] FIG. 10 is a diagram illustrating an example of a screen display in an information processing system according to an embodiment. [Figure 8]FIG. 10 is a diagram illustrating an example of a screen display in an information processing system according to an embodiment. [Figure 9] 9 is an example of a screen display in the information processing system according to an embodiment, showing only the TP section of FIG. 8. FIG. [Figure 10] 9 is an example of a screen display in the information processing system according to an embodiment, showing only the TP section of FIG. 8. FIG. [Figure 11] 9 is an example of a screen display in the information processing system according to an embodiment, showing only the TP section of FIG. 8. FIG. [Figure 12] FIG. 10 is a diagram illustrating an example of a screen display in an information processing system according to an embodiment. [Figure 13] FIG. 10 is a diagram illustrating an example of a screen display in an information processing system according to an embodiment. [Figure 14] FIG. 10 is a diagram illustrating an example of a screen display in an information processing system according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0015] (Embodiment) An embodiment of an information processing system according to the present invention will be described below with reference to Figures 1 to 14. The same or equivalent components and parts in each figure are denoted by the same reference numerals. The dimensional proportions in the drawings are exaggerated for the sake of explanation and may differ from the actual proportions.

[0016] (System Overview) First, an overview of an information processing system 10 according to this embodiment will be described. The information processing system 10 according to this embodiment is configured to load multiple predefined processing steps and execute processing according to the contents of the processing steps. Specifically, a server references pre-prepared processing step definition information and assigns each processing step to a processing entity consisting of at least one of an agent system (hereinafter simply referred to as an "agent"), another information processing system, and a user terminal operated by a user. The results of each processing step are then stored in the system as needed, and correction instructions and progress confirmations are made step by step via the user terminal, thereby efficiently generating results while inducing partial reprocessing as necessary. While this embodiment uses a method of loading pre-defined processing steps as described above, this is not limited to this, and a configuration in which new processing steps are dynamically generated in response to information entered by the user may also be used. Note that an agent is not limited to a passive terminal but refers to an entity that actively responds and acts in response to queries from a server and access control policies, including agents using trained models (so-called AI agents).

[0017] (System Configuration) Fig. 1 is a diagram showing an example of the configuration of an information processing system 10 according to this embodiment. As shown in Fig. 1, the information processing system 10 according to this embodiment includes a server 12 as an information processing device, a user terminal 14, and an agent 16, which are communicably connected to each other via a network N. The network N is, for example, a wired LAN (Local Area Network), a wireless LAN, the Internet, a public line network, a mobile data communication network, or a combination thereof.

[0018] The user terminal 14 is an example of an information processing device that is operated by a user U to input and display various information. The user terminal 14 may be a PC (Personal Computer), a smartphone, a tablet terminal, a server device, a microcomputer, a wearable device, or a combination of these.

[0019] The server 12 is an example of an information processing device that acquires information input from the user terminal 14, processes the information, and outputs the results. The server 12 may be a PC (Personal Computer), a smartphone, a tablet terminal, a server device, a microcomputer, or a combination of these. The server 12 may also be the agent described above. The specific configuration and operation of the server 12 will be described later.

[0020] As described above, in this embodiment, the agent 16 is an element that is deployed separately from the server 12 and the user terminal 14, and is an example of a subject that interacts with the system as a target entity. The agent 16 is characterized by the fact that it can be implemented in a variety of specific forms, from a software-only program form to a physical robot, an IoT device, or even the latest AI agents that utilize large-scale language models (LLMs). Possible configuration examples are shown below, but the present invention is not limited to these.

[0021] (1) Software Agent Examples of such tools include chatbots, script engines, and RPA (Robotic Process Automation) tools that operate over a network. These receive human instructions and processing requests from Server 12, accessing databases as needed and performing additional calculations and analysis. They are easy to scale out, and can be configured to run multiple instances simultaneously as microservices in a cloud environment.

[0022] (2) Robotic agents with hardware In this case, physical devices such as industrial robots on factory lines, home robots, and smart speakers function as agents. Each robot sends its own operation log and sensor information to the server 12, and switches its operating mode depending on the response content and access authority settings from the server 12, thereby realizing advanced cooperative control and dialogue with humans.

[0023] (3) IoT Device / Edge Agent In this configuration, IoT devices equipped with built-in microprocessors and sensors that measure environmental information play the role of agents. These devices transmit measurement data to server 12 via network N, and change their behavior based on the control policies and access permission settings generated by server 12. By combining edge computing, it is expected that load balancing on the server and real-time performance will be improved.

[0024] (4) Virtual agents (digital humans, etc.) Avatars and digital humans active in the metaverse or virtual space can also be agents that access the server 12 through voice recognition and natural language analysis. While interacting with the user U in the virtual space, they can dynamically change the service content and display information while receiving inquiries and access control from the server 12.

[0025] (5) Composite / Hybrid Agents It is also possible to envision a hierarchical structure, such as linking software agents with robotic agents, or deploying a parent agent that oversees multiple IoT devices. For example, in a manufacturing site, a large-scale system could be built in which multiple autonomous robots each have their own security authority, and a higher-level software agent manages a unified access control policy.

[0026] (6) Cutting-edge AI agents utilizing large-scale language models With the recent advances in generative AI technology and large-scale language models (LLMs), agents with advanced natural language processing engines at their core are emerging. These agents dynamically invoke models trained on large amounts of data in the cloud and respond to requests from users U or other agents. By combining automatic integration with external APIs and third-party services, they can autonomously process complex processing flows and are highly adaptable to environmental changes and new information. Compared to traditional agents that rely on rule-based or limited data sets, these agents can execute processes with far greater flexibility and a greater amount of knowledge and reasoning capabilities.

[0027] As described above, the concept of "agent 16" can encompass a wide range of entities, from software-based programs to robots with hardware, lightweight IoT devices, and even large-scale language model-based AI. In this embodiment, all of these are treated as examples of "target entities," making it possible to flexibly provide inquiry information, access permission settings, and collaboration processes that are tailored to the characteristics of each agent. In other words, the term "agent" used here is not limited to individual agent structures or specific implementation methods, but includes all agent forms that operate in large-scale and diverse network environments.

[0028] (Hardware configuration) 2 is a block diagram showing the hardware configuration of the server 12. The server 12 includes a processor 120, a memory 122, a storage 124, a communication I / F 126, an input / output I / F 128, and a drive device 134, which are communicatively connected to each other via a bus B.

[0029] The processor 120 controls each component of the server 12 and realizes the functions of the server 12 by loading various programs stored in the storage 124 into the memory 122 and executing them. The programs executed by the processor 120 include, but are not limited to, an operating system (OS) and a program 220 described below. Execution of these programs by the processor 120 realizes part of the state visualization method according to this embodiment. The processor 120 is, for example, a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), an application specific integrated circuit (ASIC), a digital signal processor (DSP), or a combination thereof.

[0030] The memory 122 is, for example, a read-only memory (ROM), a random access memory (RAM), or a combination thereof. The ROM is, for example, a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a combination thereof. The RAM is, for example, a dynamic random access memory (DRAM), a static random access memory (SRAM), a magnetoresistive random access memory (MRAM), or a combination thereof.

[0031] The storage 124 stores the OS, various programs described below, and various data. The storage 124 is, for example, a flash memory, a hard disk drive (HDD), a solid state drive (SSD), a storage class memory (SCM), or a combination of these.

[0032] The communication I / F 126 is an interface for connecting the server 12 to external devices including the user terminal 14 and the agent 16 via the network N and for controlling communication. The communication I / F 126 is, for example, an adapter compliant with Bluetooth (registered trademark), Wi-Fi (registered trademark), ZigBee (registered trademark), Ethernet (registered trademark), or optical communication (e.g., Fibre Channel), but is not limited to these.

[0033] The input / output I / F 128 is an interface for connecting an input device 132 and an output device 130 to the server 12. The input device 132 is, for example, a mouse, a keyboard, a touch panel, a microphone, a scanner, a camera, various sensors, an operation button, or a combination thereof. The output device 130 is, for example, a display, a projector, a printer, a speaker, a vibrator, or a combination thereof. In other words, the output device 130 is a user interface (hereinafter, sometimes simply referred to as a "user interface"). In this embodiment, a display is used as the user interface, for example.

[0034] The drive device 134 reads and writes data from and to the disk media 136. The drive device 134 is, for example, a magnetic disk drive, an optical disk drive, a magneto-optical disk drive, or a combination thereof. The disk media 136 is, for example, a compact disc (CD), a digital versatile disc (DVD), a floppy disk (FD), a magneto-optical disk (MO), a Blu-ray (registered trademark) disc (BD), or a combination thereof.

[0035] In this embodiment, the program may be written into memory 122 or storage 124 during the manufacturing stage of server 12, may be provided to server 12 via network N, or may be provided to server 12 via a non-transitory computer-readable recording medium such as disk media 136.

[0036] Furthermore, the hardware configuration of the user terminal 14 is substantially the same as the hardware configuration of the server 12 described above, and therefore a detailed description thereof will be omitted.

[0037] (Functional configuration) Next, the functional configuration of the server 12 will be described. Fig. 3 is a diagram showing an example of the functional configuration of the server 12. When executing various programs, the server 12 uses the above-mentioned hardware resources to realize various functions. The server 12 has a storage unit 22, a control unit 24, and a communication unit 20 as the functional configuration realized by the server 12. Each functional configuration is realized by the processor 120 reading and executing a program 220 stored in the memory 122 or the storage 124.

[0038] The communication unit 20 is realized by the communication I / F 126. The communication unit 20 transmits and receives information to and from the user terminal 14 via the network N. The communication unit 20 receives information input from the user terminal 14. The communication unit 20 also transmits information to the user terminal 14 and receives requests from the user U from the user terminal 14.

[0039] The storage unit 22 is realized by a memory 122 and a storage 124. The storage unit 22 stores a program 220, a trained model 222, a processing step information DB 224, a processing subject information DB 226, a processing result DB 228, a history information DB 230, a user information DB 232, and a notification information DB 234.

[0040] The program 220 is read and executed by the processor 120 of the server 12 to perform the functions of implementing various procedures of the information processing method according to this embodiment.

[0041] The trained model 222 is configured to perform advanced natural language analysis and response estimation required for processing on the server 12. During the training phase, expected processing steps and past response logs of processing subjects are prepared as initial data, and parameters within the model are updated based on a loss function that calculates the deviation from each sample labeled with a correct answer. By combining data containing contextual elements and diverse linguistic expressions that are difficult to distinguish, adjustments are made to enable more flexible and accurate context understanding. By incorporating a variety of inputs from actual operations, such as mistranslations and abbreviations, errors from the correct data are gradually discovered and parameters are corrected. During the training phase, the error from samples labeled with a correct answer is repeatedly evaluated by considering the use of operation history accumulated on the server 12 or a general natural language corpus, and internal parameters within the model, including weights and biases, are optimized. During the inference phase, the context analysis parameters obtained during the training phase are referenced, the input text is analyzed, and an appropriate output is estimated. Even when the input contains expressions that are likely to lead to ambiguous judgment, inference is made by utilizing the level of understanding of metaphors and other expressions acquired during training. Furthermore, a configuration in which a large-scale language model (LLM) is incorporated into this system is also envisioned. When introducing the system without fine tuning, inference is performed by referring to existing trained parameters, and operation can be started without the need for additional adjustments. Note that in this embodiment, the trained model 222 is configured to be stored in the storage unit 22, but this is not limiting, and the trained model may also be stored in an external server and used.

[0042] In the information processing system 10 according to this embodiment, the trained model 222 analyzes input information and operation logs in natural language and, as necessary, estimates optimal processing steps and processing details. For example, when a user inputs free text into the input unit 30 on the user interface, the trained model 222 analyzes the context to estimate the intent of the input content and can be used to allocate multiple processing steps to determine which is the optimal flow. Furthermore, when there are multiple agents that are candidates for processing, the trained model 222 can contextually understand the user's instructions and the nature of the data, and automatically suggest the steps that each agent excels at. In other words, the trained model 222 is not merely a response generation engine; it also supports automatic step selection and processing task assignment by comparing the user's input information with the system's internal processing step definitions and the history information DB 230. Furthermore, the trained model 222 may cooperate with the processing result acquisition unit 246 to provide feedback evaluation of the natural language quality of output results (e.g., documents automatically generated by agents, analysis results, etc.). In this case, the trained model 222 references previously stored processing results and user operation history as training data, enabling it to detect typos and contextual inconsistencies with high accuracy and make suggestions for improvement as needed. This feedback function contributes to optimizing the entire process as a workflow, rather than simply mining text.

[0043] Below, a specific example of a process in which the trained model 222 analyzes the context, infers the intent of the input content, and assigns the optimal flow among multiple processing steps will be described. In this embodiment, the trained model 222 has a function of analyzing natural language text input by a user and operation logs (clicks, selections, editing instructions on the UI, etc.) to infer and suggest a processing direction. Specifically, when user input information is received via the communication unit 20 of the server 12, the input information acquisition unit 242 temporarily stores the text and various operation logs in the storage unit 22. Then, in response to a call from the processing subject setting unit 244 or the processing result acquisition unit 246 in the control unit 24, the trained model 222 executes the following steps. First, as a preliminary step to natural language processing, the trained model 222 divides and normalizes the input text into words, phrases, and sentences (morphological analysis, tokenization, unnecessary word removal, etc.). The trained model 222 compares the extracted subelements (e.g., noun phrases or specific keywords) with trained embeddings (e.g., word embeddings or context vectors) to estimate process names and task categories likely related to the user's intent. For example, if the input includes the phrase "I would like to rewrite the specification by referring to the drawings," the trained model 222 extracts keywords such as "drawing," "specification," and "rewrite." It then identifies appropriate process candidates, such as "sub-process editing tasks" and "addendums to the specification," by referencing the similarity map constructed in the pre-training stage and past response logs of processing entities (e.g., correspondence with the processing process information DB 224). Next, the trained model 222 evaluates the context of the entire input sentence using these candidates and scores the priority and relevance. Here, the actual processing entity and process to be switched may be estimated with higher accuracy by taking into account user operation logs (e.g., whether the user had previously opened a drawing display screen or referred to sub-process information) and consistency with previously accumulated operation history. The model uses a neural network method that includes an encoder-decoder structure and an attention mechanism, allowing it to estimate the optimal solution based on the entire context, rather than simply matching keywords.Furthermore, the analysis results obtained by the trained model 222 are fed back to the processing entity setting unit 244 within the control unit 24. Based on these results, the processing result acquisition unit 246 automatically determines whether multiple agents, another information processing system, or the user themselves are appropriate for the task, and uses these results as information for determining the execution order and assignment entity. For example, if "Agent A, who is good at editing drawings," and "Agent B, who is good at proofreading documents," are both candidates, the trained model 222 will propose a process such as "first ask Agent A to check the drawings, and then have Agent B check the consistency of the specifications" based on the user's input text, reference status, and past success stories. In this way, the trained model 222 comprehensively judges various natural language inputs, supplemented by contextual information and operation logs, rather than simply processing strings, and finds the relationship between multiple processing steps and processing entities in real time. This significantly automates and streamlines tasks that previously required users to manually specify all steps. It also reduces agent assignment errors and task omissions, improving the accuracy of final processing results and work speed. The specific model structure and update method of the trained model 222 can be modified in various ways depending on the operational form, such as a configuration that references a large-scale language model held on-premise or on the cloud side as described above and periodically updates parameters customized for this system, or a configuration that combines an API-based LLM provided externally.

[0044] In addition, when there are multiple candidate agent groups, a specific example of a process is described below. The process contextually understands the instructions from user U and the nature of the data, and automatically proposes processes in which each agent excels. For example, in this system, each agent group is predefined with its own specialized field and specialty. For example, Agent A may excel in drawing creation and image processing, Agent B may specialize in document proofreading and editing, and Agent C may be primarily responsible for prior art searches and reference management. This attribute information and past processing history are stored in the processing entity information DB 226 and are also used as reference data when comparing agents. For example, if User U inputs, "I would like to make some adjustments to the structure of the drawing, but I would also like to rewrite the main text of the specification to make it consistent," the trained model 222 first analyzes the input statements and operation logs. If keywords like "drawing configuration," "rewrite," and "consistency" appear and past history infers that "User U seems to be currently placing importance on drawing layout," Agent A, who excels at drawing-related tasks, and Agent B, who can make detailed corrections to the specification text, emerge as likely candidates. However, Agent C, while skilled at search and reference support, does not immediately fulfill this request, so it is likely appropriate to remove him from the list of primary candidates. Next, trained model 222 further examines Agent A's and Agent B's areas of expertise and past success stories, breaks down the current task into process steps, and presents a procedure in which "Agent A will first revise the drawings, and once that work is complete, Agent B will proofread the text." If User U's input statement includes a time constraint, such as "urgent," trained model 222 can automatically determine which task to proceed with first or whether tasks can be carried out simultaneously, taking into account information previously held by each agent, such as their response speed and resource status.If user U additionally instructs the system to "finish the text more elegantly," the trained model 222 reanalyzes the request contained in the input text and, if it determines that a "polishing of the writing" process is necessary, determines that not only agent B but also agent D, for example, who is skilled in a different language, may be appropriate. This embodies an approach that differentiates agents based on whether the same "rewriting" task requires stylistic correction or legal consistency check. Ultimately, the processing subject setting unit 244 in the control unit 24 automatically generates or proposes a series of process flows to user U based on the results of this model analysis. Once the user agrees to the proposal, the processing result acquisition unit 246 can initiate the actual task. As described above, even when there are multiple candidate agents, the trained model 222 can comprehensively compare user U's request with each agent's area of ​​expertise and automatically propose the most practical and highly specialized process assignment. .

[0045] Furthermore, when change instruction information is provided, the trained model 222 may compare existing processing results and the user's intent to suggest which process should be re-executed or which parameters should be revised. For example, when regenerating independent or dependent claims or modifying the decision logic for process branching, the trained model 222 complements the context with information such as "how much the sentence structure changes before and after the modification" and "what the user's objective is," thereby helping the user smoothly carry out complex editing tasks. In this way, the trained model 222 comprehensively analyzes various input information, including natural language, in cooperation with each component of the information processing system 10 (e.g., the processing process information DB 224 and the processing result DB 228), contributing to the system in a wide range of aspects, such as automatic allocation of sub-processes and processing entities, improved quality of results, and overall workflow efficiency. In this embodiment, the trained model 222 is stored and operated in the storage unit 22. However, it may also be connected by calling a large-scale language model stored on an external server, etc., and referencing and utilizing pre-trained parameters.

[0046] The following is a specific example of a process in which, when change instruction information is given, the trained model 222 compares existing processing results with the user's intentions to determine which specific steps should be re-executed or which parameters should be modified, and provides feedback to the system. When user U inputs change instruction information expressing some dissatisfaction or additional request (e.g., "I want to re-summarize the whole thing more succinctly" or "The writing style is too stiff, so please use simpler expressions"), the trained model 222 compares the instruction content with existing processing results. Here, "existing processing results" refers to past output content and process execution logs stored in the processing result DB 228 and history information DB 230. By comparing these with the previous results, the model identifies which steps should be re-executed to meet the user's request, or whether simply adjusting specific parameters will suffice. For example, if a request such as "The sentence is too long, please summarize it," the trained model 222 references the previous processing content (document editing process) and determines that the sentence is likely to be shortened by re-invoking "Agent B, which was responsible for style conversion and paragraph restructuring." On the other hand, if the request is along the lines of "The drawing layout is incorrect, so I would like to rearrange it," the model compares the results with those of the previous drawing editing process, narrows down the problem, and concludes that reprocessing by "Agent A" is appropriate. Furthermore, if there are multiple parameter groups used in the process (e.g., the degree of compression of the summary or the drawing resolution setting), the trained model 222 quantifies and attributes expressions such as "more concise" and "higher resolution" contained in the user instructions and resets the corresponding parameters. For example, if a request is made to "make the logical structure clearer," the model suggests switching the "logical development emphasis option on" as a parameter for the prompt or editing engine. If the suggestion is approved and executed within the control unit 24, only the corresponding process will be automatically re-executed. When making such parameter changes, the model also references historical information on similar requests made in the past and the response logs of the processing entity, giving priority to agents with a proven track record and settings that are known to be highly effective for specific parameters.As a result, the trained model 222 contextually compares the change instruction information itself with the differences in the previous deliverables, presents a method to meet the request with minimal reprocessing, and induces re-execution through the control unit 24 and the processing subject setting unit 244. If the user checks this on the screen and approves the proposal, only the relevant process or parameters are changed and recalculated, making corrections more efficient than restarting the entire process from the beginning. The change instruction analysis function of the trained model 222 leads this flow of partial re-execution and parameter resetting.

[0047] The processing step information DB 224 is a database that stores information about multiple processing steps. It manages the content and progress status of each processing step and is referenced and updated when assigning a processing entity for each processing step. These processing steps include at least one of those read from a pre-prepared configuration file (e.g., a YAML file) and those dynamically generated by the trained model 222 based on input information. In this embodiment, processing step information 32, which is information about processing steps, includes the processing step name, sub-process information 42, which is information about sub-processes, and the processing results of the processing steps and sub-processes. In addition, the processing step information 32 may also include the following: For example, the processing step information 32 may include a link to information about potential personnel registered in the processing entity information DB 226 or an actual assignment history. This clarifies which personnel or agents were assigned to each process and at what point in time, facilitating tracking of change instructions and scope of responsibility. Dependencies between processes may also be included. For example, by defining cases where a process cannot begin unless another process is completed, or by defining cases where multiple processes can proceed in parallel, the entire workflow can be visualized in a unified manner, making it easier to avoid confusion when editing or reassigning processes. Furthermore, information on deadlines and priorities assigned to each task can be included. Even for specialized tasks based on special keys, a bird's-eye view of the tasks makes it easier to understand which process should be completed first, and a system can be incorporated to prompt authorized personnel via notifications, if necessary. Information on process branching conditions and input parameters can also be included. When a user or AI agent proposes additional branching conditions or when dealing with a task that branches into multiple routes, consolidating information to facilitate process branching will facilitate subsequent changes and reprocessing. Furthermore, editing history and version control information can be included on a process-by-process basis. This allows users to try multiple versions and derive the optimal configuration, ultimately improving overall processing accuracy and convenience.

[0048] The processing subject information DB 226 is a database that holds attribute information of processing subjects (target information, area of ​​responsibility, role, authority, etc.) and is used to manage the processing subject switching function and processing subject information displayed on the user interface.

[0049] The processing result DB 228 is a database that stores the output results acquired in the processing result acquisition step and the results acquired again based on the change instructions. By storing these output results (including intermediate results), the processing results can be reused even after the system is restarted.

[0050] The history information DB 230 is a database that stores output results before and after change instructions for processing steps and the history of editing operations on the user interface. It is referenced when handling branching processing, comparison processing, and differences during screen editing, which will be described later.

[0051] The user information DB 232 is a database that stores authentication information for each client and settings for each user, and is used to determine user authority and customize display content in relation to the display of the user interface.

[0052] The notification information DB 234 is a database for managing notification information, which will be described later. By storing text and addresses related to progress and changes in the processing steps, the notification information DB 234 is used to send appropriate notifications to processing entities such as users via the user terminal 14.

[0053] The control unit 24 is realized by the processor 120 reading and executing the program 220 from the memory 122 and working in cooperation with other hardware components. In this embodiment, the control unit 24 includes an input information acquisition unit 242, a processing subject setting unit 244, a processing result acquisition unit 246, a user interface control unit 248, and an output unit 250.

[0054] The input information acquisition unit 242 performs a function corresponding to the "input information acquisition step of acquiring input information" in claim 1, and in a system equipped with multiple information processing devices, also corresponds to the "input information acquisition unit for acquiring input information" in claim 15. By accepting input information transmitted from the user U or another system, etc., the input information acquisition unit 242 acquires data necessary for processing and passes it to the subsequent process. The input information may be acquired through an operation from the user terminal 14 or an external service, etc., and the acquired content is temporarily stored in the storage unit 22 and held in a form that can be referenced by the processing subject setting unit 244, the processing result acquisition unit 246, etc.

[0055] The processing subject setting unit 244 implements the step of "setting a processing subject that performs processing for each of a plurality of processing steps" as set forth in claim 1, and in terms of system configuration corresponds to the "processing subject setting unit that sets a processing subject" as set forth in claim 15. It has the function of assigning at least one of the processing subjects in a plurality of processing steps (or sub-steps), and also includes a configuration for acquiring change instruction information as set forth in claim 2 and switching the processing subject. Change instruction information is notified from a user or another step, and upon receiving this information, the processing subject is dynamically updated, keeping the allocation status always up to date.

[0056] The processing result acquisition unit 246 implements the step of "acquiring and outputting the processing results in the processing step by the processing subject" as set forth in claim 1, and in terms of system configuration corresponds to the "processing result acquisition unit that acquires and outputs the processing results" as set forth in claim 15. As set forth in claim 3, it has the function of storing the processing results in the processing result DB 228 in the storage unit 22 as storage, and utilizing the stored results even after restart. When change instruction information as set forth in claim 4 is received, it is also configured to acquire the processing results again as a branch of a sub-step, and as set forth in claim 5, compare the processing results before and after the change and output both. This realizes a flow for performing follow-up corrections and reprocessing on results obtained once, improving the flexibility of the entire process.

[0057] The user interface control unit 248 is a functional unit that collectively controls user interface display, including the "input unit display step," "processing step information display step," and "processing subject information display step" recited in claim 6. That is, in screen linkage with the user terminal 14, it provides an input screen where the user inputs (see "user interface" in FIG. 5, input unit display step in claim 6), a screen that displays a list of multiple processing step information 32 (processing step names, task names, overviews, etc.) (see FIG. 7, processing step information display step), and a screen that displays processing subject information 34 representing processing subjects (processing subject information display step). Furthermore, it also includes a function for displaying processing progress status information 38 recited in claim 7, a processing monitor information display function that displays processing monitor information 40 recited in claim 8, and a sub-process information display function that displays sub-process information 42 recited in claim 9, and is equipped with screen control logic for displaying each of these functions.

[0058] The user interface control unit 248 also includes a configuration for executing a "processing process edit display step for displaying a processing process edit screen" as recited in claim 10. A processing process edit screen 50 (see FIG. 9; in the processing process edit screen 50 in FIG. 9, the processing results of sub-processes in the processing process information 32 are edited) is generated and presented so that the user can edit the contents of the processing process, allowing various processing process information, including sub-processes, to be changed. Therefore, in accordance with the "processing process branch display step" as recited in claim 11, it is possible to implement a flow in which edited processing process information and unedited processing process information are displayed together on the user interface (see comparison screen 52 in FIG. 10) to visualize the differences between the changed portions. Furthermore, the user interface control unit 248 also centrally manages the step of displaying a notification screen 36 (see FIG. 12) showing notification information as recited in claim 12, and displays an appropriate screen to the user U depending on the processing status.

[0059] The multiple user interface-related steps defined in claims 6 to 12, such as the input unit display step and progress status information display step, are dynamically switched on the display screen of the user terminal 14. When the user U performs an editing operation or instructs reprocessing, the execution flow is updated in conjunction with the processing subject setting unit 244 and processing result acquisition unit 246 in the control unit 24, and the updated contents are immediately notified to the user interface control unit 248. Such real-time communication and event-driven processing realize a configuration that allows the user U to proceed with operations flexibly while overseeing multiple processing steps.

[0060] The output unit 250 presents the engine-side processing results of claims 1 to 5 to the outside, and also functions to integrate the final output to be reflected in the user interface-side screen display results of claims 6 to 12. The results acquired by the processing result acquisition unit 246 are presented to the user via the user interface control unit 248, and can also be configured to be written out as a separate system or file. By performing overall management of both the engine side and the user interface side, it is designed to allow the user to receive information in the format they ultimately require.

[0061] In this embodiment, software for executing the above steps or display functions can be installed in an information processing device as the program claims set forth in claims 13 and 14. That is, the same effect can be achieved by having the information processing device read and execute a program that performs processes such as the input information acquisition step, processing subject setting step, and processing result acquisition step, while also including the input unit display step and processing process information display step of the screen display system. The information processing system set forth in claims 15 and 16 also includes a configuration in which multiple information processing devices are combined via a network, each of which shares and executes part or all of the present configuration. This also includes cases in which the functions of the control unit 24 are divided and executed among multiple physical devices, and cases in which the AI ​​agent and the user terminal 14 cooperate to carry out the same processing steps.

[0062] (Processing performed by information processing system 10)

[0063] Next, the operation of the information processing system 10 will be described. Fig. 4 is a flowchart showing an example of the flow of processing by the information processing system 10. The processor 120 reads out the program 220 stored in the storage 124, expands it into the memory 122, and executes it, thereby sequentially executing the following processing. Although not shown, when the processor 120 receives operation information to terminate the operation of the information processing system 10, or information on the termination of an operation from the user terminal 14 during the ongoing determination processing (these will be simply referred to as "termination operations"), the processor 120 terminates the series of processing based on the program 220 being processed.

[0064] The processor 120 first displays the input section 30 on the user interface (step S100, see FIG. 6). This step S100 corresponds to the "input section display step of displaying the input section on the user interface" set forth in claim 6. In this embodiment, it is conceivable to use text boxes, check boxes, etc. to make it easier for the user to input input information such as ideas and parameters. This ensures operability via the user terminal 14 and provides an environment in which the user can freely input information.

[0065] Next, processor 120 checks whether input information has actually been entered (step S102). If input information has been entered (step S102: YES), the process proceeds to step S104, which will be described later. On the other hand, if input information has not yet been obtained (step S102: NO), the process returns to step S100, and the input unit continues to be displayed. In this embodiment, it is assumed that the user's operation is not complete or that input has been temporarily interrupted, and this loop is repeated until the required input is confirmed.

[0066] Once the input information is confirmed, the processor 120 acquires the input information using the input information acquisition unit (step S104). This step S104 corresponds to the "input information acquisition step of acquiring input information" referred to in claim 1. The input information may include text or file information entered by the user, or setting parameters received from other devices. The acquired content is temporarily stored in the memory unit 22 and used as information for determining the subsequent processing step settings and processor allocation.

[0067] Next, the processor 120 sets a processing step and a processing entity (step S106). This step S106 corresponds to the "processing entity setting step for setting a processing entity" in claim 1 and can also respond to a processing entity change instruction as required in claim 2. In this embodiment, processing steps are managed as "tasks," and processing entities are assigned to user U, agents, etc. based on the task content and difficulty. The assigned setting information serves as the basis for the subsequent processing execution flow, and appropriate responsibilities are determined automatically, semi-automatically, or manually by user U, etc. Note that automatic assignment does not necessarily require prior operation or confirmation by user U. For example, the system autonomously determines the optimal responsibilities by comprehensively referencing task information, the operating status of the person in charge, the agent's ability profile, and available resources in other information processing systems. For example, when a highly urgent task occurs, the system can automatically evaluate predefined priority conditions to assign an available agent, or prioritize user U himself based on the results of an expertise survey. This allows the assignment process to be completed without waiting for a request from user U. By introducing this automatic mode, the system can quickly and appropriately select the appropriate person in charge and progress the task, even in situations where user U cannot intervene every time. Furthermore, the following examples are considered for semi-automatic allocation. First, the system presents the most promising candidate based on task and processing entity information, and user U approves or fine-tunes it. Second, the system refers to user U's priority rules and organizational policies, presents a preliminary allocation, and then leaves the final decision to user U. Third, the system creates multiple candidate plans in parallel, and user U selects the best allocation plan from among them. In either case, user U can make final adjustments, allowing for flexible switching of processing entities for each process. It is also possible to use a processing entity allocation method preset by user U. For example, rules can be saved in advance, such as giving first priority to agents for tasks related to a specific field and having user U handle important or difficult tasks directly.The system automatically assigns tasks based on these rules at the start of the task, allowing the work to proceed immediately without a confirmation phase.User U can also add or change assignments at any time as needed, allowing the system to flexibly deal with changes in the situation or unexpected loads.

[0068] After setting the processing steps, the processor 120 displays multiple pieces of processing step information 32 on the user interface (step S108, see FIG. 7 ). This step S108 corresponds to the “processing step information display step of displaying multiple pieces of processing step information” mentioned in claim 6. The user can now visually understand how the information they entered has been organized into tasks, and the dependencies and priorities of each task. Furthermore, in the following step S110, processing subject information 34 is displayed on the screen. This step S110 corresponds to the “processing subject information display step of displaying processing subject information on the user interface” mentioned in claim 6, allowing the user to check the attributes and status of the person in charge or the AI ​​agent. Note that, although FIG. 7 uses the expression “you” to refer to the user U and the expression “patent specification drafting support agent” to refer to the agent, the processing subject information 34 is not limited to these expressions.

[0069] The processor 120 then determines whether a notification is necessary (step S112). If a message needs to be sent to the user terminal 14 or an alert needs to be sent to the person in charge (step S112: YES), the notification screen 36 is displayed (step S114, see FIG. 12), and the implementation corresponds to the "notification display step of displaying notification information" of claim 12. For example, it is assumed that an emergency message or a confirmation request is notified when the user U is set as the processing subject and processing by the user U is required, when the user leaves unconfirmed input information, when the agent detects an abnormality exceeding a threshold, or when new approval is required for switching between processes or branching processes. On the other hand, if a notification is not necessary (step S112: NO), the processing process is actually executed (step S116). This processing execution is also related to the previous stage of the "processing result acquisition step of acquiring and outputting the processing result" included in claim 1, and it is assumed that multiple processes are performed in parallel or in stages.

[0070] The processor 120 sequentially tracks the progress of the processing and acquires the processing status (step S118). Next, the processor 120 displays processing progress status information 38 (step S120, see FIG. 7). Step S120 corresponds to a "progress status information display step for displaying processing progress status information" as defined in claim 7, and visualizes the extent to which a task has been completed, the next scheduled task, and so on. In this embodiment, the processing progress status information 38 uses expressions such as "2 / 10," which indicates the number of completed tasks (numerator) out of the total number of substeps (denominator) in the processing step, and expressions such as "Patent specification drafting support agent is running: Browser searching..." and "Waiting for operation," which express the specific progress of the processing in natural language. However, the expressions of the processing progress status information 38 are not limited to these. The processor 120 also displays processing monitor information 40 (step S122). This step S122 corresponds to the "processing monitor information display step of displaying processing monitor information" referred to in claim 8, and enables multifaceted monitoring by displaying processing monitor information 40 that shows what calculations and analyses are being performed within the system, using at least one of text data, graphs, and numerical information.

[0071] Next, the processor 120 acquires the processing results (step S124). This step S124 corresponds to the "step of acquiring and outputting the processing results" in claim 1, and also implies the "storing in a storage unit" related to claim 3. Furthermore, the processor 120 executes a screen display of the sub-process information 42 when a sub-process exists (step S126, see FIG. 8), which corresponds to the "sub-process information display step of displaying the sub-process information" in claim 9. Here, the user can get an overview of what sub-processes are set for the processing process, including the progress status of each sub-process. Although not shown, if there are no sub-processes or if no operation is performed to display the sub-process information 42, step S126 is skipped.

[0072] The processor 120 determines whether editing is necessary (step S128). If editing is necessary (step S128: YES), the processor 120 acquires the changes (step S130) and redisplays the processing subject (step S132; the display of the processing subject information 34 in this case is not shown, but is similar to the display of the processing subject information 34 shown in FIGS. 7 to 11). In this embodiment, an implementation example is assumed in which the user U switches to edit mode when an on-screen edit button (e.g., edit button 44 in FIG. 10 ) or menu operation is performed. Editing is performed, for example, when an instruction to add or delete a processing subject is entered, a request to change the process name is entered, or a change to the processing result is desired. As an example, in FIG. 9 , a processing step edit screen 50 including a text box 48 reflecting the processing result of the sub-process is displayed for the processing result of the sub-process, and the processing result can be changed through the processing step edit screen 50. 9 illustrates a processing step editing screen 50 for editing the processing results of a sub-step. However, the processing step editing screen 50 may be used for editing the sub-step information 42 or the processing step information 32. Furthermore, when a discrepancy between a processing step being executed and another processing step is detected, or when an alert is issued indicating that an issue has arisen in cooperation with an external system and that the entire workflow should be reorganized, the system may automatically recommend transitioning to an edit mode that displays the processing step editing screen 50, which allows editing of the target portion. As a result, the processor 120 determines that editing is necessary when the corresponding instruction or detection information is obtained, and prompts the user U to transition to the processing step editing screen 50 and execute editing. The above processing corresponds to a "processing step edit display step for displaying the processing step editing screen" in claim 10. Furthermore, the system again checks whether notification is necessary (step S134). If notification is required (step S134: YES), a notification screen is displayed (step S136) to notify the relevant parties, such as the processing entity and user U, of the latest information on edits to the processing step, sub-process, etc., changes to the processing entity, etc. If notification is not required (step S134: NO), the processing step processing is re-executed (step S138).

[0073] The processor 120 determines whether a comparison is necessary (step S140). If a comparison is necessary (step S140: YES), a comparison screen 52 (see FIG. 10) is displayed, allowing the user to check the differences by displaying both pre-edit and post-edit processing information, and then the process proceeds to the next step, storage processing. Step S140 corresponds to the "processing step branch display step" of claim 11. On the other hand, if a comparison is not necessary (step S140: NO), the processing results are immediately stored (step S144). This storage process corresponds to the "storing the processing results in the storage unit" of claim 3, and is a mechanism for accumulating intermediate results in preparation for later reuse or system restart. In this embodiment, the comparison screen 52 is provided with a display change difference button 56 and a branch destination move button 58 that instructs the user to execute processing using the edited content. When the display change difference button 56 is operated, the processor 120 (user interface control unit 248) executes processing to highlight the edited portions. This makes it easier for the user U to recognize the edited portions, even if the edited portions are diverse. Furthermore, when the operation of the branch destination movement button 58 is accepted, the processor 120 (input information acquisition unit 242) processes the edited information as new input information. At that time, as shown in Fig. 11, information indicating that the corresponding processing step information has been edited or information indicating that new processing is being performed is displayed.

[0074] Next, processor 120 checks whether there are any further processing steps (step S146). If there are further processing steps (step S146: YES), the process returns to step S112 and repeats this process while controlling whether notification is necessary and the execution procedure. On the other hand, if there are no further processing steps (step S146: NO), the results of each step collected so far are integrated (step S148). In this embodiment, the integrated result corresponds to the specification draft. Then, processor 120 outputs the final result as shown in FIG. 14 (step S150). In step S150, the result can be provided in various forms, such as not only a screen display on the display of user terminal 14 but also transmission to an external system or file output, thereby completing a series of information processing. Note that, for convenience of explanation, the above-mentioned steps are described in chronological order. However, in reality, it is assumed that the steps are executed almost instantaneously using high-performance hardware resources, or that any step may be executed at any timing or in a distributed manner among multiple information processing devices.

[0075] (Effects of the embodiment) The information processing system 10 according to this embodiment is an information processing method executed by an information processing device, and includes an input information acquisition step for acquiring input information, a processing entity setting step for setting a processing entity for each of multiple processing steps configured to process the input information, and a processing result acquisition step for acquiring and outputting the processing results of the processing steps performed by the processing entities. Based on this, while sorting and processing the input information on a processing-step basis, it is possible to assign the appropriate agent, another information processing system, or at least one of the user terminals operated by the user U (i.e., processing entities) as the processing entity at that time, thereby reducing overlapping work between processes and inconsistent assignments. Even if the assigned entity is not the user, it is possible to manage which process is currently being performed by which person. This allows the user to refer to the intermediate processing status and output of each process and, if the output does not meet the required level, immediately reassign or adjust parameters. For example, even if an agent or another information processing system produces poor results, the clear assignment of the responsible party allows the user to smoothly correct the course of action, such as issuing an instruction to partially re-execute only that process. By repeating this step-by-step feedback and correction, defects in the output can be identified and quickly improved. This has the effect of further increasing the accuracy of the final output result. Note that, although the present embodiment has been described as an example of generating a patent specification draft, the present invention is not limited to this, and other processes may also be performed.

[0076] As shown in FIG. 5 , the information processing system 10 assigns a processing entity to each processing step (corresponding to 1 through 6 in FIG. 5 ) to perform processing. Therefore, tuning and modification of the processing entity can be performed independently of the information processing system 10. In other words, because the design flexibility of the processing entities is guaranteed, there is no need to significantly change the configuration of the information processing system 10 itself when reviewing or improving the design policy. Furthermore, the information processing system 10 according to the present invention can adaptively update or introduce additional functions by appropriately reviewing the processing entities assigned to specific processing steps. For example, even when migrating an agent to an optimal algorithm or switching the person in charge midway through a process to utilize user U's specialized skills, this can be done without affecting the core components of the system. Furthermore, by storing processing entity assignment information and processing results in a storage unit, flexible version management can be performed while referring to past process execution statuses and branching history. This allows for efficient understanding of results and improvements compared to the previous version, even when replacing an existing processing entity, which is advantageous for repeated verification and optimization. Furthermore, in multiple processing steps including sub-steps, it is possible to visualize and manage not only specific deliverables but also the means and process for generating the deliverables, thereby enabling highly convenient operation for at least one of the developers and the operations staff.

[0077] Furthermore, the information processing system 10 of the present invention can handle not only simple agent-based processing but also an autonomous agent system (e.g., an Agent Dispatcher using MCP+AI Gateway) equipped with an orchestrator that manages multiple workers as the processing entity. Specifically, by assigning at least one of the Agent Dispatcher, LLM Dispatcher, and Service Dispatcher as the processing entity, multiple specialized functions can be utilized in parallel or in stages, while their coordination status and execution results can be consistently managed on the information processing system 10 of the present invention. This configuration enables flexible updates without requiring major changes to the overall configuration of the information processing system 10, even when expanding the functionality or replacing algorithms within various dispatchers. For example, even if you want to replace the LLM Dispatcher with a new model optimized for advanced natural language processing or integrate the Service Dispatcher with new external APIs, you can achieve these changes while maintaining the basic framework of the present invention. In other words, the ability to simultaneously or selectively utilize multiple workers to achieve advanced task processing by leveraging their respective characteristics is one of the technical advantages of the present invention.

[0078] Furthermore, the information processing system 10 according to this embodiment acquires change instruction information for the processing entity of the set processing step in the processing entity setting step, and executes an information processing method for changing the processing entity based on the change instruction information. Based on this, it becomes possible for user U to carefully examine the output content during the process and reassign the work to another agent, another information processing system, or at least one of the user terminals. As a result, the processing entity can be flexibly switched depending on the difficulty of the task, the load situation, and the expertise of other users operating the user terminal 14, allowing work to be continued while minimizing deadlocks and malfunctions. This has the effect of further improving the accuracy of the output results.

[0079] Furthermore, the information processing system 10 according to this embodiment executes an information processing method in which, in the processing result acquisition step, the processing results are stored in a storage unit and then the stored data is reused depending on the progress and status of the processing steps. Based on this, even if the output generated by an agent or other information processing system is insufficient, the intermediate results can be retained and restarted or re-verified easily. In particular, even if operation stops due to a timeout or if the user U accesses the system from a different terminal, the results stored in the storage unit can be utilized, allowing for smooth continuation of processing from the interrupted point. Furthermore, this embodiment is characterized by the fact that, by combining the aforementioned switching function for the person in charge (processor) and the reprocessing and comparison processing described below, more flexible and detailed reprocessing can be realized than simply saving data in the storage unit. In other words, the user U can continue processing by making corrections and amendments to each subdivided step without having to redo the entire process, thereby steadily improving the accuracy of the results and ensuring reusability. This goes beyond the traditional method of recording a single document and saving it all at once, and enables gradual reprocessing in conjunction with change instructions at the process level and switching between multiple people in charge, providing continuous opportunities to improve accuracy even in highly specialized work.

[0080] Furthermore, the information processing system 10 according to this embodiment executes an information processing method in which, when change instruction information for the output processing result is acquired in the processing result acquisition step, the system reacquires the results of the processing step based on the instruction information. Based on this, it is possible to immediately reflect additional requests or corrections to the results that the user U considered to be completed, and perform reprocessing. For example, if you want to replace only a portion of a document generated by an agent, or if you want to rerun a calculation using more accurate parameters, you can simply re-call the relevant step, eliminating the need to recalculate the entire process unnecessarily and maintaining efficient work. This has the effect of further improving the accuracy of the output results.

[0081] Furthermore, the information processing system 10 according to this embodiment executes an information processing method that acquires and outputs both the results of a newly acquired processing step based on change instruction information and the results of the processing step before the change. This allows the user U to make a final decision while comparing the results before and after the change. For example, it can be used in situations where two significantly different proposals are simultaneously displayed and compared for accuracy and completeness from a professional perspective. As a result, the benefits and trade-offs associated with the changes can be immediately understood, making it easier to select the best course of action and improve accuracy. This has the effect of further improving the accuracy of the output results.

[0082] Furthermore, the information processing system 10 according to this embodiment is an information processing method executed by an information processing device, and includes an input unit display step of displaying an input unit 30 on a user interface, a processing step information display step of displaying multiple processing step information 32, and a processing subject information display step of displaying processing subject information 34 for the processing step. Based on this, the user U can input the required information using an appropriate form or options, and the entire process that has been set and the assigned person (AI agent, another information processing system, or the user themselves) are visualized in real time. As a result, work can be carried out while reducing the risk of making mistakes in the operation setup or the request destination, and deficiencies can be easily corrected before the work is passed on to the subsequent process. This has the effect of further improving the accuracy of the output results.

[0083] Furthermore, the information processing system 10 according to this embodiment executes a progress status information display step that displays the progress status of the processing steps on a user interface, allowing the user to check the current progress stage for each step. Based on this, it becomes possible to determine at a glance at which step a process is stalled or whether it is partially completed, enabling early action to be taken even if a stall or problem occurs. As a result, all steps can be carried out efficiently while further minimizing rework, thereby reducing risks to final quality. This has the effect of further improving the accuracy of the output results.

[0084] Furthermore, the information processing system 10 according to this embodiment executes a process monitor information display step of displaying process monitor information 40, which visualizes the process according to the input information, on the user interface. Based on this, it is possible to check in real time the inference and calculation status being performed by the server or agent, or the process flow being performed by other information processing systems, etc. Using the visualized information, the user U can immediately adjust the setting parameters if a bug or inefficiency is discovered during the process. This has the effect of further optimizing the overall operation while improving the accuracy of the output results.

[0085] Furthermore, the information processing system 10 according to this embodiment executes a sub-process information display step that displays sub-process information 42 in the processing process on the user interface. This provides a visualization function that not only visualizes the main processes, but also their derivative work and sub-tasks. As a result, the overall process structure becomes easier to grasp, and output accuracy can be improved by the user inspecting the progress and results of each path in detail. This has the effect of more reliably improving the completeness of the deliverable.

[0086] Furthermore, the information processing system 10 according to this embodiment executes a processing step edit display step that displays a processing step edit screen 50 for editing processing step information. Based on this, the user U can add or remove tasks, change the order of steps, etc., by operating the screen, enabling rapid redesign of the flow according to the situation. By making more appropriate process allocations and adding sub-processes based on the intermediate results, it becomes easier to eliminate shortcomings in the quality of the output. This has the effect of further improving the completeness of the deliverable.

[0087] Furthermore, the information processing system 10 according to this embodiment executes a process branch display step that displays process information after the change on the process editing screen and process information before the change on the user interface. Based on this, when the process structure is changed, the difference before and after editing can be clearly visualized. The user U can see at a glance which parts have been added, deleted, or rearranged, which is extremely useful for reconsidering the work policy. This has the effect of predicting the effect of corrections and more reliably improving overall accuracy.

[0088] Furthermore, the information processing system 10 according to this embodiment executes a notification display step of displaying a notification screen 36 showing notification information for the processing entity of the processing process. Based on this, alerts and notices can be displayed to the user U at important times in the process progress. For example, when the user needs to approve the response result from the agent, or when an error or inquiry occurs in a task assigned to another information processing system, an appropriate notification can be sent promptly to prompt a response. This further improves cooperation throughout the entire process, resulting in an effect of improving the quality of the final result.

[0089] Furthermore, the information processing system 10 according to this embodiment includes a program for causing an information processing device to execute an information processing method that includes an input information acquisition step, a processing subject setting step for setting multiple processing steps and assigning processing subjects, and a processing result acquisition step. Based on this, simply introducing the software into one or multiple device configurations can achieve the effect of implementing the process management and output accuracy improvement described in claim 1. Not only does it provide hardware, but the program that runs on it allows the user U to take the lead in gradual allocation and output correction, making it easier to obtain high-quality results. This has the effect of further improving the accuracy of results even in tasks that require specialized knowledge.

[0090] Furthermore, the information processing system 10 according to this embodiment includes a program for causing an information processing device to execute an information processing method including an input unit display step for displaying an input unit, a processing step information display step for displaying processing step information, and a processing subject information display step for displaying processing subject information. Based on this, it is possible to obtain an effect that display functions related to the user interface can also be introduced as an integrated software set. By making it easier to understand the complicated workflow, the user U can accurately operate input and task assignment, thereby achieving the effect of further improving output accuracy.

[0091] Furthermore, the information processing system 10 according to this embodiment is equipped with multiple information processing devices and has a system configuration including an input information acquisition unit, a processing entity setting unit that assigns processing entities across multiple processing steps, and a processing result acquisition unit that acquires and outputs processing results. Based on this, multiple devices can work together over a network to share tasks, resulting in enhanced process management. In particular, even in cases where collaborative work transcends geographical constraints, such as when experts and agents in remote locations work together, it is easy to ensure accuracy and task management. This has the effect of further improving the quality of output results.

[0092] Furthermore, the information processing system 10 according to this embodiment is equipped with multiple information processing devices and has a system configuration including an input unit display unit that displays the input unit, a processing step information display unit that displays information on multiple steps, and a processing entity information display unit that displays processing entity information. Based on this, even if multiple terminals or servers are distributed, the entire process and the assigned destinations can be displayed synchronously on each user interface. Since the user U can efficiently operate the system by changing the display terminal depending on their usage environment, feedback can be quickly reflected, reducing errors and improving the accuracy of the final deliverable.

[0093] In the above embodiment, an example is shown in which multiple steps, such as the "input information acquisition step of acquiring input information" recited in claim 1, etc., are mainly executed by the server 12, but this is not limiting. For example, a configuration may be adopted in which the user terminal 14 acquires input information itself and transmits it to the server 12, and the input information acquisition step is executed by the user terminal 14. Also, some or all of the processing subject setting step, processing result acquisition step, etc. may be executed by the agent 16 or another information processing system. The present invention also includes a form in which multiple information processing devices share roles and perform distributed processing of each step recited in the claims, thereby enabling a flexible system configuration to be realized as needed.

[0094] Conversely, it is also possible to configure the server 12 (information processing device) to execute each of the above steps by itself, completing all of the steps, such as obtaining input information, setting the processing subject, and obtaining and outputting processing results. For example, this also includes a configuration in which the user interface screen operated by the user is displayed on a display and input device connected to the server 12 itself, thereby performing processing without providing a separate user terminal 14 or agent 16. In this way, even if the functions set forth in claims 1 to 16 are implemented together on a single physical device, this does not depart from the technical scope of the present invention. In this case, the invention may be, for example, as follows: "An input information acquisition unit that acquires input information; a processing entity setting unit that sets a processing entity that performs processing for each of a plurality of processing steps that are set to process the input information; acquiring a processing result by acquiring and outputting a processing result in the processing step by the processing subject; An information processing device having the above. Furthermore, the invention in which processing is performed by an information processing device alone can also be described as follows. "An input unit display unit that displays an input unit that accepts input of input information on a user interface; a processing step information display unit that displays, on the user interface, a plurality of processing step information sets for processing the input information; a processing subject information display unit that displays processing subject information representing a processing subject of the processing step on the user interface; An information processing device having the above.

[0095] Furthermore, in the above-described embodiment, the term "task" refers to an individual work unit and is described as an example of a name specifically indicating the multiple processing steps described in the claims. However, the present invention is not limited to this, and processing steps may be divided into other names such as "phase," "module," or "work step." For example, even if "modules" are defined as functional units in software design and task content and processing entities are assigned to each module, each module can be treated as corresponding to multiple steps, just like a "task." Furthermore, when a process is understood as a broad stage, the term "phase" can also be used. In this way, names and levels of granularity can be selected appropriately depending on the implementation form and management requirements, without departing from the technical scope of the present invention.

[0096] Furthermore, the information processing system 10 described above may also be configured to display history information 60 for a processing step, as shown in FIGS. 12 and 13 . This configuration allows a user U to refer to past process details while performing additional processing based on other input information, even during the processing step. For example, even if a user U starts supplementary work or branching processing based on other input information, the history information 60 stored in the processing step information DB 224 can be displayed on the screen. This allows the user U to simultaneously check the reasons for changes to the processing step and intermediate results, and perform detailed difference analysis or additional comparisons as necessary, enabling flexible dynamic operation. Note that FIG. 13 shows an example of a screen display in which, when specific history information 60 is selected, the processing step information 32, processing subject information 34, processing progress status information 38, and processing monitor information 40 corresponding to the selected history information 60 are displayed.

[0097] (Variation 1) Furthermore, if we look at the above-mentioned information processing system from a different perspective, the problem (purpose) that the information processing system of this embodiment aims to solve can also be seen as "taking in content information of the invention and generating and editing a patent specification."

[0098] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information including at least one of invention content information, prior art information, technical problems, and effects; a processing entity setting step for setting a processing entity in charge of each of a plurality of processing steps set for processing the input information; a processing result acquisition step of acquiring proposed or draft information of the patent specification created or edited in each step by the processing entity and finally providing it in an outputtable format; "An information processing method including the following."

[0099] According to this configuration, the technical content of the invention entered by the user (draft claims, background art, effects of the invention, etc.) is acquired in the "input information acquisition step," and then assigned to multiple predetermined processes (description policy consideration process, claim draft process, drawing creation instruction process, etc.) in the "processing subject setting step," allowing the documentation to be shared among specialized patent specification preparation agents, the user themselves, and other information processing systems. Furthermore, by executing the "processing result acquisition step," it is possible to sequentially output the draft patent specification and various draft editing results, while also realizing a flow for making corrections and proofreading in stages.

[0100] Furthermore, if a patent specification drafting agent has a large-scale language model or natural language analysis module, it can format the output in a format that conforms to past publication formats and patent office guidelines. This automates tasks such as checking for typos and claim consistency, while generating specifications that always reflect the latest publication and examination trends, thereby improving the efficiency and accuracy of the document drafting process, which requires specialized know-how.

[0101] Furthermore, when the user enters prior art information or cited document information that needs to be reflected in the patent specification, the system analyzes that information and provides guidance to meet legal requirements, such as whether it should be included in an independent claim or not, and showing its correspondence with dependent claims.If the user instructs the agent to reprocess as needed, it can regenerate only a portion of the data, or perform branching comparisons of paragraph numbers and paragraph contents.

[0102] As described above, the modified version specialized for patent specification drafting agents has the advantage of not only generating text but also managing and editing claims with complex logical structures and explanations of related drawings in parallel. Users can efficiently check the consistency between claims and manage variations after amendments, which was previously done manually, thereby realizing faster and higher quality document creation in specialized fields and more reliably improving the accuracy of output results.

[0103] (Variation 2) Furthermore, if we look at the above-mentioned information processing system from a different perspective, the problem (purpose) that the information processing system of this embodiment aims to solve can also be seen as "obtaining comprehensive output by processing input information from multiple angles."

[0104] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a processing result acquisition step of acquiring a processing result according to a processing step set for processing the input information, and, upon acquiring change instruction information for the processing result, acquiring a processing result in the processing step based on the change instruction information; "An information processing method including the following."

[0105] This configuration eliminates the traditional one-way processing flow for obtaining processing results. Instead, it acquires modification instructions for adding additional parameters or perspectives to the output results, enabling the system to re-obtain processing results based on those instructions. In other words, by repeating branching processes based on input information, multifaceted comparison and verification becomes possible, ultimately achieving a more comprehensive output. For example, if the agent determines that the content generated by the agent is not sufficiently comprehensive, the user can input new modification instructions, adding vocabulary or evaluating it using a different algorithm in the same process, resulting in more in-depth results. This ability to apply multiple approaches in a chained manner offers the advantage of obtaining broader and more precise output for the specialized tasks and complex challenges faced by the user U.

[0106] Another feature of this configuration is that it allows for flexible management of the branching and merging of processing steps. Even if the results output do not meet expectations, the same processing steps can be reused by issuing change instructions incorporating a different perspective based on those results, eliminating unnecessary duplication of work and allowing for multifaceted verification. In addition, by assigning a different processing entity (such as an agent, another system, or a terminal operated by the user U) to each step, comprehensive consideration that combines individual areas of expertise and evaluation criteria becomes possible. If a more appropriate outcome is obtained from one of the branching results, it can be adopted and moved forward, and if necessary, re-verification can be carried out based on further change instructions.

[0107] Another example of a variation that takes advantage of this multifaceted processing mechanism is a case in which user U gradually adds additional information and branches. For example, this can easily be applied to collaborative editing work, where a user requests revisions after a rough outline has been generated, and then, after the process has progressed to a sub-process that incorporates these requests, another expert adds additional instructions to add additional perspectives. By integrating the results of each expert's unique perspective at each branch, a comprehensive deliverable incorporating highly complex knowledge can be completed. Thus, a variation based on this configuration is significant in that it does not simply end the entire process as a series of steps, but rather allows users and multiple other parties to intervene alternately to perform multifaceted processing and pursue advanced and appropriate output.

[0108] (Variation 3) Furthermore, from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can be interpreted as "integrating user U's processing steps and agent-based processing steps within the same framework, and executing multifaceted processing while flexibly switching the processing entity as needed." That is, for each processing step requiring input information, the system assigns either user U or an agent based on pre-defined allocation criteria (e.g., processing step difficulty, urgency, compatibility with specialized fields, etc.). If a change instruction is received midway through a processing step, such as "the accuracy of an agent is insufficient" or "the urgency has increased, so I would like to delegate the task to an agent," the system can instantly change the processing entity. This allows user U to obtain comprehensive and advanced output without being bound by conventional unidirectional flows, as long as multifaceted verification and reprocessing are performed in stages while optimizing the quality and processing speed of the processing results.

[0109] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a processing entity setting step of setting a processing entity for each processing step by referring to an allocation criterion according to the processing steps set for processing the input information; a step of recording the processing results obtained from the processing subject set in the processing subject setting step in a unified manner, and receiving change instruction information corresponding to the processing results not satisfying a predetermined standard, and resetting the processing subject based on an allocation standard; "An information processing method including the following."

[0110] This configuration allows the processing steps performed by user U and the execution steps performed by agents to be integrated into the same information processing method. This allows for continuous, multifaceted verification, not simply following a unidirectional flow, but also switching the appropriate processing entity depending on the task status. Specifically, if the accuracy of the processing results obtained from the processing steps initially assigned to an agent is deemed insufficient, the user can send change instruction information (e.g., "review by user U required" or "replace the agent with a different model") from the user's terminal, allowing some processing steps to be transferred back to user U or another agent. Furthermore, if user U issues a change instruction for a processing step currently in progress, such as "high urgency, so we want to complete it quickly," the assignment criteria can be checked immediately and the task can be delegated to an agent. This allows for the smooth transfer of processing from user U to an agent, or vice versa, when necessary, even while using the same processing steps. Overall, this allows for highly comprehensive and flexible task execution.

[0111] Furthermore, even when there are multiple processing steps and dependencies between them, it is possible to branch and verify multiple output proposals obtained at the merging stage, while making a choice between "User U" and "Agent" at each processing step. For example, when multiple processing steps have no dependencies, User U and the agent can run in parallel to double-check the work and then decide on the best output by consensus. Repeating this branching and merging pattern allows for multifaceted consideration that cannot be achieved with a single process, resulting in a more comprehensive deliverable. Furthermore, if requirements that were unclear at the start of a processing step become clearer during the process, this information can be added as change instructions, making it easy to have the agent re-execute only specific processing steps or transfer the work to another User U with specialized knowledge.

[0112] In this way, the information processing system according to this variation allows user U and agents to work together organically in the same workflow, freely switching the processing entity as needed. Therefore, for example, in document generation or complex analysis tasks, a so-called "hybrid approach" can be implemented at low cost, in which step-by-step feedback is used to improve completeness and comprehensiveness. As a result, compared to conventional operational models in which the user U or the agent is completely fixed in advance, this variation allows for the flexible use of both, making it easy to select the optimal processing entity depending on the workload and target accuracy. As described above, a major feature of this variation is its ability to conduct multifaceted reviews in parallel or sequentially while improving the comprehensiveness and accuracy of the processing results. It can also be adapted to cases in which user U or multiple experts gradually provide additional requirements, making it extremely useful in processes aiming for advanced and complex deliverables.

[0113] (Variation 4) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can also be understood as "automating or semi-automating the entire business process safely and efficiently by variably setting the authority of each agent according to the intelligence level and area of ​​expertise, and dynamically controlling whether approval from user U is required and the range of processing steps that the agent can be responsible for." In other words, the overall accuracy and speed of the work are optimized by limiting an agent to subtasks or draft level work if the agent is inexperienced, and requiring approval from user U for processing steps that require advanced editing authority, while allowing agents with experience to proceed with full-scale processes without approval.

[0114] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; an authority setting step of setting a plurality of authority levels, such as advanced editing authority or subtask limited authority, for a plurality of agents by referring to their respective intelligence levels, areas of expertise, and performance data; a processing subject management step of determining, based on the authority level, which of a plurality of processing steps set for processing the input information the agent can execute without approval or which processing steps must be approved; an evaluation history management step for evaluating the processing results of the processing steps handled by the agent and recording the performance such as success rate and error status; an authority level changing step of raising or lowering the authority level of the agent in question according to the performance information from the evaluation history management step; an allocation modification step of updating the range of processing steps that an agent can take charge of and whether approval is required based on the content changed by the authority level change step; "An information processing method including the following."

[0115] This configuration allows for fine-grained control over each agent's permitted scope of operations and the processes they can handle, rather than simply assigning agents to processes uniformly. Specifically, highly authorized agents can proceed with important or complex processes without user U's approval, while less advanced agents can be granted limited subtask authority and assigned only to drafting and auxiliary tasks, making risk management easier across the entire business. If an inexperienced agent repeatedly makes major mistakes, their authority can be lowered based on performance data obtained through evaluation history management, preventing them from being assigned to important processes. Conversely, if an agent consistently achieves a high success rate and short processing times, user U no longer needs to review each process. Raising their authority level expands the number of processes that can be automated, contributing to improved productivity.

[0116] Furthermore, to reinforce this configuration, it is possible to set up a learning and verification framework separate from the main business, simulating the same processing steps that agents use in actual business operations and accumulating track records. This allows the intelligence level and compatibility with the specialized field of newly introduced agents and agents immediately after updates to be safely verified before determining the authority level for actual operation. For user U, this has the advantage of being able to utilize agents while minimizing risk. Furthermore, when authority is changed dynamically, it is possible to automatically implement flow changes, such as omitting reports to user U once an agent has acquired a high level of authority, even for the same processing steps.

[0117] Additionally, by sharing history and evaluation results throughout the system, the same procedures for permission management and approval requirements can be used even as the number of agents increases as a company or organization scales. For example, by standardizing high-level permission granting standards for all agents, a consistent level of quality can be maintained across development teams and operations departments, reducing maintenance burden. Furthermore, increasing the number of processes that do not require User U's approval leads to increased automation, while allowing User U to focus on the labor hours that require important revisions and reviews. This allows for efficient use of human resources overall, while achieving both work speed and high-precision output. This modified version, with its agent permission management mechanism, realizes extremely flexible operational configurations by dynamically optimizing workflows while reflecting actual intelligence levels and learning degrees.

[0118] (Variation 5) Furthermore, from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can be interpreted as "enhancing auditing capabilities to enable detailed recording and visualization of the details of each processing step and the reasons for decision-making by the processing entity, enabling rapid and accurate subsequent tracking." In other words, even if an error or malfunction occurs during the operation of this system, a consistent log records what was input at each processing step and the logic or algorithm applied by the processing entity to make the final decision, easily achieving the transparency and clarity of responsibility required for external and internal audits, fault investigations, and other such situations. This configuration is also important for preventing black-boxing and ensuring high reliability when operating agent-type processing entities that include trained models.

[0119] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a log generation step of generating, for each of a plurality of processing steps, a log that aggregates at least one of input data and logic information used in the processing steps and execution results; a decision-making reason management step of extracting the inference process or the basis data referred to by the processing entity in charge of the processing step, and associating the log with metadata indicating the decision-making reason of the processing entity; An audit management step that allows the logs and decision-making reasons to be searched and viewed, and makes it possible to determine which processing entity used which logic in which processing step to arrive at a processing result; an error cause identification step of tracking the log of the processing step in question in the audit management step when an error or abnormality is detected, and identifying the processing entity that caused the error or the logic information applied; a responsibility range setting step of recording the decision-making domain of the processing entity in charge of the processing step based on the content of the log in order to identify the responsibility range of the processing entity; "An information processing method including the following."

[0120] According to this configuration, the log generation step first records detailed input / output and logic information for each processing step executed within the system. In this embodiment, whether the processing entity is an agent or another system module, unique IDs, version information, configured algorithm parameters, and other information can be logged. For auditability, storing the input information type and data source path along with the log information allows for easy tracking of the data source used to generate the output result when referenced externally later. Furthermore, the decision-making reason management step converts intermediate results of the inference process and reference information for similar cases into metadata, preventing the process from becoming a black box and allowing for later reevaluation of why a particular conclusion was selected and what data trends were emphasized. This approach is particularly effective in ensuring accountability in environments incorporating machine learning models and large-scale language models (LLMs).

[0121] Next, the audit management step provides a management console and API for searching and viewing the logs and metadata, displaying a list of processing steps, the ID of the corresponding processing entity, reference logic, and other information in a consolidated view. When doubts arise about a specific processing result or an error is reported, system administrators and auditors can quickly identify the relevant processing step through the management console and identify the problem by comparing the traced logs. The error cause identification step also provides a deeper level of analysis, allowing for detailed investigation of which logic version or inference module was used, or whether there was an abnormality in the input data. This allows for quick root cause investigation even in complex system structures, contributing to faster countermeasures and more efficient user support.

[0122] Finally, through the responsibility assignment step, the authority and scope of responsibility of the processing entity that performed the processing step are explicitly recorded. For example, even for tasks divided among multiple agents, the process log and metadata link information clearly clarifies which agent (or other system module) made the final decision. As a result, even in large-scale environments with numerous processing entities, responsibility can be quickly identified during the audit phase, and any suspected deficiencies can be corrected or improved to the appropriate extent. In this way, by combining detailed logs for each processing step with the visualization and storage of decision-making reasons, this variant provides a framework for ensuring reliability and transparency even in highly automated information processing systems. From the perspective of infringement detection, examining characteristic steps such as "log generation," "decision-making reason management," "audit management," and "responsibility assignment" makes it easy to identify similar system implementations.

[0123] (Variation 6) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can also be understood as "establishing a standard interface for multiple processing entities (including agents) to coordinate work while sharing expertise, and enabling the safe and smooth exchange of decision-making results and intermediate data." In other words, not only in simple one-to-one coordination, but also in large-scale environments where multiple specialized agents process data in parallel, if common specifications for intercommunication are defined, new processing entities can be easily added or replaced, and centralized management of monitoring and control will also become easier.

[0124] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a protocol management step of formulating a common protocol that standardizes the format and procedure of data transmission and reception used among a plurality of processing entities; a cooperative control step of transferring intermediate data information or decision-making results generated by at least one processing entity to another processing entity according to the common protocol; a data consistency checking step of validating the data received in the linkage control step and issuing an error response if an invalid format or inconsistency is detected; a monitor / re-instruction step of monitoring execution statuses of a plurality of processing entities from an external system via the common protocol and enabling transmission of operation information serving as a re-instruction; "An information processing method including the following."

[0125] With this configuration, the data structures and call methods to be output by each processing entity are standardized based on the common formats (e.g., JSON, Protocol Buffers, XML, etc.) and endpoint specifications established in the protocol management step. This allows intermediate data and decision-making results to be exchanged in the correct format in the collaboration control step, even between agents implemented by different vendors or in different language environments. Additionally, the data consistency verification step checks formats and types, and returns an error if there are items the receiving party cannot understand or missing values, thereby minimizing the occurrence of system failures or malfunctions due to inconsistencies.

[0126] Furthermore, in this embodiment, the monitor / redirect step allows an external system to simultaneously monitor the status and progress of each processing entity using a common protocol, and if necessary, issue operational requests such as interruption, re-execution, or task switching. This makes it easier to achieve integrated orchestration even for large-scale workflows, eliminating the need to monitor each individual agent using a separately implemented API. Furthermore, this architecture clarifies and standardizes the overall system configuration, making it easy to integrate the interface with existing protocol management units when adding or enhancing AI agents in the future by preparing modules that conform to a common format.

[0127] As described above, this variant combines a standard API protocol that facilitates collaboration between multiple agents with functions such as collaborative control, validation, monitoring, and re-instruction. As a result, (1) it streamlines the exchange of intermediate data between processing entities, (2) it reduces development and integration costs when introducing new modules, and (3) it enables external integrated control, realizing flexible operation that is less likely to cause confusion even in large-scale automated systems. Additionally, from the perspective of ease of infringement detection, checking the unique flow of "whether multiple processing entities share areas of expertise and exchange intermediate results or decision-making results through a common protocol" has the advantage of making it easy to quickly identify the implementation of similar systems.

[0128] (Variation 6) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can also be seen as "expanding the scope of AI implementation while ensuring overall work accuracy and safety by gradually managing the growth and learning status of agents, assigning major tasks to mature agents, and limiting agents in a shallow learning stage to trial operations and partial tasks." In other words, by centrally managing the lifecycle of agents, in which they repeatedly update and learn, using a registry or the like, a mechanism is provided that enables advanced AI agents to be smoothly incorporated into production operations.

[0129] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; an agent registry management step of maintaining a registry for centrally managing version information and performance indicators related to multiple agents, and when a new agent or an updated agent is registered, recording the algorithm and learning history of the agent as metadata; a learning evaluation monitor step of monitoring the learning progress and evaluation index of the agent registered in the registry and not shifting the agent to official operation until a predetermined threshold is met; When switching over to an updated version of an agent in operation, a version management step is performed in which the old and new versions are run side by side to compare and evaluate them, and if a problem is detected, the agent is restored to the old version. a task allocation control step of allocating a main task to an agent having a high evaluation index among agents currently in official operation based on the registry, and allocating a partial task to an agent having a low learning stage; an audit log management step of accumulating the agent's operational performance and evaluation log for auditing purposes and enabling reference to the version switching history and operation continuity status; "An information processing method including the following."

[0130] With this configuration, the agent registry management step centrally manages the version numbers, parameters, and learning status of agents registered in the system. By referencing this registry, operations personnel and system management departments can easily understand the expertise and model characteristics of each agent, as well as the time of its update. Furthermore, the learning and evaluation monitoring step monitors the agent's evaluation indicators (accuracy, processing speed, failure rate, etc.) as it undergoes continuous learning and additional training. This gate control prevents the agent from being introduced into the production workflow until it meets the criteria, thereby reducing the risk of inexperienced agents making mistakes. Additionally, the version management step smoothly updates existing agents while maintaining the previous version as a safety measure, allowing for a quick rollback in the event of an unforeseen problem, ensuring business continuity.

[0131] In addition, the task allocation control step refers to each agent's evaluation index and performance log, assigning primary tasks to agents with high accuracy and stability, while assigning limited tasks such as subtasks and draft areas to agents with less advanced learning stages. This allows the system as a whole to dynamically optimize its allocation, improving accuracy and saving time while minimizing the occurrence of serious errors by immature agents. Finally, the audit log management step makes it possible to refer to the agent's operation history and version switching history at any time, facilitating compliance with operational policies and handling internal and external audits. Combined with the ability to archive older versions of agents, this also provides the flexibility to retroactively investigate the causes of problems or conduct comparative analysis for experiments.

[0132] As described above, this variant comprehensively incorporates a multi-layered set of functions, such as "agent registry management," "learning and evaluation monitor," "version management," "task allocation control," and "audit log management," into the information processing device, enabling maturity management throughout the entire agent lifecycle. This allows agents to be put into production in stages while minimizing risks and errors, and the automation level of the entire process can be advanced efficiently and safely. Furthermore, from the perspective of rights enforcement, the variant incorporates measures such as "distinguishing between agents that are still learning and those that have completed evaluation, and controlling the gate for transitioning to official operation" and "performing comparative evaluations between upgraded versions and older versions," which has the advantage of making it easier to confirm the actual implementation status of similar systems and to detect infringements.

[0133] (Variation 7) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can also be interpreted as "realizing unmanned operation in which the processing steps proceed without user approval when the accuracy of the automatic processing by the agent meets a certain level, and inserting an approval flow only when the accuracy is low or does not exceed a score threshold, thereby maintaining a balance between operational efficiency and safety." This assumes that even in situations in the future when agents become sufficiently sophisticated, they will continue to perform operations automatically, leaving only the minimum necessary manual confirmation, and will gradually expand the scope of automation while making effective use of human resources.

[0134] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a result evaluation step of evaluating the processing result of the agent or the like in the processing step using an index of accuracy or reliability and determining whether the index is above or below a predetermined threshold; a mode control step of switching between an unmanned operation mode in which the processing steps are automatically completed and a hybrid control mode in which additional approval is required based on the determination of the result evaluation step; an automatic completion step in which, when the unmanned operation mode is selected, the processing entity continuously executes the next process and completes a final product without requiring confirmation by a human operator; an approval hold management step of assigning an approval waiting flag to a processing result below a threshold value when the hybrid control mode is selected, and temporarily suspending the subsequent process; "An information processing method including the following."

[0135] With this configuration, the result evaluation step first calculates the accuracy, reliability score, error rate, etc. of the results and intermediate results output by the agent and compares them with preset thresholds. If the threshold is met, the mode control step instructs an automatic completion flow (unattended operation mode), and the corresponding processing entity proceeds to the next step without user approval, thereby achieving highly efficient workflow progression. On the other hand, if a result does not meet the threshold, hybrid control mode is activated and the approval pending management step is activated. This allows for flexible responses to questionable results, such as waiting for user U's judgment and encouraging correction, reprocessing, or assignment to another agent. Furthermore, as the agent's accuracy improves sufficiently, more steps can be completed in unattended operation mode, limiting user U's intervention to exceptional cases (such as when the result falls significantly below the threshold). As a result, the burden on user U is reduced, and overall speed and cost reduction are expected. However, the approval pending management step also allows important projects and high-risk processes to be checked without being left unattended, thereby reducing the risk of quality degradation and misprocessing. With this configuration, this modified example makes it as easy as possible to implement hybrid operations that seamlessly coexist "almost complete automation driven by an agent" and "partial human approval." Also, from the perspective of patent infringement, there is an advantage that by checking whether characteristic steps are implemented, such as determining whether to switch to unmanned operation mode or hybrid control mode, or transitioning to an approval flow when the threshold is exceeded, the embodiment of the relevant system can be easily understood.

[0136] (Variation 8) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment is trying to solve can also be understood as "providing a mechanism for automatically arbitrating which result to ultimately adopt when multiple processing entities or external operational instructions exist simultaneously and their decisions or instructions conflict." In other words, by strengthening the function of detecting multiple competing plans within the system and determining which to adopt based on predefined priorities and policies, the system allows for operations that prevent confusion and errors and allow work to proceed smoothly.

[0137] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a conflict detection step of collecting decision-making results by a plurality of processing entities or operation instructions input from the outside and determining whether there are conflicting contents therein; an arbitration step in which, when the conflict occurs, the result of which processing entity is to be adopted as the final decision is determined by referring to a predefined priority setting or operation policy; a result reflection step of transferring the decision of the arbitration step to a next step and storing the arbitration content, the applied policy, etc. as a log; a policy switching step of switching the operation policy according to the operational situation and varying the priority of the processing entity to which the final decision-making authority is given in an emergency or normal situation; "An information processing method including the following."

[0138] This configuration allows for the first conflict detection step to quickly detect discrepancies between the results of multiple processing entities (e.g., Agent A and Agent B present different results) and external instructions. Then, in the arbitration step, the system automatically determines which output proposal is correct by referring to pre-defined priorities and operational policies (e.g., prioritizing the agent with the highest accuracy score, or, if scores are equal, prioritizing the agent with the most recent command time). These arbitration results are passed to subsequent processes in the result reflection step and logged, allowing for later tracking and verification of how the conflict was resolved and the criteria used to reach the final decision. Furthermore, in the policy switching step, policies can be switched in the event of an emergency or special case, allowing for flexible implementation, such as prioritizing a processing entity with higher authority or a specific external management module.

[0139] For example, under normal circumstances, a system can prioritize accuracy evaluations between autonomous agents and prioritize proposals from agents with higher accuracy. However, in the event of a system failure or unexpected behavior, a policy switching step can be implemented to allow a separate maintenance management module to make the final decision. This allows even large-scale interconnected systems to maintain order while responding to diverse scenarios and ensure fast and accurate task completion. Furthermore, thorough logging makes it possible to visualize which priority rules were applied and how agent decisions were handled versus other decisions during subsequent audits and troubleshooting. In terms of rights enforcement, external third parties can easily verify the system implementation, including clear procedures for conflict detection and arbitration steps and the automatic priority assignment process through policy switching, making it relatively easy to identify unique features and detect infringements.

[0140] (Variation 9) Furthermore, if we look at the above-mentioned information processing system from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can also be seen as "clarifying the entity responsible for agent products and establishing a monitoring and management system so that legal and ethical requirements such as attribution of copyrights and inventions of deliverables, compliance with terms of use, and privacy protection can be met." In other words, this is a mechanism that can be constructed that can manage at a system level who holds which rights for content and inventions created by automatic agent processing, and under what license conditions they can be used.

[0141] If the problem is understood as above, the invention as a means for solving the problem may be, for example, as follows: "An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a terms management step for managing contractual terms or terms of use for deliverables generated by the agent and making it possible to refer to terms defining copyright or right attribution for the deliverables; a compliance monitoring step for monitoring violations of privacy protection clauses and content guidelines, and for issuing a warning or blocking process when there is a risk of violation in the agent output; a product responsibility assignment step of assigning metadata indicating a responsible party, initial license information, etc. to the product generated by the agent; a legal notification step of presenting contract terms established from legal and ethical viewpoints to a user when using the agent and obtaining consent; an audit log retention step of storing the agent's operational performance and output history as a log, and making the log available for reference in the unlikely event of a violation of the terms of use; "An information processing method including the following."

[0142] With this configuration, first, the rights relationships of the agent output (copyright, invention attribution, secondary use permission form, etc.) are clearly defined within the system in the terms management step, and then the latest contract terms and ethical policies are always presented to users of the agent in the legal notification step, making it easy to obtain consent for continued use.In addition, with the terms compliance monitoring step in operation, detection of violations of privacy protection and content guidelines is relatively automated, and it can be designed to block inappropriate deliverables early on.

[0143] Furthermore, by assigning responsible parties and license information to agent products as metadata in the deliverable responsibility assignment step, the handling of deliverables can be carried out without misunderstanding, including subsequent secondary use and format conversion. For example, even if multiple teams within a development project reuse an agent product, the system logic can easily control which team can use it for secondary use and to what extent. Finally, by saving operation results and output logs in the audit log retention step, it becomes easier to investigate suspected violations of contract terms and to confirm legal evidence of copyright at a later date.

[0144] As a specific example of use, when a web service operator allows users to post automatically generated content through an agent, the Terms of Use Compliance Monitoring Step blocks inappropriate or defamatory language, issuing an alert and automatically deleting the content if any problematic phrases are detected. On the other hand, if an agent is tasked with a certain creative task and the resulting work is an academic paper or program code, the Work Responsibility Assignment Step embeds the code's license type as metadata, preventing users from violating rules when copying or modifying the work later. Comprehensive operation, including this mechanism, makes it possible to realize a system that easily maintains legal and ethical compliance even in a large-scale automated environment. In other words, this variant is significant in that it anticipates the agent's "handling of intellectual works" in advance and strengthens the system-wide support mechanisms for rights management and compliance with terms of use.

[0145] (Variation 10) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can also be understood as "providing a user interface that allows the user U to constantly grasp the overall system status and intervene quickly as needed, while reducing opportunities for explicit human operation as the agent proactively advances processing." In other words, while assuming autonomous decision-making and automatic execution by the agent, it is expected that by enhancing the UI / UX so that the user U can monitor the operating status in real time and perform interrupt operations only where necessary, it will be possible to prevent malfunctions and unintended processing while improving work efficiency.

[0146] If the problem is understood as above, the invention as a means for solving the problem would be, for example, as follows: "An information processing method executed by an information processing device, a dashboard control step of aggregating the work status and progress of agents in charge of a plurality of processing steps and displaying, in a dashboard format, result information of each processing step automatically executed by the agents; an interactive UI display step of visualizing, on a user interface, a summary of reasoning grounds or error recognition information regarding the processing content automatically determined by the agent, and generating an interactive interface for the user to perform an interrupt operation; an interrupt control step of, when receiving the interrupt operation, stopping or readjusting the execution of the agent based on the content of the operation and acquiring the processing result after the change; a notification output step of determining the urgency and the attention level based on the presence or absence of a change or the processing state in the interrupt control step, and displaying a push notification or a modal window to the user as necessary; a process impact visualization step of analyzing the impact of the plurality of processing steps on an upper process or a lower process and highlighting the analysis result on the dashboard; "An information processing method including the following."

[0147] According to the above configuration, the dashboard control step allows the user to grasp at a glance the list of processing steps performed by multiple agents and their current processing status. The interactive UI display step then summarizes the inference basis and anomaly detection information output by the agents and presents it on the screen, allowing the user U to quickly understand the content and, if necessary, perform interrupt operations (such as re-execution, stopping, or parameter changes) via the interrupt control step. Furthermore, the notification output step immediately notifies the user U of high-priority processes and error situations, reducing the risk of unintended agent processing runaway or failures being left unattended. The process impact visualization step also highlights the impact of changes to one process on other processes, allowing the user U to easily understand the effect of local interrupts on the overall flow and determine the optimal response. This mechanism allows the user U to appropriately intervene and control processes, even in an agent-based automated processing environment, thereby improving safety and efficiency.

[0148] (Variation 11) Furthermore, when the above-described information processing system is viewed from a different perspective, the problem (purpose) that the information processing system according to this embodiment aims to solve can also be understood as "efficiently and stably operating a large number of agents distributed across clouds and edge devices via an orchestration control unit, and automating load balancing, scale-out, and failure countermeasures as needed." In other words, a technical platform is required that maximizes overall efficiency while avoiding conflicts between agents by combining dynamic scheduling according to the load status and priority of groups of agents deployed in diverse locations, failover control when a failure occurs, and distributed transaction management.

[0149] If the problem is understood as above, the invention as a means for solving the problem would be, for example, as follows: "An information processing method executed by an information processing device, an agent / resource management step of acquiring status information corresponding to a plurality of agents placed in at least one of the cloud and the edge device, and centrally managing the load statuses and processing capabilities of the agents; an orchestration control step of dynamically controlling tasks to be assigned to the agents based on a plurality of factors including the priority of processing steps, dependency relationships, communication delay information, and risk of failure occurrence; a failover control step of transferring a task to another node or another agent in accordance with a predefined failover policy when an abnormal stop or a failure is detected during execution of the agent; a progress management step of monitoring the progress of the tasks after the reallocation or rearrangement and displaying the progress on a user interface of the user terminal as necessary; a distributed transaction control step for maintaining consistency of update processing when the update processing is executed across multiple nodes; An information processing method comprising the steps of:

[0150] With the above configuration, the agent / resource management step visualizes and comprehensively understands the CPU, GPU, and other computing power, storage usage, and available communication bandwidth of agents distributed across various locations, including the cloud and edge. The orchestration control step then automatically determines the optimal assignment based on the dependencies and urgency between processing steps, as well as the current load status of each agent, and quickly distributes tasks to multiple nodes as needed. This enables efficient resource utilization and high throughput, even in large-scale environments. Meanwhile, in the event of a failure, the failover control step quickly activates, pinpointing the interrupted part of the task and reassigning it to a backup node or another agent, thereby improving overall system availability. The progress management step constantly monitors the task relocation process, allowing users and supervisors to instantly grasp the process status via the user interface. Furthermore, the distributed transaction control step ensures final consistency even when multiple nodes perform concurrent write and update operations, preventing system inconsistencies and data inconsistencies. In this way, load balancing, scale-out, and fault prevention can be implemented comprehensively even in large-scale distributed environments, making it possible to achieve continuous, highly efficient operation even when running agents on a large scale.

[0151] (Variation 12) Furthermore, if we look at the above-mentioned information processing system from a different perspective, the problem (purpose) that the information processing system of this embodiment aims to solve can also be seen as ``continuously utilizing adjustment information for a processing entity obtained during the trial phase, design phase, etc. in another processing entity or another operating environment.''

[0152] If the problem is understood as above, the invention as a means for solving the problem would be, for example, as follows: "(Form 1) An information processing method executed by an information processing device, an adjustment information acquisition step of acquiring adjustment information for a processing subject; an adjustment information application step of applying the adjustment information to the processing subject; An information processing method including: (Form 2) further comprising an archive storage step for storing the adjustment information; the adjustment information acquisition step includes reading out the adjustment information stored in the archive. 2. An information processing method according to claim 1. (Form 3) the adjustment information includes at least one of prompt information set for the processing subject, parameters for additional learning, and a dialogue history. The information processing method according to the first and second aspects. (Form 4) further comprising an environment information acquisition step of acquiring environment information of a deployment destination where the processing entity is deployed; In the adjustment information application step, consistency between the environment information and the adjustment information is verified. The information processing method according to any one of aspects 1 to 3. (Form 5) The adjustment information includes log information generated and updated during operation of the processing entity, The processing logic of the processing subject is re-learned or re-configured in the adjustment information application step based on the log information. The information processing method according to any one of the first to fourth aspects. (Form 6) In the adjustment information acquisition step of acquiring the adjustment information, identification information for version management is assigned to the adjustment information; In the adjustment information applying step, the adjustment information is selected and applied based on the identification information. The information processing method according to any one of aspects 1 to 5.

[0153] In this modified example, the following multifaceted effects can be obtained. (1) Improving operational efficiency by inheriting adjustment information First, the adjustment information acquisition step ensures high reusability by collecting a variety of elements, such as prompt information and learning parameters generated during the operation and trial phases, as well as user interaction logs. For example, consider a situation where a user company has a processing entity (such as an AI agent or script engine) that has been highly tuned for a specific task. If this company wants to redeploy it to a different site or cloud environment, it can quickly reproduce the same level of processing quality by acquiring new adjustment information and loading it appropriately in the application step. This reduces the effort required to incur training costs from scratch in a new environment to achieve the original performance, significantly reducing the speed of implementation and the labor required for adjustment.

[0154] (2) Safety and scalability through version control As shown in form 6, when combined with a configuration that assigns identification information for version management to adjustment information, multiple versions of adjustment information can be managed in parallel. Even if an unexpected malfunction or loss of accuracy occurs as a result of the introduction of a new version of parameters during operation, stable operation can be quickly restored by simply rolling back to the previous version. Furthermore, it becomes easy to switch between multiple versions in a staging environment and a production environment and perform comparative verification, making it possible to gradually find the optimal tuning state for the processing entity.

[0155] (3) Accumulation of operational logs and continuous learning As in mode 5, if a configuration is adopted in which log information updated while the processing subject is running is reflected in the adjustment information, a major feature is that more advanced know-how and correction mechanisms are accumulated over time. As a result of incremental learning and reconfiguration, the accuracy of the entire system is continuously improved, and the final tuning state can be easily transferred to another environment as an archive, so the usefulness of this system can be maintained over the long term.

[0156] (4) Reducing problems by checking environmental compatibility As in mode 4, by simultaneously acquiring deployment destination environment information (cloud or on-premise, hardware configuration, network bandwidth, etc.) and adding a process to check consistency with the adjustment information in advance, it becomes possible to automatically switch parameters according to the application and scale. As a result, it is possible to mitigate in advance the risk of "serious errors occurring due to computational resources different from those expected" after application, which is expected to significantly improve operational stability and reduce implementation costs.

[0157] (5) A mechanism for allowing a third party to externally observe the steps of the invention Furthermore, in order to relatively easily estimate from the outside whether the steps shown in the first to sixth aspects are being performed, the following additional configuration is preferable. (a) Display on the user interface: During the adjustment information acquisition stage, messages such as "Loading adjustment information ID: XXX" or "Applied to: Processing entity A (Version: v2.0)" are displayed clearly using a progress bar, etc. This makes it easy for external auditors and users to understand that adjustment information is being acquired and applied while operating the screen, without having to look at the log. (b) Step name output to log: Unique step names such as "STEP_GET_TUNING_INFO" and "STEP_APPLY_TUNING_INFO" are recorded in the system's internal console and audit log, and success or failure is also indicated. By referencing the server log and cloud audit information, a third party can easily determine whether the flow that satisfies steps 1 to 6 above is running. (c) Assigning hash values ​​and version information: Applying hash signatures and consistent versioning to adjustment information files and outputting them to logs when they are read in. If the same hash value or version ID is used in an external environment, it is relatively easy to infer the possibility that the data in question may be used without permission as adjustment information for the present invention. (d) Post-recording on user approval screen: When applying adjustment information, the system goes through the administrator approval and review screen, and the fact is displayed on the screen and saved in the log along with the date. If such an approval flow is implemented, it is easy to identify similarities when checking similar implementations by third parties by simply comparing the behavior of the approval screen and the recording format.

[0158] (6) Overall effects By comprehensively combining the above configurations, the present modification 12 can provide the following multiple effects. - Prompt information and additional learning parameters accumulated in the past can be easily reused during deployment and operation, reducing implementation and migration costs while maintaining the accuracy of the processing body. If continuous logs and conversation history are reflected in the adjustment information, advanced automated corrections can be made over time, which is expected to improve processing quality. By implementing safety measures such as version control, risks associated with introducing new versions can be reduced, and operational stability can be improved by easily reverting to a previous version in the event of a serious error. By incorporating visualization measures in the user interface and log output, external reviewers (such as license auditors) can clearly understand which steps were performed and which version of adjustment information was loaded. As a result, it becomes easier to determine whether the steps of this invention are being implemented across the entire system. By incorporating a configuration that verifies the compatibility of the deployment environment and adjustment information in advance, the probability of achieving the performance of the processing body can be improved and the error recurrence rate can be reduced, even across a variety of infrastructures such as large-scale cloud environments and local servers.

[0159] These effects allow the user to enjoy the benefit of easily auditing external examples and checking for infringements while achieving efficient and reliable operation through the steps shown in Modifications 1 to 6. In other words, a major feature of Modification 12 is that it is designed to enable the user to discover with minimal effort whether the system is implementing the technical concept of the present invention by providing features not only in the mechanisms for storing adjustment information and reapplying learning parameters, but also in the screen display method and log recording format.

[0160] In the present invention, a series of screen display functions referred to as a "processing process information display step," "processing subject information display step," "sub-process information display step," etc. are not limited to specific screen layouts, display control timing, screen transition order, or types of components (buttons, lists, windows, etc.) used. For example, they include any form that a person skilled in the art would deem appropriate in terms of user interface design, such as a tabular display, a tree-like hierarchical display, or even a configuration in which multiple pieces of information are simultaneously displayed on a single page. Furthermore, differences such as whether the information is displayed at the top or bottom of the screen, or presented as a pop-up window or modal dialog, do not affect the technical scope of the present invention.

[0161] Furthermore, the term "display step" broadly encompasses the visualization of specific data items in a user interface and is not necessarily strictly limited to a single operation or a single screen, as the term "display" suggests. In other words, "display" also encompasses methods that allow users to view displayed information by switching between them, as well as methods such as scrolling, accordion menus, and tab switching. The format and order of the displayed content, as well as the type of physical device (e.g., smartphone, tablet, or desktop terminal), are not important. In essence, the gist of the present invention resides in a mechanism that presents multiple processing steps to users in an easy-to-understand manner, allowing the user to grasp and operate information about processing entities and sub-processes. As a result, if the user experience and functional effects intended by the present invention can be achieved simply by changing the layout and names of screen items, all of these are within the technical scope of the present application. Therefore, the "display steps" used in the claims are not intended to narrowly limit the specific form of screen control, but rather encompass all variations that embody the essential feature of the present invention, "a configuration that visualizes information to the user and allows the operator to easily recognize, select, and confirm steps and processing entities."

[0162] Furthermore, the names "processing process information display step," "processing subject information display step," "sub-process information display step," etc. are merely exemplary names, and the present application equally protects forms that use a single screen that collectively implements these display functions, and forms that display multiple screens consecutively by screen transitions. In addition, the screen examples shown in this specification (Figs. 6 to 14) are also examples for easily explaining the technical features, and it goes without saying that even if the user interface configuration differs in detail from these drawings, it can constitute a claim of the present application as long as it comprises the essence of the present invention.

[0163] Furthermore, even if a user (a user operating a user terminal) performs some input operations or screen transitions, the entire process, including these operations, can be configured to be executed under the management and control of a service provider (the provider of the system or service). In other words, when a user submits input information or presses a button on a screen, the user accesses the system via a user interface and network path presented by the service provider's system, and the user's actions are, in effect, in response to requests or instructions from the system. In such a configuration, the service provider can be considered to have essentially implemented all steps comprehensively. Even if the user appears to be operating the terminal independently, these operations are performed according to the screen configuration and control logic (e.g., input form, selection buttons, screen flow) pre-set by the system of the present invention. Therefore, the system provider can be said to have oversight and control over the final process flow and its results. Furthermore, when a user is guided through a system screen or other means, the user's actions can be considered to be performed under the "direction or control" of the service provider, and the method steps defined in the present claims can be considered to be consistently performed by a single entity (the system provider).

[0164] For example, in the user interface-related steps shown in this specification, such as the "input unit display step," "processing process information display step," and "processing subject information display step," while the user views and operates the terminal screen, the screen is generated and provided by the provider's server or software, and the server (information processing device) forcibly / exclusively obtains operation data from the user and incorporates it into the processing flow. In this state, the user's input and operation themselves merely follow the processing scenario set by the system provider, and the provider essentially controls the series of method steps as a whole.

[0165] Therefore, even if some steps involve user terminal operations, if the provider manages the steps and provides them as a service as described above, this can be considered as an "implementation by a single entity" form of implementation from the technical perspective of a person skilled in the art. Of course, even if the steps involve individual user operations, this is the same if the provider (provider of information processing equipment) controls the entire system in an integrated and continuous manner using a predetermined flow design or access control method.

[0166] As described above, although the implementation of the present invention assumes that some steps include screen operations by the user, in reality, the provider's system comprehensively controls and manages the core parts of the processing, and therefore, the series of methods can be evaluated as being performed by a single entity. Therefore, for each step defined in the claims (information acquisition step, display step, processing result acquisition step, etc.), even if an operation process on the user terminal intervenes, the provider's system leads the overall flow by bundling them, and it goes without saying that this is included in the technical scope of the present application.

[0167] Furthermore, the information processing system of the present invention is configured to function as a unified invention even when multiple physical devices or service providers operate in cooperation with one another. Specifically, even if one entity (Server A) acquires input information and sets up processing steps, while another entity (Server B, an external agent, etc.) generates part of the processing results, the entire system can be said to substantially execute the series of steps described in the claims (e.g., processing entity setting step, processing result acquisition step, user interface display step) through the orchestration function that oversees these devices and entities. In other words, if the mechanism that "appropriately assigns a series of processing steps and performs branching and reprocessing based on input information and change instruction information," which is the essence of the present invention, is managed consistently across all devices and entities, then these together constitute the information processing system / method referred to in this application.

[0168] Furthermore, even if a service provider appears to host only the user interface, the access control policy and flow management module that control the back-end of the user interface (e.g., API calls to Server A, Server B, and external agents) may actually be provided by the same provider, or may be managed collectively through contractual relationships or service specifications. In such cases, even if the service provider claims to only provide the user interface, from the user's perspective, the entire service appears to be provided by a single contact point / management entity, and the series of flows in which inputs and instructions to the user interface are internally forwarded to other devices and agents essentially function as a single system. Therefore, even if processing is physically distributed across multiple hardware devices or cloud services, there are many cases in which the processing entity setting step, processing result acquisition step, and other user interface display steps of this application are performed under the overall management of the provider.

[0169] Furthermore, even in a configuration in which multiple entities (e.g., Server A, Operating Company B, External Agent C, etc.) share part of the processing, each role is incorporated as a means or functional module for executing part or all of the processing steps defined in this invention. In other words, if the processing flow and access control defined by the business operator are thoroughly implemented as a whole, and the timing of processing by which entity, including user requests and change instruction information, is designed in an integrated manner, this will be evaluated as a single embodiment integrating the steps defined in the claims. This comprehensive coordination ensures that the visualization of processing step information, branching and comparison of sub-steps, and even reprocessing are consistently under the control of the operating side, thereby satisfying the essential features of this invention.

[0170] Therefore, the components of this specification, such as "information processing device," "agent," and "user terminal," are not necessarily limited to a single piece of hardware or service. Even if implemented by combining multiple different physical servers or external services, as long as at least the steps recited in the claims are linked as a whole and ultimately produce a certain result (the integrated operation of the processing flow and user interface display intended by the present invention), it will be treated as a single information processing system that satisfies the requirements. For example, even if some steps are outsourced to a subsystem of another company, this does not interfere with the scope of this application as long as the outsourcing entity controls the entire flow and can collect, integrate, and provide the processing results.

[0171] As mentioned above, the present invention may physically include a form in which multiple businesses and devices cooperate with each other, but as long as the cooperative process functions as a method / system in which a business integrates and manages each step described in the claims of this application, it falls within the technical scope of the present invention as a whole. Even if it appears to provide only a "user interface" or only back-end processing, if all the steps described in the claims actually operate in cooperation and the service provider defines and controls the entire process within the same flow, it can be considered that the technical means intended by this application has been implemented.

[0172] The above-described embodiments shown in the specification and drawings are merely examples intended to facilitate understanding of the present invention as defined in the claims, and are not intended to limit the scope of the invention. While the present invention refers to a series of processing steps, such as a "processing step information display step," a "processing entity information display step," a "sub-process information display step," and a "change instruction information acquisition step," these names, screen layouts, and operational procedures are merely examples and are not necessarily limited to specific names or screen transition flows. Even if the names or user interface divisions differ, as long as the user can functionally view multiple processes, sub-processes, and the processing entities responsible for them, and enter information and change instructions, all of these are within the scope of the present invention. Even if sub-process information is displayed together on the screen in a separate window or a foldable menu, or multiple steps are executed together on a single screen or with a single operation, the function of "presenting relevant information to the user and accepting references and inputs" is essentially satisfied as defined in the claims. Therefore, the assertion that differences in graphical layout, display item names, menu layout, button names, etc. alone do not constitute the scope of the patent claims is unfounded. Furthermore, with regard to the order of steps and components enumerated in the claims, the focus of the present invention is "a mechanism for visualizing multiple processes and providing instructions for changes and reprocessing as necessary." As long as it has the functionality to achieve this, adding other processes or screen elements or rearranging some steps along the way is included in the technical components intended by the present application, as long as it does not impair the overall essence. Taking these points into consideration, even if a third party makes partial changes to the names, screen layout, or operating procedures, the implementation does not deviate from the scope of the present invention as long as it functionally satisfies the technical requirements as can be read from the claims.

[0173] <Additional Notes> The present embodiment includes the following disclosure.

[0174] (Appendix 1) An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a processing entity setting step of setting a processing entity that performs processing for each of a plurality of processing steps set for processing the input information; a processing result acquisition step of acquiring and outputting a processing result in the processing step by the processing subject; An information processing method including:

[0175] (Appendix 2) The processing entity setting step acquires change instruction information for the processing entity in the processing step, and changes the processing entity based on the change instruction information. 1. The information processing method described in Appendix 1.

[0176] (Appendix 3) The processing result acquisition step stores the processing result in a storage unit, and acquires a processing result using the processing result stored in the storage unit according to the processing status of the processing step. 1. An information processing method according to claim 1 or 2.

[0177] (Appendix 4) the processing result acquisition step acquires, upon acquisition of change instruction information for the output processing result, a processing result in the processing step based on the change instruction information; An information processing method according to any one of Supplementary Notes 1 to 3.

[0178] (Appendix 5) acquiring and outputting a processing result in the processing step based on the change instruction information and a processing result in the processing step before acquiring the change instruction information; 1. The information processing method described in Appendix 4.

[0179] (Appendix 6) An information processing method executed by an information processing device, an input unit display step of displaying an input unit that accepts input of input information on a user interface; a processing step information display step of displaying, on the user interface, a plurality of processing step information sets for processing the input information; a processing subject information display step of displaying processing subject information representing a processing subject of the processing step on the user interface; An information processing method including:

[0180] (Appendix 7) a progress status information display step of displaying processing progress status information of the processing step on the user interface, 1. The information processing method described in Appendix 6.

[0181] (Appendix 8) a processing monitor information display step of displaying processing monitor information on the user interface, the processing monitor information visualizing the processing according to the input information, 8. The information processing method according to claim 6 and claim 7.

[0182] (Appendix 9) a sub-process information display step of displaying sub-process information in the processing step on the user interface, An information processing method according to any one of Supplementary Notes 6 to 8.

[0183] (Appendix 10) a processing step edit display step of displaying a processing step edit screen for editing the processing step information on the user interface, An information processing method according to any one of Supplementary Notes 6 to 9.

[0184] (Appendix 11) a processing step branch display step of displaying, on the user interface, processing step information after the change on the processing step editing screen and the processing step information before the change, An information processing method according to any one of Supplementary Notes 6 to 10.

[0185] (Appendix 12) a notification display step of displaying notification information directed to the processing subject of the processing step on the user interface, An information processing method according to any one of Supplementary Notes 6 to 11.

[0186] (Appendix 13) In the information processing device, an input information acquisition step of acquiring input information; a processing entity setting step of setting a processing entity that performs processing for each of a plurality of processing steps set for processing the input information; a processing result acquisition step of acquiring and outputting a processing result in the processing step by the processing subject; A program for executing an information processing method including the steps of:

[0187] (Appendix 14) In the information processing device, an input unit display step of displaying an input unit that accepts input of input information on a user interface; a processing step information display step of displaying, on the user interface, a plurality of processing step information sets for processing the input information; a processing subject information display step of displaying processing subject information representing a processing subject of the processing step on the user interface; A program for executing an information processing method including the steps of:

[0188] (Appendix 15) An information processing system including an information processing device, an input information acquisition unit that acquires input information; a processing entity setting unit that sets a processing entity that performs processing for each of a plurality of processing steps that are set to process the input information; acquiring a processing result by acquiring and outputting a processing result in the processing step by the processing subject; An information processing system having the above.

[0189] (Appendix 16) An information processing system including an information processing device, an input unit display unit that displays an input unit that accepts input of input information on a user interface; a processing step information display unit that displays, on the user interface, a plurality of processing step information sets for processing the input information; a processing subject information display unit that displays processing subject information of the processing step on the user interface; An information processing system having the above. [Explanation of symbols]

[0190] 10 Information Processing Systems 12 servers (information processing devices) 14 User terminal (information processing device) 30 Input section 32 Treatment process information 34. Information on the processing subject 38 Processing progress information 40 Processing monitor information 50 Processing process editing screen 220 Program 242 Input information acquisition unit 244 Processing Subject Setting Department 246 Processing result acquisition unit

Claims

1. An information processing method executed by an information processing device, an input information acquisition step of acquiring input information; a processing entity setting step of setting in advance a processing entity that performs processing for each of a plurality of processing steps set to process the input information; a processing result acquisition step of acquiring and outputting a processing result in the processing step by the processing subject; An information processing method including:

2. The processing entity setting step acquires change instruction information for the processing entity in the processing step, and changes the processing entity based on the change instruction information. The information processing method according to claim 1 .

3. The processing result acquisition step stores the processing result in a storage unit, and acquires a processing result using the processing result stored in the storage unit according to the processing status of the processing step. The information processing method according to claim 1 .

4. the processing result acquisition step acquires, upon acquisition of change instruction information for the output processing result, a processing result in the processing step based on the change instruction information; The information processing method according to claim 1 .

5. acquiring and outputting a processing result in the processing step based on the change instruction information and a processing result in the processing step before acquiring the change instruction information; The information processing method according to claim 4.

6. An information processing method executed by an information processing device, an input unit display step of displaying an input unit that accepts input of input information on a user interface when the workflow is executed; a processing step information display step of displaying, on the user interface, a plurality of processing step information sets for processing the input information when the workflow is executed; a processing subject information display step of displaying processing subject information representing a processing subject of the processing step on the user interface when the workflow is executed; An information processing method including:

7. a progress status information display step of displaying processing progress status information of the processing step on the user interface, The information processing method according to claim 6.

8. a processing monitor information display step of displaying processing monitor information on the user interface, the processing monitor information visualizing the processing according to the input information, The information processing method according to claim 6.

9. a sub-process information display step of displaying sub-process information in the processing step on the user interface, The information processing method according to claim 6.

10. a processing step edit display step of displaying a processing step edit screen for editing the processing step information on the user interface, The information processing method according to claim 6.

11. a processing step branch display step of displaying, on the user interface, processing step information after the change on the processing step editing screen and the processing step information before the change, The information processing method according to claim 10.

12. a notification display step of displaying notification information directed to the processing subject of the processing step on the user interface, The information processing method according to claim 6.

13. In the information processing device, an input information acquisition step of acquiring input information; a processing entity setting step of setting in advance a processing entity that performs processing for each of a plurality of processing steps set to process the input information; a processing result acquisition step of acquiring and outputting a processing result in the processing step by the processing subject; A program for executing an information processing method including the steps of:

14. In the information processing device, an input unit display step of displaying an input unit that accepts input of input information on a user interface when the workflow is executed; a processing step information display step of displaying, on the user interface, a plurality of processing step information sets for processing the input information when the workflow is executed; a processing subject information display step of displaying processing subject information representing a processing subject of the processing step on the user interface when the workflow is executed; A program for executing an information processing method including the steps of:

15. An information processing system including an information processing device, an input information acquisition unit that acquires input information; a processing entity setting unit that pre-sets a processing entity that performs processing for each of a plurality of processing steps that are set to process the input information; a processing result acquisition unit that acquires and outputs a processing result in the processing step by the processing subject; An information processing system having the above.

16. An information processing system including an information processing device, an input unit display unit that displays an input unit that accepts input of input information on a user interface when the workflow is executed; a processing step information display unit that displays, on the user interface, a plurality of pieces of processing step information set for processing the input information when the workflow is executed; a processing subject information display unit that displays processing subject information indicating a processing subject of the processing step on the user interface when the workflow is executed; An information processing system having the above.

Images