Information processing device, information processing method, and program

Abstract

To visualize and quantitatively evaluate the legal framework for company housing (company housing regulations, company housing systems, etc.) within companies. [Solution] The server 10 includes an evaluation standard level setting unit 132, an evaluation target level calculation unit 133, and a regulations evaluation unit 134. The evaluation standard level setting unit 132 sets multiple items (major items, medium items, minor items) to be evaluated based on the respective company housing regulations of multiple targets (Company A to Company D), and calculates a standard level (for example, a standard value for evaluation in a certain area) for each of these multiple items. The evaluation target level calculation unit 133 extracts multiple items from the company housing regulations of the target company (for example, Company A) among the targets (Company A to Company D), and calculates the item level for that company (Company A) for each of these multiple items. The regulations evaluation unit 134 evaluates the company housing regulations of the target company (for example, Company A) and outputs a company housing operation evaluation report (see Figures 7 and 8) that describes the evaluation results.

Description

Technical Field

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

Background Art

[0002] In companies that have company housing, although there are legal systems such as company housing systems and company housing regulations, these legal systems vary from company to company, and in many cases, the level of each company is unclear. Conventionally, when an employee rents a real estate that the employee desires, there is a company housing system rental company housing contract system that can efficiently search for real estate suitable for the employee (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in such a conventional system, although it is possible to search for real estate suitable for an employee, at present, there is no way to evaluate how the company housing system of one's own company compares with that of other companies. In addition, even with changes in economic situations such as income and prices, there is no comparison to determine whether the criteria defined in the regulations are appropriate, so the system itself has become rigid.

[0005] The present invention has been made in view of such a situation, and aims to visualize and quantitatively evaluate the legal systems (company housing regulations, company housing systems, etc.) of company housing in enterprises.

Means for Solving the Problems

[0006] To achieve the above object, an information processing apparatus according to an embodiment of the present invention is Based on the prescribed regulations for each of the multiple targets, a means for setting evaluation criteria levels is used to define multiple items to be evaluated and to calculate a standard level for each of those items. A means for calculating the evaluation target level, which extracts the multiple items from the predetermined regulations of the target to be evaluated among the aforementioned targets, and calculates the item level for each of the multiple items, A regulation evaluation means that compares the item level of each of the multiple items to be evaluated with the standard level of each of the multiple items, and evaluates the predetermined regulation to be evaluated based on the result of the comparison, It is equipped with. An information processing method and program corresponding to the above-mentioned information processing apparatus according to one aspect of the present invention are also provided as an information processing method and program according to one aspect of the present invention. [Effects of the Invention]

[0007] According to the present invention, it is possible to visualize and quantitatively evaluate the legal system for company housing (such as company housing regulations and company housing systems) within a company. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows a service provided by a service provider that manages a server according to one embodiment of the information processing device of the present invention. [Figure 2] This figure shows the overall configuration of the information processing system, including the server that implements the service shown in Figure 1. [Figure 3] Figure 2 shows the hardware configuration of the server in the information processing system. [Figure 4] This is a functional block diagram showing the functional configuration of a server with the hardware configuration shown in Figure 3. [Figure 5] Figures 3 and 4 show how the company housing regulations acquired by the server have been classified and digitized according to specified items. [Figure 6] Figures 3 and 4 are flowcharts showing the processing related to the evaluation of company housing regulations on the server. [Figure 7]In the servers of FIGS. 3 and 4, it is a diagram showing an example of a company housing operation evaluation report. [Figure 8] It is a diagram showing the visualized evaluation results (numerical values, graphs, general evaluations, etc.) of the continuation of the company housing operation evaluation report in FIG. 7. [Figure 9] In the information processing system of FIG. 1, it is a diagram showing the overall configuration of the company housing regulation deviation value tool in the cloud environment. [Figure 10] It is a diagram showing an example of visualization of the statistical analysis results of company housing regulations executed by the servers of FIGS. 3 and 4. [Figure 11] In the servers of FIGS. 3 and 4, it is a diagram showing a detailed analysis example in the form of a flowchart of regulation items. [Figure 12] It is a diagram showing a display example of the integrated dashboard generated by the servers of FIGS. 3 and 4. [Figure 13] It is a diagram showing a detailed implementation example of the regional comparison function by the servers of FIGS. 3 and 4. [Figure 14] It is a diagram showing the versatility and multi-regulation application example of this system by the servers of FIGS. 3 and 4. [Figure 15] In the servers of FIGS. 3 and 4, it is a diagram showing an application example to other company regulations. [Figure 16] It is a diagram showing the detailed processing of automating the calculation of company housing usage fees by utilizing AI by the servers of FIGS. 3 and 4. [Figure 17] It is a diagram showing the details of the AI machine learning processing by the servers of FIGS. 3 and 4. [Figure 18] It is a diagram showing the details of the evaluation analysis processing by the servers of FIGS. 3 and 4. [Figure 19] It is a diagram showing the details of the result output and visualization processing by the servers of FIGS. 3 and 4. [Figure 20] In the servers of FIGS. 3 and 4, it is a diagram showing the initial introduction hearing and the integrated configuration of the AI automatic processing system.

Embodiments for Carrying Out the Invention

[0009] Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, referring to FIG. 1, a service to which a server of an embodiment of the information processing apparatus of the present invention is applied will be described. FIG. 1 is a diagram showing a service provided by a service provider that manages a server of an embodiment according to the information processing apparatus of the present invention.

[0010] The service provider is, for example, a company that manages real estate such as company housing, and manages company housing where employees live under a contract with a company. Further, the service provider analyzes and evaluates housing regulations obtained from a plurality of companies, and provides each company with a housing regulation evaluation service (hereinafter referred to as "this service") for evaluating the housing regulations.

[0011] Specifically, this service is provided as follows. That is, as shown in FIG. 1, this service has steps S1 to S5. The service provider previously obtains housing regulations from a plurality of companies with which management contracts have been concluded, such as Company A to Company D, and creates evaluation items for evaluating the housing system from the housing regulations of each company.

[0012] In this service, first, in step S1, the service provider sets a plurality of evaluation items for the evaluation target based on predetermined regulations of each of a plurality of targets (for example, Company A to Company D), such as housing regulations. Specifically, the service provider sets a plurality of evaluation items for the evaluation target based on a plurality of housing regulations obtained from each of Company A to Company D. As shown in FIG. 5, the evaluation items are hierarchically divided into major items, middle items, and minor items. The major items are, for example, 1. Modernity, 2. Smooth operation,... etc. The middle items for the major item of 1. Modernity are, for example, whether the rent ceiling and contract conditions specified for each area are appropriate. The minor items for the middle item of whether the rent ceiling and contract conditions specified for each area are appropriate are, for example, whether the rent ceiling is appropriate, whether the contract deposit ceiling is appropriate, whether the short-term cancellation liquidated damages and key money, etc. are set appropriately for the area, whether contract-inadmissible conditions are set for all of "structure", "scale", and "area", whether the additional service is paid by the company,... etc.

[0013] Next, in step S2, the service provider quantifies each of the set items obtained from each of the companies A through D (quantifying items that match the content of each evaluation item as "1" and those that do not as "0").

[0014] Step S3 involves the service provider extracting multiple evaluation items from the company housing regulations of company A through D, for example, company A, and calculating the item level for each of these items for each company. Specifically, the service provider evaluates Company A's housing regulations by quantifying each of the above evaluation items (for example, counting the number of "1" answers for each of the five sub-items to determine the score for the medium-level item). For example, if there are four "1" answers in each sub-item, the score for the medium-level item will be a total of 4 points.

[0015] Step S4 involves the service provider using each company's assessment results (numerical values ​​for five sub-items and higher-level major items) and their respective standard levels to calculate the standard score and rank of each company's employee housing system. Specifically, Company A has a standard score of 45 and rank F, Company B has a standard score of 50 and rank C, Company C has a standard score of 53 and rank C, and Company D has a standard score of 65 and rank S, etc. For example, in Ward R in Tokyo, Company D's company housing regulations have a standard score of 65, but in Ward Q in Tokyo, Company D's company housing regulations have a standard score of 45. The standard scores and ranks of each company vary depending on the standard level of the area where the company housing is located.

[0016] Step S5 involves the service provider providing advice to encourage revisions to the company's housing regulations (generating a report that includes an overall assessment and specific areas for improvement) based on the assessment results (numerical values ​​at the item level) and comparisons with other companies. Specifically, the service provider generates a company housing operation evaluation report that includes evaluation results (numerical values, standard scores, ranks, etc.), an overall assessment encouraging revisions to the company housing regulations, and specific areas for improvement, based on a comparison of the evaluation results of the company being evaluated with those of other companies. For example, the sales department of the service provider can then present this company housing operation evaluation report to the relevant department at the company to encourage revisions to the company housing regulations.

[0017] Thus, this service makes it possible to visualize and quantitatively evaluate company housing regulations and systems. Furthermore, by providing a general assessment that encourages revisions to company housing regulations and suggesting specific areas for improvement, it is possible to raise the standard of company housing systems among companies and create a constant environment of competition. In addition, companies can enjoy benefits such as preventing employee turnover and promoting recruitment by improving their company housing regulations. Another possibility is that, for example, companies that already have a company housing system or regulations can use this to revise their company housing regulations. For companies without a company housing system, it becomes possible to provide a one-stop service from the creation of company housing regulations to company housing management. Services such as data analysis and evaluation of the effects of revising company housing regulations (e.g., increase in company housing users) can also be considered. Furthermore, it becomes possible to collect new data within the company, which can then be used to develop new services.

[0018] Figure 2 shows the overall configuration of an information processing system according to one embodiment of the present invention. The information processing system shown in Figure 2 is configured such that a server 10, u enterprise terminals 20-1 to 20-u (where u is any integer value of 1 or more), and a service provider terminal 30 are interconnected via a predetermined network NW such as the Internet. Enterprise terminal 20-1 is operated by company representative K1. Enterprise terminal 20-2 is operated by company representative K2. Enterprise terminal 20-u is operated by company representative KU. Company representatives K1 through KU enter into a contract with a service provider for company housing management, and select company housing from the area desired by the employee for those who wish to live in company housing, and provide the housing to the employee. In the following, when it is not necessary to distinguish between each of the corporate terminals 20-1 through 20-u individually, they will be collectively referred to as "corporate terminal 20." Similarly, when it is not necessary to distinguish between each of the corporate representatives K1 through KU individually, they will be collectively referred to as "corporate representative K." The corporate terminal 20 is an information processing device managed by each corporate employee K, and consists of, for example, a tablet terminal or a smartphone. The service provider terminal 30 is a terminal managed by the service provider's representative and exchanges information with each of the corporate terminals 20-1 to 20-u via email or other means.

[0019] Server 10 is managed by the service provider. Server 10 controls the individual operations of enterprise terminals 20-1 through 20-u and performs various processes.

[0020] Figure 3 is a block diagram showing the hardware configuration of the server in the information processing system shown in Figure 1.

[0021] Server 10 includes a CPU (Central Processing Unit) 101, ROM (Read Only Memory) 102, RAM (Random Access Memory) 103, a bus 104, an input / output interface 105, an output unit 106, an input unit 107, a storage unit 108, a communication unit 109, and a drive 110.

[0022] The CPU 101 executes various processes according to the program recorded in the ROM 102 or the program loaded into the RAM 103 from the storage unit 108. RAM103 also stores data and other information necessary for the CPU101 to perform various processes.

[0023] The CPU 101, ROM 102, and RAM 103 are interconnected via a bus 104. An input / output interface 105 is also connected to this bus 104. An output unit 106, an input unit 107, a storage unit 108, a communication unit 109, and a drive 110 are connected to the input / output interface 105.

[0024] The output unit 106 consists of a display, speakers, etc., and outputs various information as images and sounds. The input unit 107 consists of a keyboard, mouse, etc., and is used to input various types of information.

[0025] The memory unit 108 consists of a hard disk, DRAM (Dynamic Random Access Memory), etc., and stores various types of data. The communication unit 109 communicates with other devices (in the example in Figure 1, the enterprise terminal 20) via a network NW, including the Internet.

[0026] A removable media 111, such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory, is appropriately mounted in the drive 110. Programs read from the removable media 111 by the drive 110 are installed in the storage unit 108 as needed. Furthermore, the removable media 111 can store various types of data stored in the storage unit 108, just as the storage unit 108 can store them.

[0027] Figure 4 is a functional block diagram showing the functional configuration of the server in the information processing system shown in Figure 1. As shown in Figure 4, one area of ​​the storage unit 108 (Figures 3 and 4) of the server 10 is equipped with an analysis infrastructure DB 151 and a learning model 152. The analysis platform DB151 stores information from the contracting company, such as corporate information for each company, contractor information, company housing regulations, evaluation items set from each company housing regulation, numerical values ​​for each evaluation item, standard scores, features, evaluation results, and company housing operation evaluation reports. Furthermore, the analysis platform DB151 stores information related to company housing under management contracts, such as contract information, property information (detailed information on rental properties (e.g., floor plan, rent data, etc.)), address information, rent market rates, inquiry information, and open data.

[0028] The learning model 152 is an AI model that improves evaluation accuracy by using machine learning on the input information for evaluating company housing regulations and the output evaluation results from the information stored in the analysis platform DB 151.

[0029] In the CPU 101 of server 10 (see Figures 3 and 4), when performing the evaluation process for company housing, the following functions are activated: the company regulations acquisition unit 131, the evaluation standard level setting unit 132, the evaluation target level calculation unit 133, the regulations evaluation unit 134, the advice unit 135, the learning unit 136, the display control unit 137, and so on.

[0030] The Corporate Regulations Acquisition Unit 131 acquires the corporate housing regulations of each of the companies being evaluated (for example, companies A through D in Figure 1). Specifically, the Corporate Regulations Acquisition Unit 131 acquires the corporate housing regulations from each corporate terminal 20 of the companies (for example, companies A through D, etc.) with which the service provider has a contract for the management of company housing, and stores them in the Analysis Platform DB 151.

[0031] The evaluation standard level setting unit 132 sets multiple items to be evaluated (major items such as "relevance to the times," "smooth operation," "welfare benefits," "utilization of company housing outsourcing," and "average operation," as well as their respective sub-items and minor items) based on the respective company housing regulations (regulations concerning company housing) of multiple companies, and calculates a standard level (for example, a standard value for evaluation in a certain region or area) for each of these multiple items. Specifically, the evaluation standard level setting unit 132 reads the company housing regulations of multiple companies stored in the analysis platform DB 151 and quantifies the standard level for each major item. For this region, it sets a standard or central value (reference score) out of a total of 20 points that would be considered acceptable.

[0032] The evaluation target level calculation unit 133 extracts multiple items from the company housing regulations of each company being evaluated (for example, companies A through D in Figure 1) and calculates the item level for each of those items for that company. Specifically, it quantifies the correctness of the five setting items of the sub-items (1 if correct, 0 if incorrect), and adds the numerical values ​​of the five sub-items at each level of the medium-item hierarchy to obtain the numerical value for that medium-item.

[0033] The regulations evaluation unit 134 compares the item level of each of the multiple evaluation items being evaluated with the standard level of each of the multiple items, and evaluates the company housing regulations of the company being evaluated based on the results of that comparison. Specifically, the regulations evaluation unit 134 compares the item levels of multiple evaluation items extracted from the evaluation target, for example, Company A's company housing regulations, with the respective benchmark scores of multiple evaluation items calculated in advance for each region, and evaluates Company A's company housing regulations based on the results of this comparison. Alternatively, instead of comparing at the item level, comparisons may be made based on the scores of major categories. In this case, the regulations evaluation unit 134 compares the total score derived from the item level of multiple evaluation items extracted from the company housing regulations of the company being evaluated (for example, Company A) with the benchmark scores for major categories of each company that have been calculated in advance for each region, and evaluates Company A's company housing regulations based on the results of this comparison. If the evaluation is to be performed by AI, for example, the numerical values ​​(standard scores and features) of each of the multiple evaluation items are input to the learning model 152, and the evaluation results are derived and output from the learning results of the relationship between the standard scores, features and evaluation results.

[0034] The advice unit 135 generates advice information that provides advice regarding Company A's company housing regulations, based on the standard score and at least some of the features of Company A being evaluated, as well as the learning results. Specifically, in this example, the advice unit 135 inputs the standard score and features of company A into the learning model 152, and the learning model 152 outputs advice information regarding company A's housing regulations, derived from the learning results, to the display control unit 137 via the advice unit 135. The advice information is output to the display control unit 137 as a company housing operation evaluation report (see Figures 7 and 8), which includes evaluation results such as messages prompting revision of company housing regulations. The company housing operation evaluation report includes, for example, graphs such as radar charts and distribution diagrams, numerical data (overall evaluation score, standard score), rank, overall assessment, and specific areas for improvement. In this way, the evaluation results of the company housing regulations of the companies being evaluated are visualized in graphs for each of the five major categories (e.g., "relevance to the times," "smooth operation," "employee benefits," "utilization of company housing outsourcing," and "average operation rate"), and the evaluation results are shown quantitatively using numerical values, standard scores, ranks, etc., making it visually (at a glance) clear which categories the company's housing regulations are biased towards.

[0035] The learning unit 136 uses the standard scores and features of multiple targets (Company A through Company D) to perform a predetermined machine learning operation on the learning model 152. When the regulations evaluation unit 134 performs an evaluation, it inputs the standard score and features of the subject to evaluation (for example, Company A) into the learning model 152, and outputs the learning results (evaluation results) from the learning model 152. Furthermore, the advice information generated by the advice unit 135 may also be used to train the learning model 152, so that it can output better advice information.

[0036] The display control unit 137 displays the information generated by the above-mentioned units or the information output from the above-mentioned units to each company terminal 20. The display control unit 137 displays, for example, the company housing operation evaluation report 71 shown in Figures 7 and 8.

[0037] Figure 5 shows an example of an evaluation item table, in which evaluation items are hierarchically divided into major items, medium items, and minor items, and points are assigned to each item. As shown in Figure 5, in this embodiment, multiple evaluation items to be evaluated are set in the evaluation item table based on the respective company housing regulations of multiple companies. For example, major categories might include "Relevance to the Times," "Smooth Operation," "Employee Benefits," "Utilization of Company Housing Services," and "Average Operational Performance." Each major category has subcategories and their evaluation criteria. Each subcategories has further subcategories and their evaluation criteria.

[0038] Next, with reference to Figures 6 to 8, the operation of the server 10 in this information processing system for evaluating company housing regulations will be explained. Figure 6 is a flowchart showing the company housing regulations evaluation processing operation by server 10 in Figures 3 and 4. Figure 7 shows an example of a company housing operation evaluation report output as a result of the processing by server 10 in Figures 3 and 4. Figure 8 is a continuation of the company housing operation evaluation report shown in Figure 7.

[0039] In server 10, in step 101, the corporate regulations acquisition unit 131 acquires the corporate housing regulations from multiple contracted companies, sets multiple items to be evaluated based on the corporate housing regulations of the acquired companies, and calculates a standard level for each of those items. Specifically, the Corporate Regulations Acquisition Unit 131 stores multiple corporate housing regulations 61 acquired from multiple contracted companies in the analysis platform DB 151, and also examines the acquired corporate housing regulations 61, extracting evaluation items 62 from each corporate housing regulations 61 and quantifying them. The corporate housing regulations of each company are then converted into data using the quantified evaluation items. Furthermore, a thorough review of company housing regulations 61 means extracting common elements (terminology, etc.) from multiple company housing regulations 61 and classifying them as evaluation items 62.

[0040] In step 102, the regulations evaluation unit 134 extracts multiple items from the regulations of the company to be evaluated and calculates an item level for each of those multiple items. Specifically, the Regulations Evaluation Department 134 calculates a standard score for each company's company housing regulations based on data from each company's company housing regulations.

[0041] In step 103, the regulations evaluation unit 134 can extract features based on examples of revised company housing regulations, thereby calculating which of the quantified evaluation items are distinctive. These features are calculated by AI.

[0042] In step 104, the learning unit 136 performs machine learning on the learning model 152 using the standard scores and features calculated by the standard evaluation unit 134, and updates the learning model 152.

[0043] In step 105, the regulations evaluation unit 134 compares the item level of each of the multiple items in the company housing regulations of the company under evaluation with the standard level of each of the multiple items, and evaluates the company housing regulations of the company under evaluation based on the results of that comparison. Specifically, the regulations evaluation unit 134 visualizes the evaluation (analysis results) of the company housing regulations of the target company (for example, Company A) using the updated learning model 152 (results of machine learning). To visualize the evaluation (analysis results) of company housing regulations, data analysis tools such as BI tools are used. The regulations evaluation unit 134 uses the results of machine learning to analyze the level of company housing regulations of company A being evaluated (comparison with standard levels and company housing regulations of other companies, etc.), and uses the results of the analysis as advice information to create, for example, a company housing operation evaluation report. This makes it possible to visualize the evaluation of company housing regulations using standard scores and features, as well as the results of analysis using machine learning.

[0044] As shown in Figure 7, the company housing operation evaluation report 71 displays the evaluation level (the company's score) for evaluation items extracted from the company housing regulations in a table format. In addition, as shown in Figure 8, it also shows the standard score of the company housing regulations (e.g., standard score: 53), the overall evaluation showing how many points out of 100 (e.g., 59 / 100), the rank (e.g., C rank out of A rank (highest) to E rank (lowest)), and the scores for each major category (relevance to the times, smooth operation, level of welfare benefits, utilization of company housing outsourcing, average operation). Furthermore, the company housing operation evaluation report 71 also shows the actual state of company housing operation in the form of a radar chart, with scores for each major item. This allows for a quick overview of how the company's company housing regulations deviate from the standard level and how they differ from those of other companies. In addition, the company housing operation evaluation report 71 includes an overall assessment in text format. The overall assessment outlines specific areas for improvement in the company housing regulations.

[0045] In step 106, the advice unit 135 can encourage company A to revise its company housing regulations by providing the company housing operation evaluation report 71 generated by the regulations evaluation unit 134 to the company being evaluated (for example, company A) (displaying it on the company terminal 20). Furthermore, the sales department of the service provider will present Company A with a standard score of 63 for its company housing regulations, thereby encouraging Company A to revise its company housing regulations. This will help support the revision of Company A's company housing regulations. Furthermore, by providing the machine learning results64 to the service provider's sales department, the sales department can use this information to select companies that are more likely to revise their company housing regulations, thus supporting the selection of companies likely to revise their regulations.

[0046] As described above, the server 10 of this embodiment sets multiple items (major items, medium items, minor items) to be evaluated based on the respective company housing regulations of multiple companies such as Company A, Company B, Company C, and Company D, calculates a standard level (for example, a standard value for evaluation in a certain area) for each of these items, extracts multiple items from the company housing regulations of the company to be evaluated, calculates the item level of that company for each of these items, evaluates the company housing regulations of the company to be evaluated, for example, Company A, and outputs a company housing operation evaluation report (see Figures 7 and 8) that describes the evaluation results. In this way, the company housing regulations of a company can be visualized using numerical values, tables, graphs, etc., and can be evaluated quantitatively through quantification. Furthermore, the results of machine learning can be used to select companies that are more likely to revise their company housing regulations.

[0047] In addition, when a service provider enters into a company housing management contract with a company, the service provider can obtain information about the contracted company housing, such as contract information, property information, address information, rent market rates, inquiry information, and open data. By storing this information in the analysis platform DB151 and having the regulations evaluation unit 134 analyze it, the following analyses can be performed.

[0048] For example, it becomes possible to compare rents with those of other companies in the same industry. Since the "Company Housing Company Code" and "Company Housing Management Identification Company Name" are available as data, it becomes possible to compare rents with other companies in the same industry by deriving the industry classification code (medium classification) from the company name and linking them. For example, if the "Company Housing Company Code" is 10000, the "Company Housing Management Identification Company Name" is ○× Beer Co., Ltd., and the industry classification code is: https: / / www.hellowork.mhlw.go.jp / info / industry_list02.html (It is assumed that ○× Beer Co., Ltd. will be classified as "Manufacturing 10" = "Beverage, Tobacco, and Feed Manufacturing.") Furthermore, it is also possible to consider obtaining the necessary data from the corporate master data.

[0049] For example, it becomes possible to compare rents by region. Because we have data on "postal code," "prefecture_property address," and "city / ward / town / village_property address," it is possible to compare rents by region.

[0050] You can compare rental rates from other companies in the same industry across different regions. By using comparative data such as rent comparisons with other companies in the same industry and rent comparisons by region, it is possible to compare rents of competitors in the same industry by region.

[0051] It can calculate the distance between your workplace and your rental property. By using the "residential area code" and "residential area" data, it is possible to determine which branch of the company (head office, branch office, factory, etc.) the contractor works at, thus identifying the location where the contractor works. Once the workplace is identified, the distance from the rental property can be calculated, making it possible to infer things like whether the contractor chose a property far away to keep rent down.

[0052] This allows us to visualize biases in contract signing and cancellation timing. By using "contract date" and "cancellation date," it is possible to visualize the bias in contract and cancellation timing by industry. Furthermore, the following inferences become possible. For example, if there is a significant gap between the contract signing date and the move-in date in some areas, it may indicate that there are very few moving companies in that region. Alternatively, it could be inferred that the timing of the move is not the usual due to unavoidable business events related to the industry.

[0053] It is possible to analyze various expense items at the management company and real estate agent level. Analyzing data not only at the contracting company level, but also at the management company and meeting service provider level, may reveal new insights and characteristics.

[0054] It can calculate average rent based on living area and room size. We believe it is possible to calculate average rents based on living area and room size. Furthermore, by cross-referencing the rent data from our analysis platform DB151 with information published on other rental property websites, we can present more convincing rent settings.

[0055] It is possible to see trends in rent based on the age and family structure of the tenant. If data such as age and family structure can be linked, providing information tailored to the age group and family structure (single, married, with children, etc.) of the policyholder would be effective.

[0056] This allows you to understand the relationship between the property's facilities and conditions and the rent. We will focus on the relationship between property amenities and conditions (such as the presence or absence of parking, whether pets are allowed, and whether there is a balcony) and rent, to understand how much these amenities and conditions affect rent, and use this as a reference for setting rent. By including energy efficiency and eco-related amenities (such as solar panels and the presence or absence of insulation) in the analysis, we believe it will be possible to grasp the trends of environmentally conscious properties.

[0057] This allows us to understand the rental trends by occupation of the tenants. If we can collect and link data on the occupations of our tenants, we believe it will be possible to analyze rental trends based on their occupations (sales, technical, management, etc.). By understanding rental and property selection trends based on occupation, we can consider setting rental rates accordingly.

[0058] This allows us to understand the percentage of clients who work from home and their property selection trends. By collecting and linking data on the occupations of tenants, it becomes possible to analyze rental trends based on tenants' work styles (commuting to the office, working from home). A difference in rent is likely to exist between commuters and remote workers, and given the current trend of remote work becoming commonplace, it would be possible to identify demand for properties specifically tailored to remote workers (e.g., large living rooms, study spaces).

[0059] This allows you to understand the relationship between contract renewal rates and rent. By collecting and linking contract renewal rate data with rent fluctuation data, it will be possible to understand the impact of rent fluctuations on contract renewal rates, which will lead to appropriate rent setting and prevention of churn. Furthermore, if a property has a high renewal rate even with high rent, the reasons for this can be analyzed and applied to other properties or areas.

[0060] Furthermore, in this embodiment, a function can be added to automatically respond to employee inquiries using AI (for example, the AI ​​chatbot system shown in Figure 20). Specifically, the server 10 receives inquiries from employees and, based on the employee's attribute information and the prescribed regulations to be evaluated (for example, company housing regulations), it further includes a response unit (for example, an AI response function that extends the advice unit 135 in Figure 4) that automatically responds using artificial intelligence based on the conditions stipulated in the prescribed regulations. This response unit automates inquiries from employees regarding regulations, which were previously handled individually by company representatives, resulting in a significant improvement in operational efficiency. This automates a large portion of the inquiry handling tasks that previously occupied the majority of the employee's working time, and significantly reduces the average response time compared to traditional manual responses. This response unit operates on the CPU 101 of the server 10 shown in Figure 4 and is implemented by utilizing the company housing regulations data and learning model 152 stored in the analysis platform DB 151. Employee inquiries are received through communications unit 109 and analyzed by a natural language processing engine. The analyzed inquiry content is classified into categories such as "rent calculation," "contract terms confirmation," "procedures," "property search support," and "procedures for relocation" using Intent Classification, and the corresponding regulations and clauses are automatically extracted.

[0061] More specifically, the normalization process for employee attribute information converts company-specific job titles into standard job levels and classifies family structures into fixed patterns.

[0062] This normalization process absorbs differences in company-specific organizational structures and work arrangements, enabling accurate determination of regulation application through unified attribute management. Furthermore, regarding work location information, the system automatically determines the nearest station, regional classification, and transportation convenience from address information, and utilizes this information to apply regional regulation conditions.

[0063] The response unit uses artificial intelligence to calculate and respond with the maximum rent and contract terms stipulated in the prescribed regulations, based on attribute information including at least one of the employee's job title, family structure, and place of work. For example, as shown in Figure 20, when an employee asks a question, the AI ​​gathers the necessary information step by step. First, it asks the employee to select the type of relocation, then it sequentially confirms their job title, whether they own property, the location of the property, and rent information, and finally provides a concrete calculation result immediately. This AI response system goes beyond simple calculation functions and automatically performs rule interpretations that include complex conditional branching. For complex conditions, it utilizes the hierarchical structure of evaluation items set in the evaluation criterion level setting unit 132, and determines the final applicable conditions through a process of condition determination at the sub-item level, followed by aggregation to medium and major items.

[0064] Furthermore, the AI ​​estimates potential needs from employee inquiries and provides relevant information. For example, in response to inquiries about rent calculations, it automatically provides additional information such as recommended nearby properties, important points to consider when signing a contract, suggestions for optimizing the timing of moving, and conditions for combining with housing allowances. This allows employees to instantly check the company housing regulations tailored to their own circumstances, contributing to a reduction in the workload of company staff. In addition, since it is available at any time, employee convenience is greatly improved. This function utilizes the evaluation item database generated by the regulations evaluation unit 134 to perform matching processing between individual employee attributes and regulations clauses. To improve matching accuracy, a machine learning function is also incorporated that continuously reflects past inquiry history and response satisfaction data into the learning model 152.

[0065] Regarding work locations, the system automatically identifies the nearest station, regional classification, and transportation convenience score based on address information. For extracting regulatory conditions, normalized employee attributes are used as keys to identify the relevant regulatory clauses. If multiple conditions are combined, Boolean logic (AND, OR, NOT operations) is used to narrow down the relevant clauses, and the applicable clauses are determined according to priority.

[0066] Figure 9 shows the overall system configuration of the company housing regulations deviation score tool in this embodiment. This Figure 9 corresponds to an advanced form in which the overall system configuration of Figure 2 is implemented in a cloud environment. As shown in Figure 9, the company housing regulations deviation score tool is built on AWS (Amazon Web Services®) 200, a cloud computing environment, and includes key components such as an AI system 201, a data warehouse 202, and a Tableau® DB 203. This cloud infrastructure enables scalable and highly available system operation, and can handle simultaneous access from numerous companies. AI System 201 incorporates machine learning algorithms to perform advanced processing such as automated analysis of company housing regulations, extraction of evaluation items, calculation of standard scores, and feature analysis. Large-scale data processing, which was difficult in traditional on-premises environments due to processing capacity limitations, is now possible thanks to the distributed processing environment of the cloud. In contrast to server 10, this embodiment employs a cloud-native microservices architecture. This allows AI system 201 to operate as an independent container, enabling automatic scaling based on load. Specifically, a Kubernetes® cluster is built on Amazon EC2® instances, and pods are automatically increased or decreased according to the demand for evaluation processing. Amazon RDS® and Amazon S3® are used for data persistence to ensure high availability and data redundancy. The API gateway function provides unified management of access from 20 corporate terminals and offers security features such as authentication and authorization processing, rate limiting, and log management. In addition, monitoring functionality using Amazon CloudWatch provides real-time monitoring of system operation status, performance metrics, and error occurrences, and automatically sends alert notifications when anomalies are detected.

[0067] In terms of security measures, we have implemented WAF (Web Application Firewall) to defend against application-level attacks, DDoS attack countermeasures, SSL / TLS encrypted communication, network isolation using VPC, and detailed access control using IAM. We also meet legal requirements such as SOC2 Type II compliance certification, GDPR compliance, and compliance with the Personal Information Protection Act.

[0068] Furthermore, through integration with Rilonet 205, company representatives can access both the company housing regulations evaluation results and the AI ​​automated response system from a unified web interface. Rilonet 205 seamlessly integrates with existing corporate authentication systems (Active Directory, LDAP, SAML, etc.) through its single sign-on (SSO) functionality, resulting in improved usability. In disaster response, we ensure the necessary recovery systems are in place and guarantee business continuity.

[0069] Data warehouse 202 automatically integrates with the company's existing system 204 to acquire necessary data in real time. Furthermore, by pre-setting company-specific housing regulations data, DIAG information, and various other tables, the evaluation process can be streamlined. It supports a variety of interfaces for integration with existing systems. Regarding data formats, the system can handle a wide range of formats, from standard formats such as fixed-length text to custom formats specific to each company, thanks to its automatic conversion function.

[0070] Furthermore, the system incorporates a configuration that automatically retrieves open data from the web via API agreements or web scraping, as needed. This ensures that evaluations always reflect the latest market data and statistical information. Specifically, dynamic updates of evaluation criteria are achieved through automatic integration with government statistics APIs (e-Stat), real estate information site APIs, economic indicator APIs, official land price data, demographic data, and other sources.

[0071] Tableau® DB203 functions as a business intelligence tool, visualizing acquired and processed data in a variety of formats. With a rich array of visualization options including radar charts, pie charts, heatmaps, and map displays, it provides dashboards that are intuitively easy for business professionals to understand.

[0072] Data Warehouse 202 is a significantly expanded version of the analytics platform DB151. Based on Amazon Redshift, the data warehouse platform enables petabyte-scale data processing. AWS Glue® is used for ETL (Extract, Transform, Load) processing, enabling automated data acquisition, transformation, and loading from various data sources.

[0073] Data quality management maintains a high-quality data infrastructure through features such as data profiling (automatic detection of data distribution, missing values, and outliers), data lineage (complete tracking of data flow and transformation history), and data cataloging (integrated management of metadata).

[0074] The AI-powered automated data acquisition function utilizes AI-based automated data processing technology to automate the regular collection of external data. This AI-powered automated data processing technology enables unattended execution of information gathering from websites, file downloads, and data conversion.

[0075] Our machine learning pipeline utilizes managed services such as Amazon SageMaker®, AWS® Lambda, and Amazon® Step Functions to automate a series of processes including data preprocessing, feature engineering, model training, model evaluation, and model deployment. This creates a continuous learning system that automatically retrains and updates the model in response to the addition of new data or changes in specifications.

[0076] When we receive a formal order from a client company, the tool administrator will set access permissions for each company, and the dashboards built with Tableau® will be made viewable from the web application or the Relonet web platform. This access control feature provides an environment where each company can securely view only its own evaluation results while preventing the leakage of confidential corporate information. Furthermore, it is possible to restrict available functions in stages according to the size of the company and the terms of the contract. Specifically, it offers tiered service provision such as a basic plan (viewing evaluation results only), a standard plan (including comparative analysis functions), and a premium plan (including AI automated response functions and custom report creation functions). Access control is achieved through a multi-layered security configuration utilizing AWS® Identity and Access Management (IAM®) and Amazon® Cognito. A unique tenant ID is issued for each company, ensuring complete separation of data and applications. Authentication methods supported include SAML 2.0, OpenID Connect, OAuth 2.0, and multi-factor authentication (MFA), and single sign-on (SSO) integration with existing corporate authentication systems is also possible.

[0077] For access control, role-based access control (RBAC) allows for granular permission settings based on roles such as "system administrator," "department administrator," "general viewer," and "external auditor." Furthermore, additional security features such as IP address restrictions, device authentication, session management, and login attempt restrictions prevent unauthorized access.

[0078] The audit log function records all access history, operation history, and data download history, and integrates with AWS CloudTrail to maintain a complete trail of evidence necessary for security audits. To prevent log tampering, hash chain technology is employed, and a tampering detection function is also implemented.

[0079] Furthermore, it implements features to comply with international privacy regulations, ensuring the appropriate handling of personal information. Features such as data anonymization, consent management, erasure rights, and data portability meet the compliance requirements of global companies.

[0080] In terms of enterprise-level data isolation, a multi-tenant architecture completely prevents unauthorized access to other companies' data. Data from each company is logically and physically isolated, and encryption keys are managed separately for each company. Furthermore, regular penetration testing, vulnerability assessments, and security audits ensure the continuous maintenance and improvement of security levels.

[0081] Figure 10 shows an example of statistical analysis that visualizes the results of the company housing regulations analysis in flowchart format. This Figure 10 visually represents the analysis results by the evaluation standard level setting unit 132. Figure 10 shows the results of a statistical analysis regarding the setting of maximum rent in company housing regulations. The statistical analysis department 210 revealed that many companies set maximum rent within their company housing regulations, and of those, a large number of companies set the maximum rent in a way that varies depending on the conditions. This statistical analysis goes beyond simple percentage calculations, conducting multifaceted analyses such as industry-specific analysis, company-specific analysis, and regional analysis. This analysis reveals a clear ratio between companies that have provisions regarding the setting of maximum rent in their company housing regulations and those that do not. Further detailed analysis clarifies that some companies have a fixed maximum rent, while many others have a variable maximum rent that varies depending on the conditions.

[0082] The percentage display unit 211 displays these figures in a visual format, making it easier for company representatives to compare their company's situation with the industry as a whole and with other companies. The color-coded pie chart allows for a quick understanding of the ratio of "yes" and "no" responses, clearly showing your company's position. Furthermore, multi-layered analysis results, such as comparisons with competitors in the same industry, companies of similar size, and companies in the same region, are displayed in parallel.

[0083] The statistical analysis unit 210 incorporates an advanced statistical analysis engine that extends the learning model 152. Furthermore, the time-series analysis function allows for quantitative analysis of the patterns of changes in company housing regulations. Various analytical methods enable trend prediction and factor analysis of regulation changes. Specifically, we visualize time-series data such as changes in rent caps over time, trends in regional disparities, and changes in industry-specific characteristics, and use this data to predict future regulation revisions and their content.

[0084] The machine learning approach improves the accuracy of predefined classification through ensemble learning, which combines multiple algorithms. Cross-validation (5-fold cross-validation) prevents overfitting and ensures generalization performance to new company data. Prediction accuracy has improved significantly from traditional manual classification to automated classification using machine learning. Furthermore, based on past inquiries, it achieves highly accurate responses that are difficult to provide with typical AI systems. This database enables automatic identification of new policy patterns, benchmark analysis of similar companies, and prediction of industry trends.

[0085] Figure 11 shows the analysis results in a more detailed flowchart format. This Figure 11 is an evolution of the item-by-item evaluation process by the Regulations Evaluation Unit 134. The flowchart display unit 220 visualizes each item in the company housing regulations in a flowchart format that includes conditional branching, and statistically displays the selection ratio of multiple options at each branch. For example, for the item "There is a provision regarding keeping pets in company housing," it visually shows that companies are divided in a ratio of 72%:28%. This flowchart analysis handles not only simple binary branching but also complex multi-stage branching. Taking pet ownership regulations as an example, the analysis progresses hierarchically: the first stage is "whether or not regulations are documented," the second stage is "conditions for permission to keep pets" for companies that have such regulations, and the third stage is further detailed conditions. Further detailed branching reveals that the conditions for allowing pets are subdivided into categories such as "supervisor approval required," "depending on rank," and "other conditions," ultimately classifying specific operational patterns such as "supervisor approval required, but pets are allowed in company housing," "pets are allowed in company housing depending on rank," "pets are allowed under conditions other than rank and supervisor approval," and "pets are allowed unconditionally."

[0086] The statistical value display unit 221 displays the statistical values ​​at each branching point as numerical values ​​and percentages, and also displays the results of the statistical significance test.

[0087] The comparison comment section 222 automatically generates comprehensive comments.

[0088] The flowchart display unit 220 displays an appropriate flowchart based on the appropriate theory. In the statistical analysis of conditional branching, we apply multidimensional analysis techniques for categorical data. Furthermore, we use advanced methods to calculate the semantic similarity of regulation documents, enabling automatic classification and comparative analysis of similar regulations. In statistical inference, appropriate statistics are inferred using a variety of statistical methods. The filtering functions, such as specifying industry or the number of employees, allow for detailed comparisons with competitors based on industry classification codes (major, medium, and minor classifications of the Japan Standard Industrial Classification). From major classifications such as manufacturing, information and communications, wholesale and retail, finance and insurance, real estate, and services, to specific medium and minor classification levels, a more precise comparison and evaluation of competitors is achieved.

[0089] Figure 12 shows an example of the integrated dashboard display. This Figure 12 is an interactive dashboard system that significantly expands the output results from the display control unit 137. The integrated dashboard 230 is a comprehensive evaluation result display system that includes a radar chart section 231 and an evaluation details section 232.

[0090] The radar chart section 231 visualizes the characteristics of the companies being evaluated using a radar chart, allowing for a quick and easy understanding of their features. The radar chart allows for a visual understanding of relative positions. Furthermore, the trend display function allows for the confirmation of time-series changes in each evaluation item. Color coding makes it easy to identify areas for improvement at a glance. In addition, a quantitative evaluation result is displayed as part of the overall assessment. This allows companies to objectively understand where their company housing system stands within the industry. Specific ranking information is also provided.

[0091] The detailed evaluation section 232 displays detailed comments for each item, suggestions for improvement, and comparison results based on industry and number of employees. It provides analysis results from multiple perspectives, offering useful information for companies to consider specific improvement measures.

[0092] Integrated Dashboard 230 is a web-based interactive dashboard system that runs on Tableau® Server. Its responsive design, utilizing HTML5, CSS3, and JavaScript (ES6+), ensures optimal display across various devices, including desktop PCs, tablets, and smartphones. For data visualization, we utilize the latest libraries and provide interactive features such as dynamic graph updates, zoom and pan operations, data drill-down, and time-series animations. The radar chart implements advanced analytical functions such as changing the weighting of each axis, dynamically adjusting the baseline value, displaying time-series comparisons, and overlaying data with competitors.

[0093] In the evaluation details section 232, explanatory text is automatically generated from numerical data using natural language generation (NLG) technology. By utilizing large-scale language models such as OpenAI® GPT-3.5®-turbo and GPT-4®, we automatically generate personalized comments and improvement suggestions tailored to the characteristics of each company. Furthermore, multilingual support allows for multilingual display.

[0094] The real-time update function uses WebSocket communication to instantly reflect data updates on the dashboard. Furthermore, a push notification function utilizing Server-Sent Events (SSE) automatically sends alerts when important changes (rank up / down, industry ranking changes) or anomalies are detected. The export function allows for report output in various formats, including PDF® (A4 / A3 size compatible), Microsoft Excel® (xlsx® format), PowerPoint® (pptx format), and PNG / SVG images. Customizable templates (company logo, color theme, layout adjustments) support corporate branding.

[0095] Furthermore, in the "Comments from Rilo" section displayed on the right side of Figure 12, personalized advice from the applicant's expert consultant is automatically generated. This provides not only numerical evaluations but also practical improvement suggestions (such as methods for securing budgets, templates for internal approval documents, and proposed implementation schedules).

[0096] Figure 13 shows a detailed implementation example of the regional comparison function. This Figure 13 is an example of the regional analysis function performed by the regulations evaluation unit 134. The map display unit 240 displays a color-coded comparison of your company's maximum rent and the average maximum rent of RJ's existing client companies on a map of Japan. Three colors—red, yellow, and blue—visualize the differences in competitiveness by region, and the intensity of the colors further expresses the magnitude of the difference. This visual representation allows for a quick understanding of the differences in competitiveness between regions. The map allows for interactive operation via clicks and taps, and selecting a specific prefecture or municipality will display detailed information in a pop-up window. The zoom function allows for step-by-step detail, from a nationwide view to the prefecture level, municipality level, and even down to the town / district level.

[0097] The comparison table section 241 displays detailed numerical data by region in a tabular format.

[0098] The map display unit 240 is a map display system. The regional divisions support hierarchical display, providing information at an appropriate level of granularity according to the zoom level. The color-coding display allows users to select a classification method that converts continuous values ​​into discrete color classes. This design takes color vision accessibility into consideration, ensuring that the color scheme is easily distinguishable for users with all types of color blindness. In statistical analysis, we apply spatial statistical methods. We use advanced statistical techniques to perform quantitative evaluations of regional factors, etc. The comparison table section 241 displays statistical data corresponding to regional classifications in a hierarchical format. Analysis with selectable levels of detail is possible. Furthermore, by linking it with regional characteristic data, factor analysis of rent disparities is also realized.

[0099] The time series analysis function is currently performing the analysis.

[0100] External data integration ensures that analysis always reflects the latest market data. To comply with API limitations, data collection is efficient through optimization of data acquisition intervals and implementation of caching functionality. Furthermore, the comparison function, limited to competitors, extracts only companies in the same industry based on their industry classification codes, providing a more precise competitive analysis. It also supports the formulation of regional strategies that reflect the characteristics of various industries, such as manufacturing, information and communication, wholesale and retail, and finance and insurance.

[0101] Figure 14 illustrates the versatility and scalability of this system. This figure specifically demonstrates that the information processing device can be applied to various standardized evaluations.

[0102] The evaluated process flow section 250 demonstrates that it is applicable not only to company housing regulations but also to all corporate regulations (information security regulations, employment regulations, overseas assignment regulations, business trip expense regulations, telecommuting regulations, personnel and pay regulations, compliance regulations, environmental management regulations, etc.). This versatility allows companies to comprehensively evaluate and manage multiple regulations within a single system, resulting in increased efficiency and cost reduction in regulations management operations. The automatic setting function for evaluation items according to each type of regulation automatically sets the optimal evaluation axis according to the characteristics of the regulation. For example, for company housing regulations, the evaluation axis is "relevance to the times, ease of operation, level of welfare, degree of utilization of outsourcing, and average operational performance"; for information security regulations, it is "relevance to the times, effectiveness, comprehensiveness, continuity, and responsiveness"; and for employment regulations, it is "compliance with laws and regulations, clarity, comprehensiveness, effectiveness, and flexibility to the times."

[0103] The structuring processing unit 251 extracts and normalizes the specified items and converts them into an evaluable structure. Specifically, it applies various structuring methods to convert regulations written in different formats into a format that can be processed uniformly. In regulation analysis using natural language processing, important information (amounts, periods, conditions, procedures, etc.) is automatically extracted from regulation documents using various language processing techniques. Furthermore, the structure of the regulations is automatically understood through document classification (sentence types: conditional statements, procedural statements, definition statements, exception statements, etc.), and appropriate items are allocated to evaluation criteria.

[0104] The analysis method section 252 applies various analysis algorithms. These methods function complementaryly to each other, realizing a general-purpose evaluation system that does not depend on specific computational logic. The evaluation target process flow unit 250 has a modular design that generalizes the evaluation criterion level setting unit 132. Plug-in functionality tailored to each regulation type allows for dynamic loading of regulation-specific evaluation criteria and item definitions. Furthermore, a proprietary method automatically identifies relationships and common items between different regulations, enabling integrated evaluation.

[0105] The structured processing unit 251 performs automatic analysis of the regulations document using natural language processing technology.

[0106] The analysis method section 252 executes different analysis methods in parallel.

[0107] In quality assurance, we implement complete version control and experimental control of the analytical process to ensure consistency and reproducibility of standard analyses.

[0108] In this embodiment, a general-purpose mechanism independent of specific regulations or methods is provided, offering an overall process for the visualization and quantitative evaluation of regulations. This enables the realization of an integrated evaluation system that can accommodate various types of regulations.

[0109] Figure 15 shows details of the application to other regulations. This Figure 15 shows an example of applying the evaluation method used for the company housing regulations to other company regulations.

[0110] The Regulations Evaluation Department 260 performs specific evaluations of information security regulations using multiple evaluation axes, such as overall evaluation score, standard score, and rank. Specifically for each evaluation axis, the timeliness assesses the response to the latest threats, the effectiveness assesses the actual operational status of the regulations, and the comprehensiveness assesses the degree to which necessary items are covered. The employment regulations use multiple axes for evaluation, such as overall score, standard score, and rank. Compliance with laws and regulations checks the degree of conformity with relevant laws and regulations, while clarity evaluates how easy the provisions are to understand (readability index, usage rate of technical terms, etc.). The overseas assignment regulations require a detailed evaluation of multiple items. In the area of ​​life support, we evaluate the support items provided, and in the area of ​​salaries and allowances, we verify their appropriateness.

[0111] The regulation-specific display section 261 provides dedicated evaluation items and dashboards for each regulation, and displays customized analysis results according to the type of regulation.

[0112] The Regulations Evaluation Unit 260 extends the Regulations Evaluation Unit 134 and implements specialized evaluation criteria for each type of regulation. It also enables risk assessments that reflect the latest threat intelligence. In evaluating employment regulations, a check for consistency with relevant laws and regulations is automatically performed. By linking with a database of legal revision history (e-Gov legal search API integration), the system automatically calculates the legal risk assessment of the regulations and the recommended timing for updates. Furthermore, by cross-referencing with a database of labor dispute cases, it is possible to identify clauses that pose a high risk. In evaluating overseas assignment regulations, we conduct risk assessments for each destination country by linking with overseas information. Furthermore, the automatic acquisition of relevant information enables us to assess the appropriateness of allowance amounts.

[0113] The regulation-specific display section 261 provides customized visualizations tailored to the characteristics of each regulation. For multilingual support, the system enables the display of evaluation results in multiple languages ​​through an automatic translation function for regulations documents. Our business trip expense regulations include comparative evaluations of items such as transportation expense limits, accommodation expense limits, and daily allowances against regional market rates and standards set by competitors. Our telecommuting regulations include evaluation items that reflect new ways of working, such as work time management, communication environment setup, security measures, and labor management systems, enabling analyses that emphasize contemporary relevance.

[0114] Figure 16 shows the detailed processing of the first stage of the automated process for calculating company housing usage fees using AI. This figure represents a significant automation of manual processes.

[0115] In the conventional implementation flow, the company housing regulations and operational rules had to be received from the company in advance, and the implementation staff had to manually read them and extract the points necessary for calculating usage fees. However, the number of items necessary for calculation that could be gleaned from the regulations was not large, and the remaining items had to be confirmed through interviews with company representatives. This manual process relied heavily on the experience and knowledge of the implementation staff, leading to issues such as variations in quality due to reliance on individual expertise and prolonged processing times.

[0116] In the workflow after AI implementation, upon receiving company housing regulations and operational rules from the company in advance, the AI ​​automatically reads these documents and directly reflects them in the usage fee calculation specification database. AI-powered natural language processing technology significantly streamlines the process of analyzing regulations, which was previously done manually, and improves reading accuracy. Specifically, reading accuracy improves, and interview time is reduced. The usage fee calculation specification database systematically stores detailed calculation rules such as the definition of the maximum rent, the method of collecting usage fees, the daily calculation method, and the start and end dates for collection, enabling the application of consistent calculation logic in subsequent processing.

[0117] In the first stage, a pre-trained language model is utilized. Tasks such as document classification, named entity recognition, relation extraction, and summary generation of regulations documents are performed, and the conversion to structured data is automated.

[0118] The document analysis supports a variety of document formats, including PDF (Adobe PDF®, Image PDF®), Microsoft Word (docx, doc), and Excel (xlsx, xls, csv). It also enables text extraction from documents converted into images using OCR technology (Tesseract®, Cloud Vision API, Azure® Computer Vision). Document layout analysis identifies structural elements such as tables (bordered and unbordered), diagrams (flowcharts, organizational charts), and bullet points (numbered and symbolized), enabling context-aware information extraction. The automated extraction of regulation items combines various methods to achieve highly accurate information extraction. Key elements such as rent limits, contract periods, move-in conditions, and application procedures are automatically identified and stored as structured data in the specifications database. In quality control, quality checks are performed automatically. This automatically identifies items that require manual verification, enabling efficient quality assurance. The specifications database utilizes multiple hybrid configurations to achieve integrated management of structured and unstructured data. Furthermore, data version control (Git-like versioning) enables complete tracking of the specification change history.

[0119] Furthermore, to accommodate the unique characteristics of each company's regulations, the system's regulation pattern learning function enhances its ability to automatically understand company-specific expressions and conditions. This allows it to handle not only standard regulations but also complex, company-specific conditions.

[0120] Figure 17 shows the details of the AI ​​machine learning process in the second stage, where an advanced machine learning system is in operation. Based on the completed specifications, the implementation team traditionally manually built the logic and conducted validation tests based on foreseeable scenarios. During this process, the calculation tool was put into full operation after verification by the implementation manager and the group's representatives, confirming the consistency of the calculation logic with the specifications, and finally receiving confirmation from the company's representative. In the conventional process, errors occurred due to human oversight. In particular, errors due to human oversight were frequent in the processing of complex conditional branching.

[0121] After implementing AI, it becomes possible to build accurate logic by consulting with the AI ​​in a chat format during the logic construction process. The AI ​​refers to accumulated past cases and best practices to propose the optimal calculation logic. Furthermore, since validation tests are automatically performed by AI to ensure no assumptions are missed, human errors are reduced and quality is improved. This automation streamlines the creation of calculation logic that aligns with specifications, drastically reducing the time required from several weeks to just a few days. Furthermore, continuous automated processing ensures that logic construction continues even late at night and on holidays.

[0122] The second stage of machine learning processing employs a multifaceted approach. Existing computational logic examples are used as training data to automatically learn computational patterns. Data cleaning, data augmentation, and annotation are performed to ensure the quality of the training data. The deep learning architecture utilizes CNNs (Convolutional Neural Networks: ResNet-50, EfficientNet) for document structure feature extraction, RNNs (Recurrent Neural Networks: LSTM, GRU) for sequence pattern learning, and Transformers (BERT, GPT®) for automatic identification of important elements using attention mechanisms. In particular, it has learned the conditional branching patterns specific to company housing regulations with high accuracy. The automated machine learning system automatically performs feature selection, model selection, and hyperparameter optimization. Model performance is evaluated using a comprehensive assessment based on multiple metrics. To ensure Explainable AI (XAI), we visualize the basis for prediction results using various methods. This creates a system that allows humans to understand and verify the AI's decision-making process. The Continual Learning feature enables the model to be continuously updated using new data. Various methods are used to acquire new knowledge while retaining past learning. Furthermore, through fine-tuning based on the applicant's extensive operational experience, the system achieves highly accurate responses that are difficult to achieve with typical AI systems. This experience enables the automatic determination of complex conditions and the handling of unfamiliar questions.

[0123] Figure 18 shows the details of the evaluation and analysis process in the third stage. In this stage, the analysis capabilities of the advice unit 135 are significantly expanded.

[0124] In the conventional workflow, irregular requests required the person in charge to manually check the specifications and perform calculations and checks manually. Traditionally, this type of retrospective calculation process has been time-consuming on average and carried a high risk of calculation errors. Furthermore, retrospective calculations require the recalculation of multiple related calculation items, necessitating an accurate understanding of the dependencies between each item and the determination of the processing order. Traditional manual processing methods were prone to secondary errors due to overlooking these dependencies.

[0125] After implementing AI, users can receive questions and answers from the chat-based AI, enabling quick and accurate calculations to be reflected in the data. During retrospective calculations, the AI ​​accesses past calculation results, identifies changes, and performs automatic calculations. The AI ​​refers to historical data stored in the usage fee calculation file server and automatically performs calculations such as calculating the difference before and after changes, determining the retrospective period, and recalculating all related calculation items. Processing time has been reduced compared to before, and calculation accuracy has also been greatly improved. This feature significantly streamlines complex retrospective calculations and recalculations due to condition changes, resulting in reduced human error and shorter processing times. The third stage of evaluation and analysis integrates advanced analytical methods. The validity of calculation results is automatically verified through statistical quality control based on past calculation history data. Change impact analysis automatically identifies the impact of one change on other calculation items. The prescribed algorithm enables the complete identification and prioritization of items that require recalculation. The data integrity check automatically detects integrity violations and suggests corrections. It also ensures complete traceability of calculation results. Real-time analysis enables real-time analysis of large amounts of data updates. Furthermore, it enables the detection and automatic response to complex event patterns. Predictive analytics uses various analytical methods to predict future calculation patterns and outliers. Proactive alerting based on these predictions enables early detection and countermeasures against problems.

[0126] Furthermore, the learning function for handling irregular situations accumulates and learns from past response cases, continuously improving the accuracy of automated responses when similar irregular cases occur. As a result, cases that required human judgment the first time can be handled automatically from the second time onward.

[0127] Figure 19 shows the details of the result output and visualization process in the fourth stage. In this stage, a multimodal output system, which is a significantly advanced version of the display control unit 137, is in operation.

[0128] Previously, only monthly usage fee calculation data was delivered, and the service company's personnel had to handle the process themselves. In particular, when dealing with inquiries from residents, it was necessary to explain the technical terms and calculation logic of the regulations in a way that was easy for general employees to understand, and the inconsistency in the quality of service due to individual differences in explanation skills was also a challenge.

[0129] After implementing AI, the company housing manager can now check the calculation of usage fees not only monthly, but also whenever they need to. Furthermore, inquiries from residents regarding company housing usage fees can also be handled by the AI. This automates many inquiry-handling tasks, allowing the manager to focus on higher-value tasks. Specifically, when a resident asks a question to the AI ​​in a chat format, the AI ​​gathers the necessary information step by step and then immediately provides a detailed calculation result. The fourth stage of visualization processing enables data output in various formats through integration with business intelligence (BI) tools. Native integration with BI tools such as Tableau® (Server / Desktop), Microsoft® Power BI®, Qlik Sense®, and Looker (Google® Cloud) allows for the creation of flexible dashboards tailored to the needs of businesses. Data connectivity supports standard interfaces. The conversational AI system integrates natural language understanding and natural language generation to achieve natural dialogue with humans. Through dialogue state management, it can handle complex inquiries spanning multiple turns. We can respond appropriately to inquiries in stages. Multimodal output enables the delivery of information in various formats, including text, audio, images, and video. The personalization feature provides individually optimized answers based on the user's attributes, such as past inquiry history, job title, department, and usage patterns. In terms of accessibility, the system implements features such as text-to-speech for the visually impaired, color-adjustable color schemes for the colorblind, text enlargement for the hearing impaired, and voice input for people with physical disabilities. This achieves universal design.

[0130] Furthermore, by analyzing the history of employee inquiries, it is possible to quantitatively understand which parts of the company housing regulations are difficult to understand and when inquiries are concentrated, which can then be used as reference information when the advisory department 135 provides advice on revising the regulations.

[0131] Figure 20 shows the structured process of initial interviews and the integrated configuration of an automated data acquisition system using artificial intelligence. Figure 20 shows a further automated version of the corporate regulations acquisition unit 131.

[0132] During the initial implementation, a comprehensive interview will be conducted regarding the definition of the rent cap, the detailed calculation method for collecting usage fees, and the start and end dates for collection. Traditionally, interviews by experienced personnel were required, and rework due to missed information during the interview process was a common occurrence in projects. Specifically, we will check which items are included in the definition of the maximum rent, and for items that are not included in the definition of rent, such as "water charges (fixed rate)," "utilities (fixed rate)," "neighborhood association fees," "fire insurance premiums," and "furniture and appliance rental fees," we will set one of the following for each: "company-funded," "individual burden / company reimbursement," or "individual burden / individual payment." Furthermore, a detailed review of how any amount exceeding the upper limit will be handled is also necessary. Regarding the collection of usage fees, detailed operational rules will be established, including the monthly calculation method at the time of move-in and move-out, the method for calculating fractional amounts of usage fees, and the start and end dates for collection.

[0133] In this embodiment, the AI ​​systematically manages these interview items and automatically generates a step-by-step question flow (decision tree format), thereby preventing missed interview points and achieving efficient information gathering. Furthermore, a function is implemented to automatically suggest customized question sets tailored to the company's characteristics, based on past interview results with similar companies.

[0134] The AI-powered automated data acquisition system enables the automatic extraction of data from a company's existing systems. Specifically, it significantly reduces manual data entry by automatically acquiring data from systems such as human resources systems, accounting systems, and real estate management systems.

[0135] The AI ​​chatbot function provides an interactive interface, as shown in the right side of Figure 20, that allows employees to intuitively check usage fees. It also provides detailed calculation results. Standardizing the initial consultation items ensures consistent implementation quality and efficiency. These consultation items include the usage fee calculation period, payroll deduction timing, definition of rent cap, burden allocation for various expense items, monthly calculation method, fractional calculation method, and collection start date. These standardized interview items ensure consistent quality in implementation work, regardless of the skill level of the implementation staff, resulting in improved customer satisfaction and shorter implementation periods. Furthermore, the digitization and database creation of interview results streamlines the process of sharing information with similar companies and facilitating the search and reference of past case studies.

[0136] Furthermore, in this embodiment, a function can be added to automatically respond to employee inquiries using AI. Specifically, the server 10 receives inquiries from employees and further includes a response unit that automatically responds using artificial intelligence based on the employee's attribute information and the prescribed regulations to be evaluated, according to the conditions stipulated in the prescribed regulations.

[0137] This response unit automates inquiries from employees regarding regulations, which were previously handled individually by company representatives, resulting in a significant improvement in operational efficiency. The response unit uses artificial intelligence to calculate and respond with the maximum rent and contract terms stipulated in the prescribed regulations, based on attribute information including at least one of the employee's job title, family structure, and place of work.

[0138] Furthermore, in embodiments of the present invention, a configuration is provided in which the evaluation process by the prescribed evaluation means performs the first to fourth stages of processing. As the first stage of processing, we will obtain the prescribed regulations for multiple targets and extract multiple items. As the second stage of processing, the acquired data according to the prescribed regulations is normalized and quantified. As the third stage of processing, we perform feature extraction and calculation of standard scores using machine learning. The fourth stage of processing involves visualizing and outputting the evaluation results.

[0139] Furthermore, this embodiment provides a configuration in which the evaluation process by the prescribed evaluation means (for example, the prescribed evaluation unit 134 in Figure 4) performs the processing of the first to fourth stages. The first stage of processing involves obtaining prescribed regulations for multiple targets (e.g., company housing regulations) and extracting multiple items (e.g., corresponding to Figure 16), the second stage involves normalizing and quantifying the data of the obtained prescribed regulations (e.g., corresponding to Figure 17), the third stage involves extracting features and calculating standard scores using machine learning (e.g., corresponding to Figure 18), and the fourth stage involves visualizing and outputting the evaluation results (e.g., corresponding to Figure 19).

[0140] These four stages of processing systematically execute the automated processes from the first to the fourth stage, as shown in Figures 16 to 19, achieving significant improvements in efficiency and accuracy compared to conventional manual processing. The processing content at each stage is interconnected, and the results of the processing in the previous stage are used as input for the next stage, creating an integrated system configuration.

[0141] Furthermore, in embodiments of the present invention, a configuration is also provided that further includes an automatic acquisition means (for example, the AI ​​automatic reading function shown in Figure 16, the AI ​​automatic processing system shown in Figure 20) that automates the acquisition of predetermined regulations (for example, company housing regulations) from multiple targets using artificial intelligence. This automated data acquisition method allows for the automatic extraction of regulatory data from a company's existing systems, acquisition of market information from external data sources, and digitization and structuring of regulatory documents, significantly streamlining traditional manual data collection processes.

[0142] In embodiments of the present invention, a general-purpose regulations evaluation system is provided that includes at least one of the following as predetermined regulations: company housing regulations, information security regulations, work rules, overseas assignment regulations, and business trip expense regulations (for example, the general-purpose system shown in Figures 14 and 15). This allows companies to manage and evaluate diverse regulations in an integrated manner within a single system, enabling them to simultaneously achieve cost reductions and improved operational efficiency.

[0143] Furthermore, multiple items are hierarchically structured into major items, medium items, and minor items (for example, the evaluation item table shown in Figure 5), and the evaluation target level calculation means (for example, the evaluation target level calculation unit 133 in Figure 4) provides a configuration that evaluates the evaluation results of minor items with binary values, calculates the evaluation value of a medium item by summing the evaluation values ​​of minor items belonging to the same medium item, and calculates the evaluation value of a major item by summing the evaluation values ​​of medium items belonging to the same major item. This hierarchical evaluation allows for multi-level assessments, from detailed to general, enabling companies to comprehensively understand their own regulations from various perspectives.

[0144] In embodiments of the present invention, a configuration is also provided that further includes a regional comparison display means (for example, the map display unit 240 in Figure 13) that compares the monetary conditions stipulated in a predetermined regulation to be evaluated (for example, company housing regulations) with the actual market price in each region using color coding on a map. This feature allows companies to visually understand how competitive their regulations are compared to the local market, and can be used to develop regional strategies.

[0145] Furthermore, a configuration is also provided that includes a means for comparing multiple targets by industry based on industry classification codes (e.g., Japan Standard Industrial Classification) and performing comparative evaluations with other companies in the same industry. This allows companies to conduct precise comparative analyses with competitors that have similar business models, enabling them to efficiently secure a competitive advantage and identify areas for improvement.

[0146] Embodiments of the present invention also provide a configuration built on a cloud computing environment (e.g., AWS® 200 in Figure 9) that includes an artificial intelligence system (e.g., AI system 201 in Figure 9), a data warehouse (e.g., data warehouse 202 in Figure 9), and a dashboard function (e.g., Tableau® DB 203 in Figure 9). This cloud-based architecture enables the delivery of high-quality services to numerous businesses while ensuring scalability, availability, and security.

[0147] Furthermore, the system includes external data acquisition means (for example, the function for linking with online web data shown in Figure 9) that automatically acquires open data through API integration or web scraping, and the evaluation standard level setting means (for example, the evaluation standard level setting unit 132 in Figure 4) is also provided in a configuration that calculates the standard level using external data. This feature allows for the setting of dynamic evaluation criteria that always reflect the latest market data and statistics, enabling the continuous provision of highly accurate evaluations that adapt to changing times. Furthermore, a configuration is also provided that includes a visualization means (for example, the flowchart display unit 220 in Figure 11) that visualizes multiple items in a flowchart format including conditional branching and statistically displays the selection ratio of multiple targets at each branch. This makes it possible to display the complex condition settings of regulations in an intuitively easy-to-understand format and support corporate decision-making. The system further includes a display control means (e.g., display control unit 137 in Figure 4) that executes control to display the results of the evaluation by the regulations evaluation means (e.g., regulations evaluation unit 134 in Figure 4) on an integrated dashboard (e.g., integrated dashboard 230 in Figure 12) that includes at least one of a radar chart, pie chart, heat map, and map display. This configuration provides an environment in which companies can analyze their regulations from multiple perspectives through various visualization methods. By further incorporating access control measures that allow setting access permissions for each company via a web application or dedicated system and controlling the viewing of evaluation results, it is possible to ensure the proper management and security of confidential corporate information. The learning mechanism (e.g., the learning unit 136 in Figure 4) continuously updates the machine learning model (e.g., the learning model 152 in Figure 4) using new evaluation performance data, thereby continuously improving evaluation accuracy and the ability to adapt to new regulation patterns. This continuous learning function enables adaptation to changing times, new legal systems, and the diversifying regulation patterns of companies, allowing for the continued provision of highly accurate evaluation services over a long period. As described above, the present invention makes it possible to visualize and quantitatively evaluate the legal system for company housing (company housing regulations, company housing systems, etc.) within a company. Furthermore, advanced functions such as AI-powered automated response, multi-stage processing, general regulation support, hierarchical evaluation, regional and industry-specific comparison, cloud infrastructure, external data linkage, and continuous learning enable the provision of a comprehensive and precise regulation evaluation system that was previously difficult to achieve.

[0148] Although embodiments of the present invention have been described above, the present invention is not limited to the embodiments described above, and any modifications, improvements, etc. that can achieve the objectives of the present invention are included in the present invention.

[0149] For example, the cloud system configuration shown in Figure 9 is merely an example for achieving the objectives of the present invention and is not particularly limited. It is also possible to use cloud platforms other than AWS® (such as Microsoft Azure®, Google® Cloud Platform, IBM® Cloud, etc.).

[0150] Furthermore, the functional configuration diagrams shown in Figures 10 to 20 are merely illustrative and not particularly limiting. In other words, it is sufficient for the information processing device to be equipped with the functionality to execute the series of processes described above as a whole, and the functional blocks and databases used to realize this functionality are not particularly limited to the examples shown in the figures.

[0151] Furthermore, the location of functional blocks and databases is not limited to the diagram and can be arbitrary. Furthermore, a single functional block and database may be configured as a standalone hardware unit, on separate hardware, as standalone software, or as a combination of these.

[0152] When the processing of each functional block is to be executed by software, the programs that make up that software are installed on a computer or other device from a network or storage medium. A computer may be a computer built into dedicated hardware. Alternatively, a computer may be a computer capable of performing various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.

[0153] Recording media containing such programs consist not only of removable media distributed separately from the main device to provide the program to the user, but also of recording media provided to the user in a state where they are pre-installed in the main device.

[0154] In this specification, the step of describing a program to be recorded on a recording medium includes not only processes that are performed chronologically in that order, but also processes that are not necessarily performed chronologically, but are executed in parallel or individually.

[0155] Furthermore, in this specification, the term "system" refers to an overall system composed of multiple devices, means, etc.

[0156] In the above embodiment, company housing regulations were used as an example, but company housing is part of a company's employee benefits and the standards vary from company to company. However, this method can be applied not only to company housing but also to any other regulations such as "overseas assignment regulations," "business trip expense regulations," "personnel and salary regulations," "work rules," "information security regulations," and "telecommuting regulations." Furthermore, in the embodiment described above, for example, five items were given as an example of evaluation items, such as sub-items, but this is merely an example. The names and contents of the major, medium, and minor items are also merely examples.

[0157] In summary, the information processing device to which the present invention applies only needs to have the following configuration, and various embodiments can be adopted. In other words, the information processing device to which the present invention is applied (for example, the server 10 in Figure 3, the cloud system in Figure 9, etc.) Based on the respective prescribed regulations (e.g., company housing regulations) of multiple entities (e.g., companies A through D), a means for setting evaluation criteria levels (e.g., evaluation criteria level setting unit 132 in Figure 4) sets multiple items to be evaluated and calculates a standard level (e.g., a standard value within a certain area) for each of those multiple items. From the aforementioned targets (for example, companies A through D), the evaluation target level calculation means (for example, the evaluation target level calculation unit 133 in Figure 4) extracts the aforementioned multiple items from the aforementioned prescribed regulations (for example, company housing regulations) of the evaluation target (for example, company A) and calculates the item level for each of the aforementioned multiple items, The system includes a regulations evaluation means (for example, the regulations evaluation unit 134 in Figure 4) that compares the item level of each of the multiple items of the subject to evaluation (for example, Company A) with the standard level of each of the multiple items, and evaluates the prescribed regulations (for example, company housing regulations) of the subject to evaluation (for example, Company A) based on the results of the comparison.

[0158] In this way, by setting multiple items to be evaluated based on the respective prescribed regulations (e.g., company housing regulations) of multiple entities (e.g., companies A through D), calculating a standard level for each of these items, extracting multiple items from the prescribed regulations (e.g., company housing regulations) of the entity being evaluated (e.g., company A), calculating an item level for each of these items, comparing the item level of each of the items in the entity being evaluated with the standard level of each of those items, and evaluating the prescribed regulations (e.g., company housing regulations) of the entity being evaluated (e.g., company A) based on the results of this comparison, the company housing regulations and company housing system in the entity being evaluated (e.g., company A) can be made visible and evaluated quantitatively.

[0159] The aforementioned regulation evaluation means (for example, the regulation evaluation unit 134 in Figure 4) can perform the evaluation of the aforementioned prescribed regulations (for example, company housing regulations) by calculating a deviation score from the item level of each of the items subject to evaluation, based on the standard level of each of the items subject to evaluation, and relating the prescribed regulations of the subject to evaluation (for example, company A).

[0160] The aforementioned regulation evaluation means (for example, the regulation evaluation unit 134 in Figure 4) can perform the aforementioned evaluation of the aforementioned regulation (for example, the company housing regulations) by selecting one or more items that are characteristic of the aforementioned regulation (for example, the company housing regulations) of the subject of evaluation (for example, Company A) from among the multiple items as characteristic items, and by calculating characteristic quantities based on each of the item levels of the one or more such characteristic items.

[0161] The system further includes a learning means (for example, the learning unit 136 and the learning model 152 in Figure 4) that performs predetermined machine learning using the aforementioned standard scores and features for multiple targets and outputs the learning results. This allows the results of machine learning to be used to select targets (for example, companies) that are more likely to revise their company housing regulations.

[0162] The system further includes an advice means (for example, the advice unit 135 in Figure 4) that generates advice information (for example, a message encouraging revision of the company housing regulations) that provides advice on the predetermined regulations (for example, company housing regulations) of the evaluation target (for example, company A) based on the learning results, and at least a portion of the standard score and features of the evaluation target (for example, company A). This allows the system to encourage companies to revise their company housing regulations to make them better.

[0163] The aforementioned subject is a company, The aforementioned prescribed provisions are the company's regulations concerning employee housing (for example, employee housing regulations).

[0164] By further incorporating a response mechanism (an independent AI automated response system linked to a data warehouse) that receives inquiries from employees and automatically responds using artificial intelligence based on the employee's attribute information and the aforementioned prescribed regulations subject to evaluation, it becomes possible to automate employee inquiries regarding regulations that were previously handled individually by company personnel, thereby achieving a significant improvement in operational efficiency.

[0165] The response means calculates the maximum rent and contract terms stipulated in the prescribed regulations based on attribute information including at least one of the employee's job title, family structure, and place of work, using artificial intelligence, and responds accordingly. This allows employees to immediately check the contents of the company housing regulations that suit their own conditions, and also contributes to reducing the workload of company personnel.

[0166] The aforementioned regulation evaluation means (for example, the regulation evaluation unit 134 in Figure 4) executes the first to fourth stages of processing as part of the evaluation process. As the first stage of processing, the predetermined regulations for the multiple targets and the multiple items are extracted (for example, the first stage in Figure 16), As the second stage of processing, the acquired data according to the predetermined rules is normalized and digitized (for example, the second stage in Figure 17), As the third stage of processing, the feature quantities are extracted and the standard scores are calculated using machine learning (for example, the third stage in Figure 18), As the fourth stage of processing, the results of the evaluation are visualized and output (for example, the fourth stage in Figure 19), thereby achieving significant efficiency improvements and accuracy enhancements compared to conventional manual processing through a systematic processing process.

[0167] By further incorporating an automated acquisition means (for example, the AI ​​automatic reading function shown in Figure 16, or the RPA system in Figure 20) that automates the acquisition of predetermined data from the aforementioned multiple targets using artificial intelligence, the conventional manual data collection work can be made significantly more efficient.

[0168] The aforementioned prescribed regulations include at least one of the following: company housing regulations, information security regulations, work rules, overseas assignment regulations, and business trip expense regulations (for example, the general-purpose correspondence shown in Figures 14 and 15). This enables companies to comprehensively manage and evaluate diverse regulations within a single system.

[0169] The aforementioned items are hierarchically organized into major items, medium items, and minor items (for example, the evaluation item table shown in Figure 5), The evaluation target level calculation means (for example, the evaluation target level calculation unit 133 in Figure 4) evaluates the evaluation results of the sub-items as binary values, calculates the evaluation value of the sub-item by summing the evaluation values ​​of the sub-items belonging to the same medium-item, and calculates the evaluation value of the medium-item by summing the evaluation values ​​of the medium-items belonging to the same major-item, thereby enabling multi-stage evaluation from the detailed level to the general level.

[0170] By further providing a regional comparison display means (for example, the map display unit 240 in Figure 13) that compares the monetary conditions stipulated in the aforementioned prescribed regulations, which are subject to evaluation, with the actual market price in each region, using color coding on a map, companies can visually grasp how competitive their regulations are compared to the regional market.

[0171] By further incorporating industry-specific comparison tools that classify the aforementioned multiple targets by industry based on industry classification codes and perform comparative evaluations with other companies in the same industry, companies can conduct precise comparative analyses with other companies in the same industry that have similar business models.

[0172] Built on a cloud computing environment (e.g., AWS® 200 in Figure 9), and including an artificial intelligence system (e.g., AI system 201 in Figure 9), a data warehouse (e.g., data warehouse 202 in Figure 9), and dashboard functionality (e.g., Tableau® DB 203 in Figure 9), it is possible to provide high-quality services to numerous companies while ensuring scalability, availability, and security.

[0173] Furthermore, it includes external data acquisition means (for example, the function to link with online web data shown in Figure 9) that automatically acquires open data through API integration or web scraping. The evaluation standard level setting means (for example, the evaluation standard level setting unit 132 in Figure 4) calculates the standard level using the external data, thereby enabling the setting of dynamic evaluation standards that always reflect the latest market data and statistical information.

[0174] By further providing a visualization means (for example, the flowchart display unit 220 in Figure 11) that visualizes the aforementioned multiple items in a flowchart format including conditional branching and statistically displays the selection rate of the aforementioned multiple items at each branch, the complex condition settings of the regulations can be displayed in an intuitively easy-to-understand format, thereby supporting corporate decision-making.

[0175] The system further includes a display control means (e.g., display control unit 137 in Figure 4) that executes control to display the results of the evaluation by the aforementioned regulation evaluation means (e.g., regulation evaluation unit 134 in Figure 4) on an integrated dashboard (e.g., integrated dashboard 230 in Figure 12) that includes at least one of a radar chart, pie chart, heat map, and map display, thereby providing an environment in which companies can analyze their regulations from multiple perspectives through various visualization methods.

[0176] By further providing access control means that allow access rights to be set for each company via a web application or dedicated system and control the viewing of the evaluation results, appropriate management and security of corporate confidential information can be ensured.

[0177] The learning means (for example, the learning unit 136 in Figure 4) can continuously improve evaluation accuracy and its ability to handle new standard patterns by continuously updating the machine learning model (for example, the learning model 152 in Figure 4) using new evaluation performance data.

[0178] The information processing method to which the present invention is applied is: In an information processing method executed by an information processing device, Based on the prescribed regulations for each of the multiple targets (for example, company housing regulations), multiple items to be evaluated are set, and for each of these items, a standard level is calculated in the evaluation standard level setting step (for example, step S101 in Figure 6), From the aforementioned subject, the evaluation target level calculation step (for example, step S102 in Figure 6) involves extracting the aforementioned multiple items from the aforementioned prescribed regulations (for example, company housing regulations) of the evaluation target (for example, company A), and calculating the item level for each of the aforementioned multiple items. The system may include a regulations evaluation step (e.g., step S105 in Figure 6) which compares the item level of each of the multiple items of the subject to evaluation (e.g., Company A) with the standard level of each of the multiple items, and evaluates the prescribed regulations (e.g., company housing regulations) of the subject to evaluation (e.g., Company A) based on the results of the comparison.

[0179] The program to which this invention applies is: On the computer, Based on the prescribed regulations for each of the multiple targets (for example, company housing regulations), multiple items to be evaluated are set, and for each of these items, a standard level is calculated in the evaluation standard level setting step (for example, step S101 in Figure 6), From the aforementioned subject, the evaluation target level calculation step (for example, step S102 in Figure 6) involves extracting the aforementioned multiple items from the aforementioned prescribed regulations (for example, company housing regulations) of the evaluation target (for example, company A), and calculating the item level for each of the aforementioned multiple items. The control process can be made to execute a regulation evaluation step (for example, step S105 in Figure 6) which includes comparing the item level of each of the multiple items of the subject to evaluation (for example, company A) with the standard level of each of the multiple items, and evaluating the prescribed regulations (for example, company housing regulations) of the subject to evaluation (for example, company A) based on the results of the comparison. [Explanation of Symbols]

[0180] 10...Server, 20, 20-1, 20-u...Enterprise terminal, 30...Service provider terminal, 101...CPU, 102...ROM, 103...RAM, 104...Bus, 105...Input / Output interface, 106...Output unit, 107...Input unit, 108...Storage unit, 109...Communication unit, 110...Drive, 111...Removable media, 131...Enterprise regulations acquisition unit, 132...Evaluation standard level setting unit, 133...Evaluation target level calculation unit, 134...Regulation evaluation unit, 135...Advice unit, 136...Learning unit, 137...Display control unit, 151...Analysis platform DB, 152...Learning model, 200...AWS (Registered Tableau® (Registered Trademark), 201...AI System, 202...Data Warehouse, 203...Tableau® DB, 204...Web Application, 205...Rilonet, 210...Statistical Analysis Department, 211...Percentage Display Department, 220...Flowchart Display Department, 221...Statistical Value Display Department, 222...Comparison Comment Department, 230...Integrated Dashboard, 231...Radar Chart Department, 232...Evaluation Details Department, 240...Map Display Department, 241...Comparison Table Department, 250...Evaluation Target Process Flow Department, 251...Structured Processing Department, 252...Analysis Method Department, 260...Other Regulations Evaluation Department, 261...Regulation-Specific Display Department, NW...Network

Claims

1. Based on the prescribed regulations for each of the multiple targets, a means for setting evaluation criteria levels is used to define multiple items to be evaluated and to calculate a standard level for each of those items. A means for calculating the evaluation target level, which extracts the multiple items from the predetermined regulations of the target to be evaluated among the aforementioned targets, and calculates the item level for each of the multiple items, A regulation evaluation means that compares the item level of each of the multiple items to be evaluated with the standard level of each of the multiple items, and evaluates the predetermined regulation to be evaluated based on the result of the comparison, An information processing device equipped with the following features.

2. The aforementioned regulation evaluation means performs the evaluation of the aforementioned prescribed regulation by calculating a deviation score for the aforementioned prescribed regulation relative to the aforementioned prescribed regulation, based on the aforementioned standard level for each of the aforementioned items, and from the respective item levels of the aforementioned items that are to be evaluated. The information processing apparatus according to claim 1.

3. The aforementioned regulation evaluation means, based on the results of the comparison, selects one or more items from the plurality of items that are characteristic of the prescribed regulation being evaluated as characteristic items, and calculates feature quantities based on the item level of each of the one or more characteristic items, thereby performing the evaluation of the prescribed regulation. The information processing apparatus according to claim 2.

4. A learning means that performs predetermined machine learning using the aforementioned standard scores and features for each of the multiple targets and outputs the learning results. The information processing apparatus according to claim 3, further comprising:

5. An advice means that generates advice information indicating advice for the predetermined regulations of the subject to evaluation, based on the standard score and at least a portion of the features of the subject to evaluation, and the learning results. The information processing apparatus according to claim 4, further comprising:

6. The aforementioned subject is a company, The aforementioned prescribed regulations are the regulations concerning company housing. The information processing apparatus according to any one of claims 1 to 5.

7. A response means that receives inquiries from employees and, based on the attribute information of the employee and the aforementioned prescribed regulations subject to evaluation, automatically responds using artificial intelligence according to the conditions stipulated in the aforementioned prescribed regulations. The information processing apparatus according to claim 6, further comprising:

8. The response means calculates and responds to the employee's attribute information, which includes at least one of the employee's job title, family structure, and place of work, using artificial intelligence to determine the maximum rent and contract terms stipulated in the prescribed regulations. The information processing apparatus according to claim 7.

9. The aforementioned regulation evaluation means executes the first to fourth stages of processing as part of the evaluation process. As the first stage of processing, the predetermined regulations for the multiple targets and the multiple items are extracted. As the second stage of processing described above, the acquired data according to the predetermined regulations is normalized and digitized. As the third stage of processing, the feature quantities are extracted and the standard scores are calculated using machine learning. As the fourth stage of processing, the results of the evaluation are visualized and output. The information processing apparatus according to claim 3.

10. The system further includes an automated acquisition means that uses artificial intelligence to automate the acquisition of predetermined data from the multiple targets. The information processing apparatus according to claim 1.

11. The aforementioned prescribed regulations include at least one of the following: company housing regulations, information security regulations, work rules, overseas assignment regulations, and business trip expense regulations. The information processing apparatus according to claim 1.

12. The aforementioned items are hierarchically organized into major items, medium items, and minor items. The evaluation target level calculation means evaluates the evaluation result of the sub-item as a binary value, calculates the evaluation value of the sub-item by summing the evaluation values ​​of the sub-items belonging to the same medium-item, and calculates the evaluation value of the medium-item by summing the evaluation values ​​of the medium-items belonging to the same major-item. The information processing apparatus according to claim 1.

13. The system further includes a regional comparison display means for comparing the monetary conditions specified in the aforementioned prescribed regulations for the subject of evaluation with the actual market price in each region, using color coding on a map. The information processing apparatus according to claim 1.

14. The system further includes a means for classifying the aforementioned multiple targets by industry based on industry classification codes and for performing comparative evaluations with other companies in the same industry. The information processing apparatus according to claim 1.

15. Built on a cloud computing environment, it includes artificial intelligence systems, data warehouses, and dashboard functions. The information processing apparatus according to claim 1.

16. It further includes external data acquisition means that automatically acquire open data through API integration or web scraping. The evaluation standard level setting means calculates the standard level using the external data. The information processing apparatus according to claim 1.

17. The system further includes visualization means for visualizing the aforementioned multiple items in a flowchart format including conditional branching, and for statistically displaying the selection ratio of the aforementioned multiple items at each branch. The information processing apparatus according to claim 1.

18. The system further includes a display control means that performs control to display an integrated dashboard in which the results of the evaluation by the prescribed evaluation means are shown, in a format including at least one of a radar chart, a pie chart, a heat map, and a map display. The information processing apparatus according to claim 1.

19. The system further includes access control means that allows setting access rights for each company via a web application or dedicated system and controls the viewing of the evaluation results. The information processing apparatus according to claim 1.

20. The learning means continuously updates the machine learning model using new evaluation performance data. The information processing apparatus according to claim 4.

21. In an information processing method executed by an information processing device, Based on the prescribed regulations for each of the multiple targets, multiple items to be evaluated are set, and a standard level is calculated for each of those items in the evaluation standard level setting step. A step to calculate the evaluation target level, in which the multiple items are extracted from the predetermined regulations of the target to be evaluated among the aforementioned targets, and the item level is calculated for each of the multiple items, A rule evaluation step which involves comparing the item level of each of the multiple items to be evaluated with the standard level of each of the multiple items, and evaluating the prescribed rule to be evaluated based on the result of the comparison, Information processing methods including

22. On the computer, Based on the prescribed regulations for each of the multiple targets, multiple items to be evaluated are set, and a standard level is calculated for each of those items in the evaluation standard level setting step. A step to calculate the evaluation target level, in which the multiple items are extracted from the predetermined regulations of the target to be evaluated among the aforementioned targets, and the item level is calculated for each of the multiple items, A rule evaluation step which involves comparing the item level of each of the multiple items to be evaluated with the standard level of each of the multiple items, and evaluating the prescribed rule to be evaluated based on the result of the comparison, A program that executes control processes, including those mentioned above.

Images