AI implementation effect simulation system, AI implementation effect simulation program, and AI implementation effect simulation method

Abstract

We will simulate the effects of AI implementation and improve convenience for those considering AI adoption. [Solution] The system includes multiple job name selection units 32a that accept job type selections, a number of employees input unit 32b, a desk work ratio input unit 32c, an hourly wage input unit 32e that can be registered for each job name selection unit 32a, a calculation unit 22, and a simulation results screen 50 that outputs the simulation results, and an AI reduction rate is further applied to the desk work ratio. Furthermore, the simulation results screen 50 can display a monthly cost reduction amount display unit 52, an annual cost reduction amount display unit 53, an investment recovery period display unit 54, and a personnel reduction equivalent display unit 55. Moreover, it is possible to input the job name selection unit 32a, the number of employees input unit 32b, the desk work ratio input unit 32c, and the hourly wage input unit 32e with AI assistance.

Description

Technical Field

[0001] The present invention relates to an AI introduction effect simulation system, an AI introduction effect simulation program, and an AI introduction effect simulation method for quantitatively estimating the return on investment (ROI) of introducing artificial intelligence (AI) in an organization.

Background Art

[0002] In recent years, the utilization of artificial intelligence (AI) in organizations such as corporations and associations has started to progress. If it is possible to grasp whether an effect commensurate with the investment required for AI introduction can be obtained, this will be useful as a reference for AI introduction examiners such as corporate managers, association chairpersons, or persons with payment authority when making decisions on AI introduction.

[0003] However, there has been a problem that there are few means for presenting investment effects as objective numerical values at management meetings or for internal approval procedures. In addition, individual diagnosis by consulting firms has the problem of requiring significant cost and time.

[0004] As a system for performing reduction simulation, there has been proposed a system that generates a plurality of simulation results using AI based on input data and presents the results to a user (see Patent Document 1).

[0005] However, this system predicts the effect of energy efficiency improvement measures and other environmental reduction measures selected by a user and generates simulation results, and is not capable of simulating the effect of AI introduction in an organization.

Prior Art Literature

Patent Literature

[0006]

Patent Document 1

Summary of the Invention

Problem to be Solved by the Invention

[0007] In view of the above-mentioned problems, this invention aims to simulate the effects of AI implementation and improve convenience for those considering AI implementation. [Means for solving the problem]

[0008] This invention comprises a plurality of occupation type designation units that accept the designation of occupation types, a number of people designation unit that accepts the designation of the number of people, a target work ratio designation unit that accepts the designation of the target work ratio for which working time can be reduced by introducing AI, a calculation unit that performs simulation calculations based on the designated information, and a simulation result output unit that outputs the results of the simulation performed by the calculation unit. The number of people designation units and the target work ratio designation units are provided so that one each can be designated for one target work ratio designation unit. The calculation unit performs calculations for the occupation type unit using the number of people designated in the number of people designation unit and the target work ratio designated in the target work ratio designation unit as elements, integrates the calculated values ​​for all occupation types, and outputs the time or cost that can be reduced by introducing AI as a reduction effect to the simulation result output unit. The system includes a message display area where the AI ​​asks the user about the occupation type, number of people, and percentage of the target tasks; a text input field that accepts natural language input from the user as input content to be sent to the AI; and an application instruction unit that applies the response text generated by the AI ​​as a response to the input content to the occupation type specification unit, the number of people specification unit, and the target task percentage specification unit. The present invention is characterized by being an AI implementation effect simulation system, an AI implementation effect simulation program, and an AI implementation effect simulation method. [Effects of the Invention]

[0009] This invention provides an AI implementation effect simulation system, an AI implementation effect simulation program, and an AI implementation effect simulation method that simulate the effects of AI implementation and improve convenience for those considering AI implementation. [Brief explanation of the drawing]

[0010] [Figure 1] Block diagram showing the system configuration of an AI-implemented ROI simulation system. [Figure 2] Data structure diagram for the industry adjustment table. [Figure 3] Data structure diagram for the scale classification table. [Figure 4] A functional block diagram of the ROI simulation program that runs on the user's terminal. [Figure 5] A diagram illustrating the screen configuration of the simulation input screen displayed on the user's terminal. [Figure 6] A diagram illustrating the screen layout of the simulation results screen displayed on the user's terminal. [Figure 7] Screen layout diagram showing the external AI usage screen open. [Figure 8] Screen layout diagram showing the AI ​​usage screen on the HP website. [Modes for carrying out the invention]

[0011] Hereinafter, one embodiment of this invention will be described with reference to the drawings. In this embodiment, the type of business is referred to as the "industry," and the type of occupation is referred to as the "job." [Examples]

[0012] Figure 1 is a block diagram showing the system configuration of AI-implemented ROI simulation system 1. The AI-implemented ROI simulation system 1 consists of a management server 2 and multiple user terminals 8 connected to the internet 1a.

[0013] The management server 2 is composed of a web server computer and has the following hardware elements: a control unit 2a which consists of a CPU, ROM, RAM, etc., and performs various calculations and control operations; an input unit 2b which consists of a touch panel, keyboard, mouse, push buttons, or a combination thereof, and accepts input by touch operation; a display unit 2c which consists of a liquid crystal display or organic EL display, etc., and displays images such as characters and diagrams; a communication unit 2d which consists of a LAN board or WiFi unit, etc., and performs wired or wireless communication; and a storage unit 2e which consists of a hard disk or flash memory, etc., and allows reading and writing of information.

[0014] A program 2f and data 2g are stored in a storage unit 2e. Accordingly, a control unit 2a of the management server 2 operates in accordance with the program 2f, and operates by appropriately referencing and updating the data stored in the data 2g. The program 2f and the data 2g may be stored in various storage media, and may be installed or copied from the storage medium to the storage unit 2e of the management server 2.

[0015] The user terminal 8 is configured of an information communication terminal such as a personal computer, a tablet terminal, or a smartphone, and includes as hardware components: a control unit 8a configured of a CPU, a ROM, a RAM, and the like and executing various calculations and control operations; an input unit 8b configured of a touch panel, a keyboard, a mouse, a push button, or a combination of a plurality of the foregoing, and accepting an input by a touch operation; a display unit 8c configured of a liquid crystal display, an organic EL display, or the like, and displaying images such as characters and drawings; a communication unit 8d configured of a LAN board, a WiFi unit, or the like, and executing wired or wireless communication; a storage unit 8e configured of a hard disk, a flash memory, or the like, and allowing reading and writing of information; and an imaging unit 8h that captures an image using an imaging element such as a CCD or CMOS.

[0016] A program 8f such as a web browser or a mailer (e-mail transmission / reception application), and various types of data 8g are stored in the storage unit 8e. Accordingly, the control unit 8a of the user terminal 8 operates in accordance with the program 8f, and operates by appropriately referencing and updating the data stored in the data 8g. The control unit 8a of the user terminal 8 receives HTML from the management server 2 via the Internet 1a, and transmits input information from the input unit 8b input by a user to the management server 2. The program 8f and the data 8g may be stored in various storage media, and may be installed or copied from the storage medium to the storage unit 8e of the user terminal 8.

[0017] FIG. 2 is a data configuration diagram of an industry correction table 24a stored as a type of the data 2g in the storage unit 2e of the management server 2. The industry correction table 24a stores the number of people by occupation, the desk work ratio, and default hourly wages corresponding to six industries. When a user selects an industry preset, the occupation in each row of the occupation table input unit 21b is updated, and the number of people, desk work ratio, and hourly wage for each occupation row are automatically updated to the values in this table. In addition, portions indicated by "-" in the industry correction table 24a contain NULL values, indicating that this occupation does not exist in the default items for that industry. This enables arbitrary setting of both the number and types of configurable occupations for various industries.

[0018] Figure 3 is a data configuration diagram of the scale classification table 24b stored as a type of data 2g in the storage unit 2e of the management server 2. The scale classification table 24b stores default introduction cost values for four phases corresponding to three classifications: small-scale (30 or fewer people), medium-scale (31 to 100 people), and large-scale (101 or more people). The scale determination means 22a automatically determines the classification based on the total number of registered people in the occupation table input unit 21b, and sets the corresponding default value in the introduction cost setting unit 35, which serves as an AI introduction cost specification unit. However, if a user manually changes a cost, that manually changed value takes precedence.

[0019] Figure 4 is a functional block diagram of each functional unit realized when the browser of the user terminal 8 receives and expands an HTML file through the operation of an AI introduction ROI simulation program that outputs the HTML file as a type of program 2f for the management server 2. Figure 5 is a configuration diagram of the simulation input screen 30 displayed on the browser of the user terminal 8. Figure 6 is a configuration diagram of the simulation result screen 50 displayed on the browser of the user terminal 8.

[0020] As shown in Figure 4, the input unit 21 includes an industry selection unit 21a, an occupation table input unit 21b, and a cost input unit 21c.

[0021] The industry selection unit 21a displays several industry buttons 31a to 31f to accept the selection of an industry (business type). In this embodiment, six types of industry buttons 31a to 31f are displayed: construction / equipment work, manufacturing, retail / food service, service industry, professional services / consulting, and office-related. When a user presses any of the buttons, the control unit 2a refers to the industry correction table 24a and lists the occupations recorded as the standard occupation configuration for that industry in all rows of the occupation table input unit 21b. It then updates the number of people, desk work percentage (DW percentage), and hourly wage for each occupation to the values ​​registered in the industry correction table 24a. Note that even after selecting an industry, each parameter (number of people, desk work percentage, hourly wage) can be manually changed by the user, and the industry selection only functions as a batch setting of default values.

[0022] The job type table input section 21b is a table-format input interface that forms the core of the input section 21. It assigns one row to each job type (occupation type) and allows for the addition and deletion of multiple rows. Each row includes a job type name selection section 32a for selecting or freely entering a job type, a number of employees input section 32b for entering the number of people in that job type within the organization, a desk work percentage input section 32c for entering the desk work percentage of that job type as a percentage from 0 to 100, a fixed value AI reduction rate display section 32d, and an hourly wage input section 32e for entering the average hourly wage for that job type within the organization. An estimated annual income may also be displayed next to the average hourly wage. In this case, it can be calculated and displayed by appropriate calculations such as multiplying the average hourly wage by 12 months and then by 160 hours.

[0023] The cost input unit 21c is the part that receives input to each field of the implementation cost setting unit 35, and includes a conceptual diagnosis cost input unit 35a, a PoC cost input unit 35b, a main development cost input unit 35c, and a monthly operating cost input unit 35d.

[0024] The calculation unit 22 includes a scale determination means 22a, a reduction time calculation means 22b, a cost reduction calculation means 22c, a payback period calculation means 22d, and a conversion calculation means 22e.

[0025] The scale determination means 22a obtains the sum of the number of employees column in the job type table input unit 21b (i.e., the sum of the values ​​in the number of employees input unit 32b for each row) and automatically determines the scale category as follows: 30 employees or less is small-scale, 31 to 100 employees is medium-scale, and 101 employees or more is large-scale. Based on the determination result, it reads the corresponding default cost values ​​(concept diagnosis cost, PoC cost, actual development cost, monthly operating cost) from the scale category table 24b, sets them in the respective fields of the cost input unit 21c, and displays them. The scale determination means 22a re-evaluates each time the contents of the job type table input unit 21b are updated, and if the scale category changes, the default cost values ​​(concept diagnosis cost, PoC cost, actual development cost, monthly operating cost) are updated. However, if the user has manually changed the costs, those manually changed values ​​take precedence.

[0026] The time reduction calculation means 22b is activated when the simulation execution button 36 is pressed. It applies the following formula 1 to all rows of the job type table input unit 21b to calculate the monthly time reduction Hi for each job type, and stores the total value as ΣHi. A base value of 160 hours per person is used for monthly working hours. [Formula 1: Monthly time reduction for job type i (Hi)] Hi = Ci × 160 × (Di ÷ 100) × 0.35 *Note: Ci represents the number of employees in job category i, Di represents the percentage of desk work in job category i, and 0.35 represents the AI ​​reduction rate constant.

[0027] The cost reduction calculation means 22c uses the monthly reduction time Hi and hourly wage Wi for each job type calculated by the reduction time calculation means 22b to perform calculations according to the following formula 2 and calculate the monthly cost reduction amount S (in ten thousand yen). In addition, the annual cost reduction amount is calculated by multiplying the monthly cost reduction amount S by 12. [Formula 2: Monthly cost reduction amount S (10,000 yen)] S = Σ(Hi × Wi) ÷ 10,000 *Note: Wi represents the hourly wage (in yen) for job type i.

[0028] The payback period calculation means 22d obtains the sum of the concept diagnosis cost, PoC cost, and actual development cost set in the cost input unit 21c as the total initial cost F, and calculates the investment payback period P (months) by performing the calculation according to the following formula 3 using the monthly operating cost M and the monthly cost reduction amount S. If the difference between S and M is 0 or less, that is, if the monthly reduction amount does not exceed the monthly operating cost, the investment payback period P is displayed as "cannot be calculated". [Formula 3: Payback period P (months, rounded up)] P = F ÷ (SM) *Note: F represents the total initial cost (in 10,000 yen), and M represents the monthly operating fee (in 10,000 yen).

[0029] The conversion calculation means 22e receives the total monthly reduction time Hi value ΣHi calculated by the reduction time calculation means 22b, and calculates the equivalent value for new sales visits (ΣHi ÷ 3, unit: visits / month), the equivalent value for proposal / quote creation (ΣHi ÷ 5, unit: items / month), and the equivalent value for employee training / development (ΣHi ÷ 8, unit: days / month), respectively. In other words, the conversion calculation means 22e in this embodiment outputs the converted values ​​for new sales visits (1 visit = 3 hours), proposal / quote creation (1 item = 5 hours), and employee training / development (1 day = 8 hours), but it is not limited to these and can be set to any appropriate time conversion. This converts the effect of time reduction into an indicator of workload that managers can intuitively understand and provide. In addition, the equivalent number of personnel E (people) to be reduced is calculated using the following formula 4. [Formula 4: Equivalent number of personnel to be reduced (number of people)] E = ΣHi ÷ 160

[0030] The display unit 23 includes a result display means 23a, a graph display means 23b, and an approval document output means 23c.

[0031] The result display means 23a reflects each calculated value calculated by the calculation unit 22 in the respective display units of the simulation result screen 50 (monthly time reduction display unit 51, monthly cost reduction amount display unit 52, annual cost reduction amount display unit 53, investment recovery period display unit 54, personnel reduction equivalent display unit 55, implementation cost breakdown display unit 56, and time utilization conversion display unit 57). The JavaScript engine of the user terminal 8 displays the results by updating the text of each element through DOM manipulation, and in mobile environments (screen width less than 900px), smooth scrolling to the simulation result screen 50 is also performed.

[0032] The graph display means 23b visualizes the trend of monthly cumulative cost reductions over the calculated investment recovery period as a line graph or bar graph, intuitively indicating the timing of investment recovery. The graph is drawn on the client side using the browser's Canvas API or SVG.

[0033] The approval document output means 23c calls the browser's window.print() function when the print / output button 60 is pressed, and hides the industry preset selection section 31, simulation execution button 36, print / output button 60, inquiry CTA, and navigation elements by setting display:none using the @media print media query. It then formats the input values ​​from the simulation input screen 30 and all calculation results from the simulation results screen 50 into a single document and outputs it in approval document format. This allows individuals in organizations considering AI implementation to submit simulation results as an approval document or an attachment to the approval document when seeking approval from their superiors.

[0034] The memory unit 24 stores the industry correction table 24a, the size classification table 24b, and the default job values ​​24c. The default job values ​​24c store the number of employees, the percentage of desk work, and the standard hourly wage for each job used in the initial display of the job table input unit 21b. In this embodiment, the initial values ​​are set as follows: Field / Manual labor: 15 people, DW 10%, hourly wage 2,000 yen; Office / Accounting / General Affairs: 4 people, DW 80%, hourly wage 2,200 yen; Sales / External Relations: 3 people, DW 50%, hourly wage 2,500 yen; Management / Executive: 2 people, DW 60%, hourly wage 3,500 yen; Professional / Practical work: 1 person, DW 85%, hourly wage 3,500 yen; Others: None (to be added as needed).

[0035] As shown in Figure 5, the simulation input screen 30 includes an industry preset selection unit 31, a job type table input unit 32, an employee count display unit 33, a company-wide average hourly wage display unit 34, an implementation cost setting unit 35, and a simulation execution button 36.

[0036] The industry preset selection unit 31 is a selection unit in which six industry buttons 31a to 31f for construction / equipment work, manufacturing, retail / food service, service industry, professional services / consulting, and office-related industries are arranged horizontally. Each button is a clickable rectangular element that displays the industry name. When pressed, the button changes to a selected state (active display), and the industry selection unit 21a is activated, updating all rows in the job table input unit 32 to the standard values ​​for that industry.

[0037] The job type table input unit 32 is a table-format input unit in which each job type is arranged in rows, and each row has a job type name selection unit 32a, a number of employees input unit 32b, a desk work percentage input unit 32c, an AI reduction rate display unit 32d, and an hourly wage input unit 32e as columns. At the bottom of the table is an add button 32g for adding rows (adding job types), and at the end of each row is a delete button 32f for deleting the row.

[0038] The job title selection unit 32a is a pull-down input unit that allows the user to select one of several preset job titles or to enter a free-form text. In this embodiment, a total of six types are set: field / manual labor, office / accounting / general affairs, sales / external relations, management / executive, professional / practical work, and other (free input). When "other (free input)" is selected, the pull-down input unit switches to a text box input unit, which accepts free text input from the user and treats the entered text as the job title set by the user.

[0039] The employee count input section 32b is a numerical input box that allows the user to input the number of employees belonging to the relevant job category as an integer value in units of employees. It accepts only integer values ​​of 0 or greater, and the total number of employees display section 32h and the employee total number display section 33a of the employee average hourly wage display section 33 are automatically updated each time the input value changes. In addition, the scale determination means 22a performs a re-evaluation, and if the scale classification changes, the default value of the implementation cost setting section 35 is updated.

[0040] The desk work percentage input section 32c is an input section that allows users to input the percentage of their work time for the relevant job that is spent on desk work (information processing, document creation, communication, etc., tasks that can be replaced by AI) as an integer value (%) between 0 and 100. It is implemented as a numerical input box, and input is restricted if a value less than 0 or greater than 100 is entered. If an industry is selected by the industry preset selection section 31, the corresponding value in the industry correction table 24a is automatically set.

[0041] The AI ​​reduction rate display unit 32d is a read-only display unit that shows the percentage reduction in desk work time due to the introduction of AI (AI reduction rate). In this embodiment, a fixed value of 35% (0.35) is displayed. The system is configured not to accept changes by the user, and the cell is clearly indicated as unchangeable by visual means such as graying it out. This fixed value was set as the conservative median of empirical data from McKinsey Global Institute (2023), Peng et al. (2023), and Deloitte Tohmatsu (2024), but it is not limited to this and can be set to any appropriate value. In this embodiment, the AI ​​reduction rate displayed in the AI ​​reduction rate display unit 32d may also be configured to be changeable by the user. In this case, it is preferable to accept input in percentage format between 0 and 100, and to accept arbitrary numerical settings for each row.

[0042] The hourly wage input section 32e is a numerical input box that allows the user to input the hourly wage per employee for the relevant job type as an integer value in yen. The input value is used to calculate the weighted average hourly wage in the company-wide average hourly wage display section 33b, and is also referenced in the calculation of the monthly cost reduction amount S. In this embodiment, the initial values ​​differ for each job type: 2,000 yen for field / manual labor, 2,200 yen for office / accounting / general affairs, 2,500 yen for sales / external relations, and 3,500 yen for management / executive staff and professional / practical staff.

[0043] The employee total display unit 33a is a read-only display field that displays the total number of employees, calculated by summing the number of employees for all job categories entered in the employee count input unit 32b of the job category table input unit 32. It is updated in real time whenever the contents of the job category table input unit 32 change. The displayed employee total field also has a batch overwrite function that accepts direct input from the user, and the batch overwritten value is used as the input value for the scale determination means 22a.

[0044] The company-wide average hourly wage display section 33b is a read-only display field that calculates the weighted average hourly wage (hourly wage for all job types × total number of employees ÷ total number of employees) from the input values ​​of the number of employees input section 32b and the hourly wage input section 32e of the job type table input section 32, and displays it as a reference value. It is intended for users to check the average labor cost level for the entire company.

[0045] The implementation cost setting unit 35 is a collection of input units that allow users to input the costs required for an AI implementation project in four phases, and includes a conceptual diagnosis cost input unit 35a, a PoC cost input unit 35b, a main development cost input unit 35c, and a monthly operation cost input unit 35d. Each field is automatically set to a default value corresponding to the scale category determined by the scale determination means 22a, but the user can change it to any value.

[0046] The Conceptual Diagnosis Cost Input Section 35a is a numerical input box that allows users to input the costs required for the conceptual planning and current situation diagnosis phases of an AI implementation project in units of 10,000 yen. In this embodiment, the default values ​​for each scale are set as follows: Small scale: 400,000 yen, Medium scale: 1,000,000 yen, Large scale: 1,500,000 yen.

[0047] The PoC cost input section 35b is a numerical input box that allows users to input the costs required for the Proof of Concept (PoC) phase in units of 10,000 yen. PoC is a phase in which a small-scale proof-of-concept experiment is conducted before full-scale development. In this embodiment, the default values ​​for each scale are set as follows: small scale: 700,000 yen, medium scale: 2,000,000 yen, and large scale: 3,500,000 yen.

[0048] The development cost input section 35c is a numerical input box that allows users to input the cost required for the full-scale development and implementation phase of the AI ​​system in units of 10,000 yen. In this embodiment, the default values ​​for each scale are set as follows: small scale: 2 million yen, medium scale: 5.5 million yen, and large scale: 17.5 million yen.

[0049] The monthly operating cost input section 35d is a numerical input box that allows users to input monthly expenses in units of 10,000 yen, such as license fees, maintenance fees, and cloud infrastructure fees, which are incurred continuously each month after the AI ​​system starts operating. This value is used as the monthly cost in the calculation of the investment recovery period and therefore directly affects the monthly income and expenditure calculation. In this embodiment, the default values ​​for each scale are set as follows: small scale: 230,000 yen, medium scale: 450,000 yen, and large scale: 1,300,000 yen.

[0050] The simulation execution button 36 is a button element located at the bottom of the simulation input screen 30. When pressed by the user, it sequentially activates each calculation means of the calculation unit 22 (reduction time calculation means 22b, cost reduction calculation means 22c, payback period calculation means 22d, and conversion calculation means 22e), updating all display sections of the simulation results screen 50 with the latest calculated values.

[0051] As shown in Figure 6, the simulation results screen 50 includes a monthly time reduction display unit 51, a monthly cost reduction amount display unit 52, an annual cost reduction amount display unit 53, an investment recovery period display unit 54, a personnel reduction equivalent display unit 55, an implementation cost breakdown display unit 56, a time utilization conversion display unit 57, a job-specific time reduction display unit 58, a calculation formula display unit 59, and a print / output button 60.

[0052] The monthly time reduction display unit 51 is a display field that displays the total monthly time reduction Hi value ΣHi for all job types, calculated by the time reduction calculation means 22b, in the format "○○ hours". This value represents the total amount of work time that the AI ​​replaces for the entire organization in one month. It may also be displayed in the format "○○ hours / month".

[0053] The monthly cost reduction display unit 52 is a display field that displays the monthly cost reduction amount S calculated by the cost reduction calculation means 22c using formula 2 in the format "○○ million yen". This value represents the reduction effect equivalent to labor costs, calculated by multiplying the monthly reduction time by the hourly wage for each job type. It may also be displayed in the format "○○ million yen / month".

[0054] The annual cost reduction display unit 53 is a display field that displays the annual cost reduction amount, which is calculated by multiplying the monthly cost reduction amount S calculated by the cost reduction calculation means 22c by 12, in the format "○○ million yen". It is provided as an indicator that is useful for comparison with the annual budget. It may also be displayed in the format "○○ million yen / year".

[0055] The investment recovery period display unit 54 is a display field that displays the investment recovery period P calculated by the recovery period calculation means 22d using the formula 3 in the format "○○ months". If the monthly cost reduction amount does not exceed the monthly operating expenses (S ≤ M), it displays "Calculation not possible".

[0056] The personnel reduction display unit 55 is a display field that displays the personnel reduction E calculated by the formula 4 in the format "time equivalent to XX people". This value is obtained by dividing the total monthly reduction time by the monthly working hours per person (160 hours), and is provided so that managers can intuitively grasp the reduction effect in terms of the number of people. It may also be displayed in the format "equivalent to XX people".

[0057] The implementation cost breakdown display unit 56 displays a list of costs for each phase set in the cost input unit 21c, and also includes an initial cost total display unit 56a, a monthly operating cost display unit 56b, and a monthly net reduction amount display unit 56c. The initial cost total display unit 56a displays the sum of the values ​​from the concept diagnosis cost input unit 35a, the PoC cost input unit 35b, and the main development cost input unit 35c, while the monthly operating cost display unit 56b displays the value from the monthly operating cost input unit 35d.

[0058] The initial cost total display unit 56a is a display field that displays the total cost of the three phases set in the conceptual diagnosis cost input unit 35a, the PoC cost input unit 35b, and the main development cost input unit 35c shown in Figure 5, in the format "Total: XX million yen (excluding tax)". This value is used as the initial cost total F in the payback period calculation means 22d.

[0059] The monthly operating expense display section 56b is a display field that confirms and displays the current value of the monthly expenses set in the monthly operating expense input section 35d. It is displayed as part of the breakdown of implementation costs so that the user can make a final confirmation, as it is used as a value to calculate the investment recovery period.

[0060] The monthly net reduction display unit 56c displays the value obtained by subtracting the monthly operating expenses M from the monthly cost reduction amount S (SM) as the "monthly net reduction amount".

[0061] The time utilization conversion display unit 57 is a display unit that displays three types of conversion values ​​calculated by the conversion calculation means 22e side by side, and includes a sales visit conversion display unit 57a, a proposal creation conversion display unit 57b, and a training / development conversion display unit 57c.

[0062] The sales visit conversion display section 57a ​​is a display field that shows the value obtained by dividing the total monthly reduction time ΣHi by 3 hours per visit in the format "equivalent to XX new sales visits per month". 3 hours is a conversion factor based on the average time required for sales visit activities (travel, negotiation, and report preparation).

[0063] The proposal creation conversion display section 57b is a display field that shows the value obtained by dividing the total monthly reduction time ΣHi by 5 hours per item in the format "Equivalent to creating XX proposals / quotes per month". 5 hours is a conversion factor based on the average man-hours required to create a proposal / quote.

[0064] The training / development conversion display section 57c is a display field that shows the value obtained by dividing the total monthly reduction hours ΣHi by 8 hours per day in the format "Equivalent to XX days / month of employee training / development". 8 hours is a conversion factor based on the standard working hours per day.

[0065] The job-specific time reduction display unit 58 displays a value calculated using the formula: number of people × 160 hours × percentage of desk work × AI reduction rate of 35%.

[0066] The calculation formula display unit 59 shows the formula used to calculate each simulation result and the basis for determining the AI ​​reduction rate.

[0067] The print / output button 60 is a button element located at the bottom of the simulation results screen 50. When pressed by the user, it activates the approval document output means 23c. The approval document output means 23c calls the browser's window.print() function and uses the @media print media query to hide non-calculated elements, then prints or saves as a PDF the approval document format with the input values ​​and calculation results formatted. This allows for printing and saving as a PDF only after the content has been neatly adjusted for printing and PDF, rather than simply printing or saving the screen display as is.

[0068] The AI ​​reduction rate of 0.35 (35%) was set as the conservative median of the publicly available measured and estimated reduction effects (9% to 55%). By using a fixed value, arbitrary manipulation of the calculations is prevented, and a well-founded conservative estimate is provided. However, this is not the only option; the fixed value can be changed, or the user can change it, or any other configuration is possible as appropriate.

[0069] With the above configuration and operation, the AI ​​implementation ROI simulation system according to the present invention enables even business owners, managers, or personnel in charge of promoting AI implementation who lack specialized knowledge to complete highly accurate ROI calculations that reflect industry characteristics and job structure in a short amount of time.

[0070] The simulation results can be printed or saved as a PDF, allowing them to be immediately used as approval documents. Furthermore, while default implementation cost estimates are provided based on scale, users can change these figures to their own specifications.

[0071] Furthermore, by adopting a conservative reduction rate backed by empirical research as the AI ​​reduction rate, it is possible to provide well-founded estimates for management decisions. In addition, instead of using the desk work percentage as is, by applying the AI ​​reduction rate in addition to it (multiplicatively in this example), it is possible to obtain highly reliable simulation results adjusted by the AI ​​reduction rate while using the desk work percentage, an indicator that is intuitively easy for those considering AI implementation to understand.

[0072] Furthermore, by fixing the AI ​​reduction rate at a conservative median (35%) based on empirical research, and adjusting the proportion of desk work and hourly wages, it is possible to make calculations that are more in line with the actual organizational situation, thus preventing errors in management decisions due to overestimation.

[0073] Because the estimated implementation cost is automatically set according to the size category based on the number of employees entered, even those considering AI implementation for the first time can immediately grasp the cost and payback period.

[0074] Since the calculation results can be output in a printable style as an approval document format, internal decision-making documents can be created without the need to request external consultants, contributing to the efficiency of the organization's operations.

[0075] This invention is not limited to the configuration of the embodiments described above, and many other embodiments can be obtained. For example, the system is configured to multiply the proportion of desk work by a uniform AI reduction rate, but it is not limited to this. Different AI reduction rates can be set for each job type, or the system can be configured to allow setting work types such as computer work, paperwork, and telephone support for desk work, and the AI ​​reduction rate can be set individually for each work type. This allows for more accurate calculations. [Examples]

[0076] Next, we will describe Example 2, which adds AI-based input support to Example 1 described above. In Example 2, a function is added in which AI supports the input of each item into the industry preset selection unit 31 and the job type table input unit 32 as described in Example 1. All displays and operations after the industry preset selection unit 31 and the job type table input unit 32 have been entered remain the same as in Example 1.

[0077] Figure 7 shows the screen configuration when the external AI usage screen 80 is open, and Figure 8 shows the screen configuration of the in-house AI usage screen 90. Below these, the simulation input screen 30 described in Example 1 is displayed, but the illustration and description are the same as those described in Example 1, so they are omitted here.

[0078] As shown in Figure 7, the AI ​​Assist Button 70, which serves as the first accordion control unit, is displayed in the upper section. The AI ​​Assist Button 70 is composed of a horizontal strip element including an icon labeled "AI ASSIST," an explanatory text section, and an unfolding guidance text section. In the initial display state of this page, the external AI Assist Button 73, the in-HP AI Assist Button 75, the external AI usage screen 80, and the in-HP AI usage screen 90 are all set to a hidden state (hereinafter referred to as the "hidden state").

[0079] When the user clicks the AI ​​assist button 70, the corresponding expansion animation (max-height transition) is executed, and the external AI assist button 73 and the internal AI assist button 75 transition to a visible expanded state. At the same time, the arrow element 71 within the AI ​​assist button 70 rotates 180 degrees to visually indicate that it is in the expanded state (the equilateral triangle state shown in Figure 7).

[0080] When the user clicks the AI ​​assist button 70 again, the external AI assist button 73 and the in-HP AI assist button 75 return to their retracted state, and if there is an external AI usage screen 80 or an in-HP AI usage screen 90 that was previously deployed, these also transition to their retracted state in conjunction. At this time, the arrow element 71 within the AI ​​assist button 70 rotates 180 degrees to visually indicate that it is in the retracted state (an inverted triangle state, upside down compared to Figure 7).

[0081] The second accordion control unit is an input method selection interface that is displayed when the AI ​​assist button 70 is deployed, and has an external AI usage screen 80 and an internal AI usage screen 90.

[0082] When the user clicks the external AI assist button 73, the external AI usage screen 80 transitions to the expanded state. At this time, if the internal AI usage screen 90 is already in the expanded state, the internal AI usage screen 90 transitions to the collapsed state before the external AI usage screen 80 is expanded.

[0083] Similarly, when a user clicks the HP AI Assist button 75, the HP AI usage screen 90 transitions to an expanded state, and if the external AI usage screen 80 is already expanded, it transitions to a collapsed state. In other words, the external AI usage screen 80 and the HP AI usage screen 90 cannot be expanded simultaneously; they are expanded exclusively. If a button that is already expanded is clicked again, the corresponding panel transitions to a collapsed state.

[0084] The external AI usage screen 80 is an operation screen (Copy & Paste panel) that uses an external AI service to generate URLs, and is labeled "AI ASSIST Copy & Paste". It controls the display of the external AI usage screen 80.

[0085] The AI ​​usage screen 90 within the HP is an interactive input screen (Interaction panel) using the AI ​​chat function built into this page, and is labeled "AI ASSIST Interaction," controlling the display of the AI ​​usage screen 90 within the HP.

[0086] The external AI usage screen 80 provides a method for generating simulator input data using external AI services (ChatGPT®, Gemini, Claude, etc.; hereinafter referred to as "external AI"), and includes a prompt issuing unit 81 and a URL input application unit 85.

[0087] The prompt issuing unit 81 includes a copy button 83 for copying a predetermined prompt text to the clipboard and a preview display area 84 (not shown in the illustration) for the prompt content. The prompt issued here is a prompt that asks the user for the industry, job type, number of people and hourly wage for each job type, and then outputs a URL that includes this information (industry, job type, number of people and hourly wage for each job type) as parameters. As a result, the external AI can handle this process appropriately and can also provide support, including estimations, even if the user gives an ambiguous answer, by saying something like, "Generally, the values ​​are like this, what do you think?"

[0088] The URL input application unit 85 includes a text input field 86 for inputting a URL output by an external AI, and an application button 87 which acts as an application instruction unit for analyzing the URL and applying it to the simulator.

[0089] The operating procedure for the external AI usage screen 80 is as follows: When the user operates the copy button, a predetermined prompt text is written to the user's terminal clipboard. This prompt text includes the following steps, display format, format of the final output URL, how to write about registered job types, how to write about unregistered job types, a registered job type code correspondence table, industry-specific desk work percentage correction values, industry ID correspondence table (which industry ID corresponds to which industry name), estimation criteria for desk work percentage and hourly wage for unregistered job types (custom job types), an example URL, and notes. This allows the external AI to be controlled to output values ​​in a format that can be appropriately entered into each element to be input on the simulation input screen 30 (see Figure 5) described in Example 1.

[0090] The user pastes the prompt text written to the clipboard into the chat interface of an external AI (such as ChatGPT, Gemini, or Claude) of their choice and sends it.

[0091] The external AI, following the instructions written in the prompt text, generates URL parameters including industry, job code, number of people, percentage of desk work, and hourly wage, and outputs them as a URL in a predetermined format.

[0092] The user pastes the URL output by the external AI into the text input field 86 of the URL input application section and operates the apply button 87.

[0093] The URL input application unit 85 analyzes the query parameters from the input URL, converts them into job title labels by referring to the role code map table, sets numerical values ​​(number of people and hourly wage) for each row (job title) on the simulation input screen 30, inputs the total number of employees, the company-wide average hourly wage, and the company size that can be calculated from this information, along with the concept diagnosis cost, PoC cost, actual development cost, and monthly operating cost set for this company size, into the total number of employees display unit 33a, the company-wide average hourly wage display unit 33b, the concept diagnosis cost input unit 35a, the PoC cost input unit 35b, the actual development cost input unit 35c, and the monthly operating cost input unit 35d, and activates the calculation unit 22 (see Figure 4), which operates when the simulation execution button 36 (see Figure 5) is clicked, to display the calculation results on the simulation results screen 50 (see Figure 6).

[0094] The prompt text includes URL parameter formatting specifications (registered job code, custom job code, industry key, desk work percentage, and hourly wage specification), allowing the service provider to leverage external AI without incurring API usage. This has the effect of providing AI input assistance to users while keeping the service provider's API costs to zero.

[0095] The AI ​​usage screen 90 within the HP provides a method for generating simulator input data through natural language interaction with the AI ​​on this page, and includes a message display area 91, a text input field 86, a send button 87, an apply button (not shown) as an application instruction unit, and a reset button 88.

[0096] The message display area 91 is an area that displays response messages from the AI ​​and user input messages in chronological order.

[0097] The text input field 86 is a text area element that accepts natural language input from the user.

[0098] The submit button 87 is a button element that sends the input content to an AI registered as an AI that communicates with HP via a proxy server.

[0099] The Apply button (not shown in the illustration) is a button element that reflects the input data generated by the AI ​​onto the simulator, and is dynamically displayed when the AI ​​returns a response in a predetermined JSON format.

[0100] The reset button 88 is a button element that initializes the conversation history and redisplays the initial message.

[0101] The operating procedure for the AI ​​usage screen 90 on the HP website is as follows: When the AI ​​usage screen 90 on the HP is first deployed, an initial message from the AI ​​(requesting confirmation of the industry) is displayed in the message display area 91. When a user enters information such as industry and employee composition in natural language into the text input field 86 and operates the send button 87 or presses the Enter key in the input section 8b, the input content is sent to the registered AI via the proxy server.

[0102] Registered AIs return the response text to the AI ​​usage screen 90 on the website. If the AI's response includes a confirmation sheet format, the AI ​​will display a list of industry name, job title, number of people, percentage of desk work, hourly wage, and estimated annual income (hourly wage x 160 hours x 12 months), and ask the user to confirm the contents.

[0103] When the user confirms the content and enters an OK response in the text input field 86 and presses the submit button 87, the AI ​​generates a response in a predetermined JSON format (a format in which apply_simulation is included in the action field).

[0104] The AI ​​usage screen 90 within the HP detects a response in JSON format, displays the explanatory text excluding the JSON block in the message display area 91, and dynamically generates and displays an apply button (not shown).

[0105] When the user presses the Apply button, the JSON data is parsed, and the industry preset, custom industry name, and job configuration are set in each element of the simulation input screen 30 (see Figure 5). Then, the calculation unit 22 (see Figure 4) is activated, and the calculation results are displayed on the simulation results screen 50 (see Figure 6).

[0106] Furthermore, since the AI ​​operations performed on the AI ​​usage screen 90 within the HP utilize API keys and incur costs, access restrictions are in place to prevent excessive use by setting rate limits per IP address and a monthly token limit. This prevents various users from unnecessarily consuming AI tokens and incurring excessive costs.

[0107] With the above configuration, users can input the information necessary for the simulation with the support of AI. Therefore, even users who are unsure what to enter and with what values ​​in each item of the simulation input screen 30 can input the appropriate job titles and values ​​with the support of AI and easily obtain appropriate simulation results. In particular, once the AI ​​support is complete, the user can simply press the apply button 87 (or the apply button (not shown)) and the information determined with AI support (job title, number of people, percentage of desk work, hourly wage) will be entered into each item of the simulation input screen 30 (job title selection section 32a, number of employees input section 32b, desk work percentage input section 32c, hourly wage input section 32e), eliminating the need for cumbersome tasks such as copy and paste, and allowing for comfortable use. Furthermore, once this information is entered and determined, the calculation unit 22 automatically calculates and inputs the values ​​to be entered into the total number of employees display unit 33a, the company-wide average hourly wage display unit 33b, the concept diagnosis cost input unit 35a, the PoC cost input unit 35b, the actual development cost input unit 35c, and the monthly operating cost input unit 35d. The system then transitions to the state where the simulation execution button 36 is pressed, and the simulation results screen 50 is displayed. Therefore, users utilizing AI support can obtain simulation results with all items entered and calculations completed simply by completing the interaction with the AI ​​and clicking the apply button (not shown).

[0108] The two-stage accordion-type interface allows the AI ​​input assistance area to be hidden in the initial display state, thus providing AI input assistance functionality without compromising the visibility of the main functions of the simulation input screen 30 (simulator page).

[0109] Furthermore, by controlling the display of the external AI usage screen 80 and the internal AI usage screen 90 exclusively, it is possible to prevent users from becoming confused by displaying both panels simultaneously and to ensure clarity of operation.

[0110] Furthermore, since using the external AI utilization screen 80 eliminates API costs for the service provider, the service provider's cost burden can be minimized even when a large number of users utilize the service.

[0111] Furthermore, the AI ​​usage screen 90 on the HP conceals the API key by using a proxy server, and excessive API usage is suppressed by IP address-based rate limits and monthly token limits, allowing the AI ​​dialogue function to be provided within a reasonable cost range.

[0112] Furthermore, on the AI ​​usage screen 90 within the HP, the AI ​​returns data in a predetermined JSON format, and by operating the apply button, it is automatically set in the simulator's input form. This mechanism allows users to run simulations using only natural language conversation and perform ROI calculations without requiring any technical knowledge of form input.

[0113] Alternatively, instead of the copy button 83, you could provide an AI call button to open the AI ​​in a new window or tab. In this case, you could prepare buttons for each type of AI, such as a "ChatGPT button," a "Gemini button," and a "Claude button," and when one of them is clicked, the AI ​​will open in a new window (or new tab) on the web with the prompt copied. This would reduce the number of copy operations by one. Furthermore, for AIs that accept prompt input (such as ChatGPT), you could configure it so that the prompt is already pasted into the input field when opened. This would reduce the number of operations by another.

[0114] In the correspondence between this invention and its embodiments, the occupation type designation unit of this invention corresponds to the occupation name selection unit 32a of the occupation table input unit 21b and the occupation table input unit 32, and similarly, The calculation unit corresponds to the calculation unit 22, the time reduction calculation means 22b, or the cost reduction calculation means 22c. The number of people specification section corresponds to the number of enrolled input section 32b, The target work ratio specification unit corresponds to the desk work ratio input unit 32c, The simulation result output unit corresponds to the simulation result screen 50. The AI ​​implementation effect simulation system corresponds to the AI ​​implementation ROI simulation system 1. The AI ​​implementation effect simulation program corresponds to the AI ​​implementation ROI simulation program, This invention is not limited to the above-described embodiment and can take on various other forms. [Industrial applicability]

[0115] This invention can be widely used by companies, organizations, and other entities in manufacturing, construction, retail, professional services, and other sectors when making investment decisions regarding the introduction of AI. [Explanation of Symbols]

[0116] 1…AI-implemented ROI simulation system 21b... Job Category Table Input Section 22...Arithmetic section 22b...Reduction time calculation means 22c... Cost reduction calculation means 32... Job Category Table Input Section 32a... Job Title Selection Section 32b... Input section for the number of enrolled students 32c... Input section for the percentage of desk work 50...Simulation results screen

Claims

1. Multiple occupation type designation departments that accept occupation type designations, The section that accepts requests for the number of people, The Target Task Ratio Designation Department accepts designations for the percentage of tasks that can be reduced in working hours through the introduction of AI, A calculation unit that performs simulation calculations based on specified information, The system includes a simulation result output unit that outputs the results of the simulation performed by the calculation unit, The aforementioned number of people designation unit and the aforementioned target work ratio designation unit are provided so that one designation unit can be set for each of the aforementioned target work ratio designation units. The calculation unit performs calculations for the occupational type unit using the number of people specified in the number of people specification unit and the target work ratio specified in the target work ratio specification unit as elements, integrates the calculated values ​​for all occupational types, and outputs the time or cost that can be reduced by introducing AI as a reduction effect to the simulation result output unit. A message display area where the AI ​​asks the user about the occupation type, number of people, and percentage of the work covered, The input content to be sent to the aforementioned AI includes a text input field that accepts natural language input from the user, The system includes an application instruction unit that applies the response text generated by the AI ​​in response to the aforementioned input content to the occupation type specification unit, the number of people specification unit, and the target work ratio specification unit. AI implementation effect simulation system.

2. It includes an AI implementation cost designation unit that accepts designations for the AI ​​implementation costs required for AI implementation, The system includes an investment recovery period calculation unit that calculates the investment recovery period, which indicates how long it will take to recover the AI ​​implementation costs based on the aforementioned reduction effects. The AI ​​implementation effect simulation system according to claim 1.

3. Computers, Multiple occupation type designation departments that accept occupation type designations, The section that accepts requests for the number of people, The Target Task Ratio Designation Department accepts designations for the percentage of tasks that can be reduced in working hours through the introduction of AI, A calculation unit that performs simulation calculations based on specified information, It functions as a simulation result output unit that outputs the results of the simulation performed by the aforementioned calculation unit. The aforementioned number of people designation unit and the aforementioned target work ratio designation unit are provided so that one designation unit can be set for each of the aforementioned target work ratio designation units. The calculation unit performs calculations for the occupational type unit using the number of people specified in the number of people specification unit and the target work ratio specified in the target work ratio specification unit as elements, integrates the calculated values ​​for all occupational types, and outputs the time or cost that can be reduced by introducing AI as a reduction effect to the simulation result output unit. A message display area where the AI ​​asks the user about the occupation type, number of people, and percentage of the work covered, The input content to be sent to the aforementioned AI includes a text input field that accepts natural language input from the user, The system includes an application instruction unit that applies the response text generated by the AI ​​in response to the aforementioned input content to the occupation type specification unit, the number of people specification unit, and the target work ratio specification unit. AI implementation effect simulation program.

4. Multiple occupation type designation departments that accept occupation type designations, The section that accepts requests for the number of people, The Target Task Ratio Designation Department accepts designations for the percentage of tasks that can be reduced in working hours through the introduction of AI, A calculation unit that performs simulation calculations based on specified information, Using an AI implementation effect simulation system equipped with a simulation result output unit that outputs the results of the simulation performed by the calculation unit, The aforementioned number of people designation unit and the aforementioned target work ratio designation unit are provided so that one designation unit can be set for each of the aforementioned target work ratio designation units. A message display area where the AI ​​asks the user about the occupation type, number of people, and percentage of the work covered, The input content to be sent to the aforementioned AI includes a text input field that accepts natural language input from the user, The system includes an application instruction unit that applies the response text generated by the AI ​​in response to the aforementioned input content to the occupation type specification unit, the number of people specification unit, and the target work ratio specification unit. The calculation unit performs calculations for the occupational type unit using the number of people specified in the number of people specification unit and the target work ratio specified in the target work ratio specification unit as elements, integrates the calculated values ​​for all occupational types, and outputs the time or cost that can be reduced by introducing AI as a reduction effect to the simulation result output unit. A method for simulating the effects of introducing AI.

Images