Matching support device and matching support method

Abstract

To provide an improved technique in a matching system.SOLUTION: A matching support device according to one embodiment includes: first acquisition means for acquiring first tag information, which indicates one of multiple predefined items and corresponds to a new matter that an enterprise intends to request; second acquisition means for acquiring, for each of multiple partner candidates who intend to accept an order for the new matter, second tag information being the tag information and corresponding to the partner candidate; and output means for inputting, for the multiple partner candidates, the first tag information and the second tag information into AI, and outputting partner candidates selected using information output from the AI as partner candidates suitable for the new matter.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to a matching support device and a matching support method. [Background technology]

[0002] Technologies for matching between companies are known. For example, Patent Document 1 discloses a system that extracts words that represent the characteristics of a company from company information and performs matching using the extracted words. Patent Document 2 discloses a system that extracts core phrases from company information and performs matching using the extracted core phrases. Patent Document 3 discloses a system that infers topics from information on matching applicants / targets and performs matching using the inferred topics. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-193537 [Patent Document 2] International Publication No. WO2022 / 113286 [Patent Document 3] Patent No. 6802334 Summary of the Invention [Problem to be solved by the invention]

[0004] In the technologies of Patent Documents 1 to 3, when extracting or estimating words / core phrases / topics that represent characteristics from corporate information, inappropriate words / core phrases / topics may be extracted, which could hinder matching.

[0005] In contrast to this, the present invention provides a technique for performing matching using tags that are more suitable for matching. [Means for solving the problem]

[0006] One aspect of the present disclosure provides a matching support device having: a first acquisition means for acquiring first tag information, which is tag information indicating any of a plurality of predefined items and corresponds to a new project that an enterprise intends to request; a second acquisition means for acquiring second tag information, which is the tag information, corresponding to each of a plurality of partner candidates intending to receive the new project; and an output means for inputting the first tag information and the second tag information for the plurality of partner candidates to an AI, and outputting partner candidates selected using information output from the AI ​​as partner candidates suitable for the new project.

[0007] This matching support device has a display control means for displaying on a display means a case input screen for inputting case information relating to the new case, the case information including free input fields, and the first acquisition means may acquire the first tag information obtained by inputting at least a portion of the case information input via the case input screen into the AI.

[0008] This matching support device may have a writing means for writing the case information entered via the case input screen into a case database, and the first acquisition means may acquire the first tag information corresponding to the new case from the case database.

[0009] This matching support device may have a tagging request means that requests the AI ​​to generate tags from at least a portion of the case information recorded in the case database, and the writing means may write tag information obtained in response to the request into the case database as tag information corresponding to the case information.

[0010] This matching support device has a display control means for displaying on a display means a partner information input screen for inputting partner information relating to the plurality of partner candidates, the partner information including free-entry fields, and the second acquisition means may acquire the second tag information obtained by inputting at least a portion of the partner information input screen entered via the partner information input screen into the AI.

[0011] This matching support device may have a writing means for writing the partner information input via a partner information input screen into a partner database, and the second acquisition means may acquire the second tag information corresponding to the plurality of partner candidates from the partner database.

[0012] This matching support device may have a tagging request means that requests the AI ​​to generate a tag from at least a portion of the partner information recorded in the partner database, and the writing means may write the tag information obtained in response to the request into the partner database as tag information corresponding to the partner information.

[0013] The first tag information may include business-related tags related to business skills relevant to the execution of the new project, and the second tag information may include business-related tags related to the current occupations or past work histories of the multiple partner candidates, and the AI ​​may perform matching using the business-related tags included in the first tag information and the business-related tags included in the second tag information.

[0014] The first tag information may include non-work-related tags relating to matters other than the work skills, and the second tag information may include non-work-related tags relating to matters other than the occupations and past work history of the multiple partner candidates, and the AI ​​may perform matching using the non-work-related tags included in the first tag information and the non-work-related tags included in the second tag information.

[0015] This matching support device has a receiving means for receiving an evaluation of the partner candidate who has received the new project from a requester of the new project while the partner candidate who has received the new project is working on or after the new project has been completed, a display control means for displaying on a display means a project input screen for inputting project information related to the new project, the project information including free input fields, and a writing means for writing the project information input via the project input screen into a project database, and the writing means may write the evaluation received by the receiving means into the project database.

[0016] The output means may input the first tag information, the second tag information, and the evaluation for the plurality of partner candidates into an AI, and output a partner candidate selected using the information output from the AI ​​as a partner candidate suitable for the new project.

[0017] The matching support device may include a learning means for training the AI ​​using the evaluations recorded in the project database as training data.

[0018] Another aspect of the present disclosure provides a matching support method including the steps of: acquiring first tag information, which is tag information indicating any of a plurality of predefined items and corresponds to a new project that an enterprise intends to request; acquiring second tag information from the tag information, for each of a plurality of partner candidates intending to receive the new project, corresponding to the partner candidate; and inputting the first tag information and the second tag information for the plurality of partner candidates into an AI, and outputting partner candidates selected using information output from the AI ​​as partner candidates suitable for the new project. [Effects of the Invention]

[0019] According to the present invention, matching can be performed using tags that are more suitable for matching. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a diagram showing an overview of a matching support system 1 according to an embodiment. [Figure 2] FIG. 1 is a diagram showing the functional configuration of a matching support system 1. [Figure 3] FIG. 2 is a diagram illustrating an example of the hardware configuration of a server 10. [Figure 4] FIG. 2 is a diagram illustrating an example of the hardware configuration of a user terminal 30. [Figure 5] 4 is a sequence chart illustrating the operation of the matching support system 1. [Figure 6] 10 is a sequence chart illustrating a process for registering a new matter. [Figure 7] FIG. 3 is a diagram illustrating an example of a case input screen 32 for a new case. [Figure 8] FIG. 2 is a diagram illustrating an example of data registered in a case database DB1. [Figure 9] 10 is a sequence chart illustrating a partner registration process. [Figure 10] FIG. 4 is a diagram illustrating a partner information input screen 42. [Figure 11] FIG. 2 is a diagram illustrating an example of data registered in a partner database DB3. [Figure 12] 10 is a sequence chart illustrating details of a matching process. [Figure 13] FIG. 10 is a diagram illustrating an example of an outline of matching. [Figure 14] FIG. 1 is a diagram illustrating an overview of skill matching. DETAILED DESCRIPTION OF THE INVENTION

[0021] 1. Configuration FIG. 1 is a diagram showing an overview of a matching support system 1 according to one embodiment. The matching support system 1 provides a service (hereinafter simply referred to as a "matching service") that supports matching between a requester and a partner. Users of the matching support system 1 are broadly divided into (a) requesters, (b) partners, and (c) consultants. A requester is a business operator who is about to request or has actually requested a project, specifically, a corporation such as a company, or an organization without legal personality. A project is a task that is the subject of a request, such as marketing research or consulting. A partner is a person who accepts a project, specifically, for example, a freelance worker or a company. A consultant is a person who supports matching between a requester and a partner. A consultant is, for example, an employee of the operator of the matching support system 1 or a person commissioned by the operator.

[0022] The matching support system 1 has a server 10, an AI 20, a user terminal 30, a user terminal 40, and a user terminal 50. The server 10 is an information processing device that functions as a server for the matching service and is an example of a matching support device. The AI ​​20 is an AI that supports matching. The user terminal 30 is an information processing device used as a terminal by a requester. The user terminal 40 is an information processing device used as a terminal by a partner. The user terminal 50 is an information processing device used as a terminal by a consultant.

[0023] FIG. 2 is a diagram illustrating the functional configuration of the matching support system 1. The matching support system 1 includes a storage unit 11, a first acquisition unit 12, a second acquisition unit 13, an output unit 14, and a processing unit 99. The storage unit 11 stores various data and programs. In this example, the storage unit 11 stores a project database DB1, a client database DB2, and a partner database DB3 as databases used for matching. These databases will be described in detail later. The first acquisition unit 12 acquires first tag information, which is tag information indicating one of multiple predefined items and corresponds to a new project that an operator intends to request. The second acquisition unit 13 acquires second tag information, which is tag information, corresponding to each of multiple partner candidates who intend to accept a new project. The output unit 14 inputs the first tag information and second tag information for the multiple partner candidates to the AI ​​20 and outputs partner candidates selected using the information output from the AI ​​20 as partner candidates suitable for the new project. The processing unit 99 performs various processes.

[0024] In this example, the matching support system 1 further includes a display control means 15 and a writing means 16. The display control means 15 causes the display means 31 to display a case input screen 32 for inputting case information 502 related to a new case, the case information 502 including free input fields. Here, the first acquisition means 12 acquires first tag information obtained by inputting at least a portion of the case information 502 input via the case input screen 32 into the AI ​​20. Specifically, the first acquisition means 12 acquires first tag information corresponding to the new case from the case database DB1.

[0025] The matching support system 1 further includes a tagging request means 17. The tagging request means 17 requests the AI ​​20 to generate tags from at least a portion of the case information 502 recorded in the case database DB1. The writing means 16 then writes the tag information obtained in response to the request into the case database DB1 as tag information corresponding to the case information 502.

[0026] Furthermore, the display control means 15 causes the display means 41 to display a partner information input screen 42 for inputting partner information 552 related to multiple partner candidates, the partner information 552 including free-entry fields. The writing means 16 writes the partner information 552 input via the partner information input screen 42 into the partner database DB3. Here, the second acquisition means 13 acquires second tag information obtained by inputting at least a part of the partner information input screen 42 input via the partner information input screen 42 into the AI ​​20. Specifically, the second acquisition means 13 acquires second tag information corresponding to multiple partner candidates from the partner database DB3.

[0027] The tagging request means 17 requests the AI ​​20 to generate a tag from at least a part of the partner information 552 recorded in the partner database DB3. The writing means 16 then writes the tag information obtained in response to the request into the partner database DB3 as tag information corresponding to the partner information 552.

[0028] Furthermore, the matching support system 1 includes a receiving means 18 and a learning means 19. The receiving means 18 receives an evaluation (or feedback evaluation) of a partner candidate who has received an order for a new job from a requester of the new job, during or after the new job is completed. Here, the writing means 16 writes the evaluation received by the receiving means 18 into the job database DB1. The learning means 19 uses the evaluation recorded in the job database DB1 as training data to train the AI ​​20.

[0029] In this example, the storage means 11, the first acquisition means 12, the second acquisition means 13, the output means 14, the display control means 15, the writing means 16, the tagging request means 17, the receiving means 18, the learning means 19, and the processing means 99 are implemented in the server 10. The display means 31 is implemented in the user terminal 30. The display means 41 is implemented in the user terminal 40.

[0030] Furthermore, the AI ​​20 includes a machine learning model 21, a machine learning model 22, a machine learning model 23, and a machine learning model 24. These will be described in detail later.

[0031] 3 is a diagram illustrating an example of the hardware configuration of the server 10. The server 10 is a computer device having a CPU (Central Processing Unit) 101, a memory 102, a storage 103, and a communication IF 104. The CPU 101 is a processing device that performs various processes according to a program. The memory 102 is a main storage device that functions as a work area when the CPU 101 executes a program. The storage 103 is a non-volatile auxiliary storage device that stores various data and programs. The communication IF 204 is a device that communicates with other information processing devices according to a predetermined communication standard (for example, Ethernet (registered trademark)).

[0032] In this example, the storage 103 stores a program (hereinafter referred to as the "server program") for causing a computer to function as the server 10 in the matching support system 1. When the CPU 101 is executing the server program, the CPU 101 is an example of the first acquisition means 12, the second acquisition means 13, the display control means 15, the writing means 16, the tagging request means 17, the receiving means 18, the learning means 19, and the processing means 99, at least one of the memory 102 and the storage 103 is an example of the storage means 11, and the communication IF 204 is an example of the output means 14.

[0033] FIG. 4 illustrates an example of the hardware configuration of the user terminal 30. The user terminal 30 is a computer device, such as a personal computer (PC), smartphone, or tablet terminal, that includes a CPU 301, memory 302, storage 303, a communication IF 304, an input device 305, and an output device 306. The CPU 301 is a processing device that performs various processes according to programs. The memory 302 is a main storage device that functions as a work area when the CPU 301 executes a program. The storage 303 is a non-volatile auxiliary storage device that stores various data and programs. The communication IF 304 is a device that communicates with other information processing devices according to a predetermined communication standard (e.g., Ethernet (registered trademark), IEEE 802.11, 4G, or 5G). The input device 305 is a device that inputs data or user instructions to the user terminal 30 and includes, for example, at least one of a keyboard, a touch screen, and a microphone. The output device 306 is a device that outputs data or information and includes, for example, at least one of a display and a speaker.

[0034] Although detailed description will be omitted, user terminal 40 and user terminal 50 also have the same hardware configuration as user terminal 30 (reference symbols are shown in parentheses in FIG. 4). In this example, storage 303 stores a program (hereinafter referred to as "requester terminal program") for causing a computer to function as user terminal 30 in matching support system 1. When CPU 301 is executing the requester terminal program, output device 306 (specifically, a display) is an example of display means 31.

[0035] In this example, the storage 403 stores a program (hereinafter referred to as a "partner terminal program") for causing a computer to function as a user terminal 40 in the matching support system 1. When the CPU 401 is executing the partner terminal program, the output device 406 (specifically, a display) is an example of display means 41.

[0036] In this example, step S503 stores a program (hereinafter referred to as the "consultant terminal program") for causing the computer to function as the user terminal 50 in the matching support system 1. When the CPU 101 is executing the partner terminal program, the output device 506 is an example of the display means 51 (specifically, a display).

[0037] 2. Operation 2-1. Project registration and matching 5 is a sequence chart illustrating the operation of the matching support system 1 according to one embodiment. The matching support system 1 can be used by a client, a partner, and a consultant at any time, but for the sake of explanation, an example will be described in which processing is performed in the following order: a client registers a case, a potential partner registers a profile, and then matching takes place.

[0038] In step S101, the user terminal 30 accesses the server 10 to register a new case. In this example, the matching service is a membership-based service. A company that is a client (or intends to become a client) registers as a member with the matching service in advance. When registering as a member, the server 10 displays an input screen for member registration on the user terminal 30. The client inputs the requested information on this input screen. The user terminal 30 transmits this information to the server 10. The server 10 registers the received data in the client database DB2.

[0039] The requester database DB2 stores multiple records. Each record contains, for one member, the member's member ID (i.e., identification information) and attribute data. The attribute data is data that indicates the member's attributes, and includes, for example, at least one of the following: member type, name, address, industry, and contact information. Of these, member type indicates the type of member, i.e., whether the member is a requester.

[0040] Each record in the client database DB2 further includes a consultant ID 505. The consultant ID 505 is identification information of the consultant in charge of that client. In this example, a consultant is assigned to each client. The consultant in charge of a particular client is basically fixed, and the same consultant will be in charge of that client continuously. This consultant is called the consultant in charge.

[0041] Each record in the client database DB2 further includes a client interview history 503. The client interview history 503 is a record of interviews between the client and the consultant. The client interview history 503 includes, for each of a plurality of interviews held at different dates and times, the date and time of the interview and a summary of the contents of the interview. This summary is in the form of free input (or free text).

[0042] In step S102, the client and the consultant in charge register a new case.

[0043] Figure 6 is a sequence chart illustrating the process of registering a new case. A client who wishes to request a case contacts a consultant and requests an interview to register the new case. The consultant then meets with the client. This interview can be either online or face-to-face.

[0044] During the consultation, the consultant accesses the server 10 from his / her own user terminal 50 to register a new case. Upon receiving the access for registering a new case, the server 10 transmits data for the case input screen 32 for registering the new case to the user terminal 50 (step S201). Upon receiving this data, the user terminal 50 displays the case input screen 32 (step S202). Here, the case input screen 32 is also displayed on the user terminal 30.

[0045] FIG. 7 is a diagram illustrating an example of a case input screen 32 for a new case. This input screen has multiple areas, including area 321 and area 322. Area 321 has an input field for the identification name of the new case. Area 322 has an input field for the recruitment conditions for that case. The recruitment conditions are divided into multiple items, and in this example, they include industry, job type, job area, background / challenges, operating medium, business skills, personality, required skills, welcome skills, requested tasks, and NG items. Of these items, the job area is selected from multiple predefined options (i.e., multiple choice), but the other items are free-entry items. Hereinafter, the information requested in area 321 and area 322 will be collectively referred to as case information 502.

[0046] The area 323 includes UI objects, in this example, buttons, for issuing instructions to the server 10 or the AI ​​20. The registration button 831 is a button for issuing an instruction to the server 10 to register the case information 502.

[0047] Referring again to Figure 6, the consultant in charge inputs the case information 502 while interviewing the client (step S203). After the interview, the consultant in charge presses the registration button 831 to instruct the server 10 to register the data (step S204). The data registration instruction sent to the server 10 includes at least a portion of the case information 502.

[0048] When the server 10 receives a data registration instruction from the user terminal 50, it registers the case information 502 included in the instruction in the case database DB1 (step S205).

[0049] FIG. 8 is a diagram illustrating data registered in the case database DB1. The case database DB1 includes multiple records. Each record includes, for one case, case identification information 501 and case information 502. The case identification information 501 is identification information that uniquely identifies the case. The case information 502 includes, in addition to the information received from the user terminal 50 in step S204, an interview history and tag information. The interview history includes a client interview history 503 and a partner interview history 504. The client interview history 503 is a history of an interview between the consultant and the client. The partner interview history 504 is a history of an interview between the consultant and the partner. Details of these interview histories will be described later. The tag information is information indicating tags extracted from the case information 502. Here, a tag is a character string used to organize or classify the case information 502, and is sometimes called a label or keyword.

[0050] The tag information includes one or more tags. In this example, the tag information includes tags related to skills. In the matching support system 1, multiple business skills (the business skills referred to here are a different concept from the business skills in the example of FIG. 7) are defined in advance. That is, the server 10 stores a database (not shown) that defines business skills. One tag indicates one skill out of these multiple business skills.

[0051] Also in this example, tag information is generated from project information 502. Each record in the project database DB1 includes a flag indicating whether tag information has been generated. The initial value of this flag is "not yet generated," indicating that tag information has not yet been generated. When tag information is generated by the process described below, the value of this flag is rewritten to "generated." Each record also includes a consultant ID 505.

[0052] Referring again to Figure 6, when a predetermined event occurs, the server 10 requests the AI ​​20 to tag the case information 502 (step S206). This event may be, for example, an event that the current time reaches a predetermined time (e.g., 2:00 AM). Alternatively, this event may be another event, such as an event that a predetermined time has passed since the last time the case information 502 was tagged. Furthermore, "tagging" refers to generating tags from input information (here, the case information 502).

[0053] When requesting tagging, the server 10 extracts untagged case information 502 from the case database DB1. The server 10 inputs a tag generation request including the case identification information 501 and the extracted case information 502 to the AI ​​20. When the tag generation request is input, the AI ​​20 outputs a set of tags corresponding to the case information 502 included in the input request (step S207).

[0054] The AI ​​20 includes a model trained to output a tag corresponding to an input string, including a free input. In other words, the machine learning model 21 tags the free-input string. The AI ​​20 uses the machine learning model 21 to generate one or more tags from the project information 502. In this example, the machine learning model 21 performs tagging for each predefined tag item (hereinafter referred to as a "tagged item"). This definition is provided, for example, by the management company or operating company of the matching support system 1. The correspondence between the tagged items and each item in the project information 502 is defined in a database (not shown). In one example, the tagged items include nine items: "job type," "job area," "operation medium," "business skills," "personality," "essential skills," "requested tasks," "NG items," and "feedback." In this example, the tagged items are classified into "business-related tags" and "non-business-related tags." Business-related tags indicate matters directly related to the project, while non-business-related tags indicate more general matters not directly related to the project. This classification is exclusively defined by a database (not shown). For example, "job type," "job area," "operational media," "business skills," "essential skills," and "requested work" are job-related tags, while "personality," "NG items," and "feedback" are non-job-related tags. Among these, for example, "job type" corresponds to the items "industry" and "job type" in the project information 502, and "job area" corresponds to the item "job area" in the project information 502. For each tagged item, its value (possible value) is defined in the database (not shown). For example, the values ​​defined for the tagged item "job type" include "director," "planner," "media consultant," "designer," and "operation operator." When an arbitrary character string is input as the value of an item in the project information 502 corresponding to a certain tagged item, the machine learning model 21 selects and outputs one or more words (values) corresponding to the character string from the defined tag values.

[0055] The AI ​​20 transmits data including the extracted tags and the case identification information 501 to the server 10 as a response to the tag generation request (step S208). This response includes tagged items and their values. In the following, the tag of the tagged item "essential skills" will be referred to as "essential skills," and the tag of the tagged item "job area" will be referred to as "job area skills." Essential skills represent business skills that are considered essential for the execution of the case. Job area skills represent business skills related to the job area of ​​the case. Job area refers to the range or area of ​​activities or responsibilities within the case.

[0056] When the response is received from the AI ​​20, the server 10 registers (i.e., writes) one or more tags included in the response as tag information in a record stored in the case database DB1 that corresponds to the case identification information 501 included in the received response (step S209). This completes the new case registration process.

[0057] Referring again to Figure 5, in step S103, the partner and consultant register the partner's profile. New partner registration is independent of new case registration and can be performed at any time, but for convenience, it will be described below after new case registration.

[0058] 9 is a sequence chart illustrating the partner registration process. As already explained, this matching service is a membership-based service. A person who wishes to register as a partner (hereinafter referred to as "applicant") accesses the server 10 from a user terminal 40 to apply for registration as a partner in the matching support system 1. When the server 10 receives access for partner registration, it transmits data to the user terminal 40 to display a partner information input screen 42 for inputting information required for registration (step S301).

[0059] Upon receiving the data from the server 10, the user terminal 40 displays the partner information input screen 42 on the display means 41 (step S302). The applicant inputs the requested information (hereinafter referred to as "partner information 552") on the partner information input screen 42 (step S303).

[0060] FIG. 10 is a diagram illustrating an example of the partner information input screen 42. The partner information input screen 42 has multiple areas, areas 821 to 823. Area 821 has an input field for the identification name of the new partner. Area 822 has an input field for the partner's attributes. The attributes are divided into multiple items, and in this example, they include employment status, side job experience, self-introduction, specialty industry, specialty work, and achievements. These items are free-entry items. At least some of these items are further divided into sub-items. For example, employment status includes four sub-items: current occupation, current company name, work experience, and past work and industry.

[0061] Area 823 has an input field for information about the business skills possessed by the applicant. As already explained, business skills are predefined in the matching support system 1. In area 823, these multiple business skills are presented along with multiple options for their level (or degree of attainment, achievement, or proficiency). The applicant selects the level of the business skills they believe they possess from the options.

[0062] Business skills are defined in a database (not shown) in the matching support system 1. Business skills are designed to suit the business (e.g., marketing) targeted by the matching support system 1. Business skills are divided into major and minor categories. Major categories include, for example, "Affiliate Marketing," "App Development," "App Advertising," and "ToC Advertising Strategy Planning." Each of these major categories has its own minor categories defined. For example, the major category "Affiliate Marketing" has the minor categories "Planning," "Operational Direction," and "Proposal / Review." The major category "App Development" has the minor categories "Requirements Definition," "Basic Design," "Detailed Design," "Development / Implementation / Testing," "Store Application," and "Maintenance / Operation." The default value for each minor category is "N / A" or blank. Five levels are set as the minor category values, and applicants can select the level they believe applies to them from a pull-down menu. For example, an applicant who believes that he or she is good at planning affiliate marketing would select level 5 for the sub-item "Planning" of the major item "Affiliate Marketing." In this way, the applicant enters information about his or her own business skills.

[0063] Referring again to Figure 9, when the entry is complete, the applicant instructs the server 10 to provisionally register the partner information 552 (step S304). When provisional registration is instructed, the server 10 writes the instructed partner information 552 together with the applicant's partner identification information 551 into the partner database DB3 (step S305). At this point, the applicant is still provisionally registered, and a flag indicating that the registration is provisional is written into the partner database DB3.

[0064] In this example, in the matching support system 1, an interview with a consultant is required to register a partner. A consultant in charge is assigned to each provisionally registered applicant. The consultant in charge interviews the applicant (step S306). During (or after) the interview, the consultant accesses the server 10 from the user terminal 50. Upon receiving access from the user terminal 50, the server 10 transmits a partner information input screen to the user terminal 50 (step S307). The partner information input screen includes an area for displaying partner information 552 for the applicant that is provisionally registered in the partner database DB3. The consultant in charge adds to or modifies the partner information 552 based on the content of the interview (step S308).

[0065] In this example, the information that the consultant adds or modifies to partner information 552 includes an evaluation of the partner. The evaluation is broadly divided into two categories: business skills and personality. Business skills are general skills required for project execution. Business skills include six categories, such as presentation skills, logical thinking, coaching skills, self-motivation, document creation skills, and communication skills. Note that the business skills evaluated here are different from the business skills that the partner enters when registering. Personality is a more general personality category that is separate from project execution. Personality includes seven categories, such as studious, patient, caring, straightforward, cooperative, active, and reliable. Both business skills and personality are evaluated on a five-point scale, and the consultant in charge subjectively scores the partner during the interview.

[0066] When the addition or correction of the partner information 552 is completed, the consultant instructs the server 10 to register the partner information 552 (step S309). Upon receiving an instruction from the user terminal 50, the server 10 registers the added or corrected partner information 552 in the partner database DB3 (step S310).

[0067] When a predetermined event occurs, the server 10 requests the AI ​​20 to tag the partner information 552 (step S311). This event may be, for example, an event that the current time reaches a predetermined time (e.g., 2:00 AM). Alternatively, this event may be another event, such as an event that a predetermined time has passed since the last time the partner information 552 was tagged.

[0068] When requesting tagging, the server 10 extracts untagged partner information 552 from the partner database DB3. The server 10 inputs a tag generation request including the partner identification information 551 and the extracted partner information 552 to the AI ​​20. When the tag generation request is input, the AI ​​20 generates a set of tags corresponding to the partner information 552 included in the input request (step S312).

[0069] In this example, the AI ​​20 includes a model trained to output a tag corresponding to an input string, including a free input, in other words, a machine learning model 22 that tags the free input string. The AI ​​20 uses the machine learning model 22 to generate one or more tags from the partner information 552. In this example, the machine learning model 22 performs tagging for each predefined tagging item. At least some of these tagging items are common to the tagging items used to tag the project information 502. The correspondence between these tagging items and each item of the partner information 552 is defined in a database (not shown).

[0070] The AI ​​20 transmits the generated tag set to the server 10 (step S313). The server 10 registers the tag set received from the AI ​​20 as tag information in the record of the partner in the partner database DB3 (step S314). This completes the new partner registration process.

[0071] FIG. 11 is a diagram illustrating an example of data registered in the partner database DB3. The partner database DB3 includes multiple records. Each record includes partner identification information 551 and partner information 552 for one partner. The partner identification information 551 is identification information that uniquely identifies the partner. The partner information 552 includes a profile, business skills, an evaluation, an interview history, and tag information. The profile and business skills are information entered by the partner in step S303 (some information may be added or corrected by the consultant in step S308). The evaluation is an evaluation of the business skills and personality entered by the consultant in step S308. The interview history includes the partner interview history 504. The tag information is information indicating tags generated from the partner information 552. Like the case information 502, the tag information is categorized into business-related tags and non-business-related tags.

[0072] Referring again to Figure 5, when a predetermined event occurs, the server 10 matches the job with a partner (step S104). The event that triggers the matching is, for example, an event that the current time reaches a predetermined time (for example, 2:00 AM). Alternatively, this event may be another event, such as an event that a predetermined time has passed since the previous matching was performed.

[0073] FIG. 12 is a sequence chart illustrating the details of the matching process. In step S401, the server 10 selects one case to be subjected to the matching process from among multiple cases registered in the case database DB1. Hereinafter, the case to be subjected to the matching process will be referred to as the "target case." In the case database DB1, each record has a flag indicating whether matching has been completed. The initial value of this flag is "uncompleted," indicating that matching has not been completed, but once the matching process for that case is completed, the value of the flag is rewritten to "completed." The server 10 refers to this flag to identify cases for which matching processing has not been completed, and selects one case from among them.

[0074] In step S402, the server 10 selects, as a plurality of partner candidates, a plurality of partners to be subjected to the matching process from a plurality of partners registered in the partner database DB3. The selection of the plurality of partner candidates is performed, for example, based on the partner information 552. In one example, the server 10 selects, as a plurality of partner candidates, partners whose industry included in the partner information 552 is the same as the industry included in the project information 502. Alternatively, the server 10 selects, as a plurality of partner candidates, partners whose service provision areas or unit cost prices satisfy the client's conditions (in this case, the service provision areas or unit cost prices are registered in the partner database DB3). Note that, in step S402, no particular narrowing down of partner candidates is performed, and all partners registered in the partner database DB3 may be selected as a plurality of partner candidates.

[0075] In step S403, the server 10 sends a matching request to the AI ​​20 for one partner candidate (hereinafter referred to as the "target partner candidate") selected from the multiple partner candidates selected in step S402. This matching request includes project identification information 501 and at least a portion of the project information 502, as well as partner identification information 551 and at least a portion of the partner information 552 of the target partner candidate. At least a portion of the project information 502 includes first tag information. The first tag information is tag information related to the target project and is extracted from the project database DB1. The first tag information includes, for example, business-related tags and non-business-related tags of the target project. At least a portion of the partner information 552 includes second tag information and an evaluation. The second tag information is tag information related to the target partner candidate and includes business-related tags and non-business-related tags. The evaluation is an evaluation of business skills and personality. The second tag information and the evaluation are extracted from the partner database DB3.

[0076] Upon receiving the matching request, the AI ​​20 performs a matching process. In this example, the matching process is divided into two stages: matching based on skills (hereinafter referred to as "skill matching") (step S404) and matching based on personality (hereinafter referred to as "personality matching") (step S405).

[0077] FIG. 13 is a diagram illustrating an example of an outline of matching. When tag information (first tag information) of a target case and tag information (second tag information) of a target partner candidate are input, the AI ​​20 first performs skill matching. When a business-related tag in the first tag information and a business-related tag in the second tag information are input, the machine learning model 23 outputs the degree of match (or suitability) between them. The machine learning model 23 narrows down the multiple partner candidates to a predetermined number (e.g., 10 people) based on the degree of match.

[0078] FIG. 14 is a diagram illustrating an example of an overview of skill matching. In skill matching, target partner candidates are classified into five ranks, from rank 1 to rank 5. Rank 1 is the group with the lowest match, and rank 5 is the group with the highest match. For example, a target partner candidate with a required skill match rate of 90% or more is classified as rank 5, and a target partner candidate with a required skill match rate of 30% is classified as rank 3. For example, a target partner candidate with a required skill match rate of 0% but a job-skill match rate of 70% is classified as rank 2.

[0079] Referring again to FIG. 13, personality matching is performed on the target partner candidates narrowed down to a predetermined number. When the non-work-related tags from the first tag information and the non-work-related tags and evaluations from the second tag information are input, the machine learning model 24 outputs the degree of match between them. For partner candidates with the same rank among the narrowed down partner candidates, the AI ​​20 ranks them according to the degree of match in personality matching. The server 10 repeatedly executes the processes of steps S403 to S405 for all of the multiple partner candidates selected in step S402. Note that while FIG. 12 illustrates an example in which target partner candidates are identified one by one and the process is repeated, data on all of the multiple partner candidates may be input to the AI ​​20.

[0080] 12 again. The AI ​​20 transmits the matching result to the server 10 (step S406). The matching result includes partner identification information 551 of the multiple partner candidates (ranked) extracted by the matching process. The server 10 writes the received matching result into the record of the target case in the case database DB1 (step S407).

[0081] Referring again to FIG. 5, in this example, the matching process starts and ends regardless of instructions from the user terminal 30 (for example, automatically in the middle of the night). The user terminal 30 can access the server 10 at any time to check the matching results (step S105). When accessed by the user terminal 30, the server 10 reads the matching results for the project related to the requester from the project database DB1. The server 10 transmits the read matching results to the user terminal 30 (step S106). In step S107, the user terminal 30 displays the partner candidates included in the matching results. This screen includes, for example, a list of partner candidates. When the requester performs an operation to select one partner candidate from the list, the user terminal 30 displays detailed information about the selected partner candidate. The detailed information includes at least a portion of the partner information 552 recorded in the partner database DB3.

[0082] The client selects the partner candidate that the client believes is most suitable for the project from the list of extracted partner candidates. The user terminal 30 transmits the identification information of the selected partner candidate to the server 10 (step S108). Upon receiving a notification from the user terminal 30 that a partner candidate has been selected, the server 10 notifies the user terminal 50 of the consultant in charge (step S109). The consultant in charge contacts the selected partner candidate to make a final confirmation as to whether or not the partner candidate is willing to accept the project. Once the consultant in charge confirms the partner candidate's willingness to accept the project, the consultant in charge notifies the server 10 that the partner candidate's willingness to accept the project has been confirmed (step S110). Upon receiving the notification from the user terminal 50, the server 10 registers the partner candidate in the project database DB1 as an official partner for the project (step S111).

[0083] In this example, matching between potential partners and cases is automatically performed based on freely entered case information and tag information obtained from freely entered partner information. Since both business-related and non-business-related tags are taken into consideration during matching, more comprehensive matching is possible, taking into account the characteristics or attributes of the case and partner.

[0084] Feedback Once a formal partner is selected for a project, the project progresses between the requester and the partner. The requester and the partner each periodically (e.g., once a week or once a month) input feedback about the project to the matching support system 1. The server 10 records the feedback input by the requester and the partner as an interview history in the project database DB1, the requester database DB2, and the partner database DB3. In the project database DB1, the interview history of the requester and the partner is recorded in the record of the project. In the requester database DB2, the interview history of the requester is recorded together with project identification information 501 in the record of the requester. In the partner database DB3, the interview history of the partner is recorded together with project identification information 501 in the record of the partner. The interview history of the client recorded in the project database DB1, the interview history of the client recorded in the client database DB2, and the interview history of the client recorded in the partner database DB3 may all be the same or may be partially different. Similarly, the interview history of the partner recorded in the case database DB1 and the interview history of the partner recorded in the partner database DB3 may be entirely the same or partly different.

[0085] The format of the interview history is defined in the matching support system 1. The interview history includes, for example, evaluations and free-form fields for predetermined evaluation items. The evaluation items include, for example, four items: "deadline," "accuracy of work," "communication," and "skill sufficiency." Each item is evaluated on a five-point scale. The evaluation is directed at the partner. In other words, the requester's feedback evaluates the other party, while the partner's feedback is a self-evaluation.

[0086] When the client and partner input their feedback, the consultant in charge will interview each of them. During the interview, the consultant in charge accesses the server 10 from the user terminal 50 and conducts the interview while referring to the interview history (feedback). If the consultant in charge feels it is necessary, he or she can amend or add to the contents of the interview history (feedback) recorded in the database.

[0087] At a predetermined timing (for example, periodically), the server 10 feeds back the interview history recorded in the case database DB1 to at least one of the machine learning model 23 and the machine learning model 24. Here, feeding back the interview history to the machine learning model means causing the machine learning model to learn (additional learning) using data obtained from the interview history as training data. In one example, this learning is performed as follows.

[0088] First, the server 10 calculates an overall score for the work of the partner who accepted the case based on the evaluation in the interview history. This overall score corresponds to the satisfaction level of the client when the case requested by the partner is completed. The evaluation in the interview history is numerical data, and a formula for calculating the overall score from the evaluation is defined in the server 10. The server 10 sends a learning request to the AI ​​20. This learning request includes tag information for the case, tag information for the partner, and the overall score. The AI ​​20 uses the received data as training data to learn. In one example, the business-related tags in the tag information for the case and the tag information for the partner are provided as training data to the input layer, and the overall score is provided as training data to the output layer, and the machine learning model 23 is trained. Furthermore, the non-business-related tags in the tag information for the case and the business-related tags and evaluation in the tag information for the partner are provided as training data to the input layer, and the overall score is provided as training data to the output layer, and the machine learning model 24 is trained.

[0089] Furthermore, the evaluation in the interview history may be used in the matching process between the case and the partner, similar to the evaluation in the partner information 552. According to this example, feedback from the client or partner or the results of the interview with the consultant can be reflected in the learning of the AI ​​20.

[0090] 3. Variations The present invention is not limited to the above-described embodiment, and various modifications are possible. Some modifications will be described below. Two or more of the following items may be used in combination.

[0091] (1) Data used for matching processing The data used in the matching process is not limited to those exemplified in the embodiments. Other data may be used in addition to or instead of the data exemplified in the embodiments. For example, the information input to AI20 is not limited to tag information; untagged information, such as the partner's profile (age, etc.), project conditions (delivery date, compensation, etc.), and the client's profile (capital, years in business, etc.), may be input to AI20 and used for matching. Alternatively, client information recorded in client database DB2 may be tagged, and the tag may be input to AI20 and used for matching.

[0092] (2) Details of the matching process The details of the matching process are not limited to those exemplified in the embodiment. For example, the AI20 is not limited to a process that performs two stages of matching using business-related tags (or skill matching) and matching using non-business-related tags (or personality matching). For example, the AI20 may extract partner candidates using skill matching alone. Alternatively, the AI20 may have a machine learning model that outputs a comprehensive match degree using business-related tags and non-business-related tags.

[0093] Furthermore, the data used in the matching process is not limited to those exemplified in the embodiments. In the embodiments, at least a portion of the data used in the matching process may be replaced with other data or omitted. Alternatively, at least a portion of the client information may be used in the matching process. In this case, the client information is tagged in advance, similar to the partner information. AI20 performs matching using this tag information. In this case, the machine learning model of AI20 has been trained in advance using the tags of the client information.

[0094] (3) AI In the embodiment, an example in which the AI ​​20 performs several processes has been described, but these processes may be implemented by dividing them into multiple AI engines. Alternatively, the AI ​​20 in the embodiment may be interpreted as a collection of multiple AI engines each performing a different process.

[0095] (4) User registration and case registration The specific method by which a client or a partner registers as a user in the matching support system 1, or the specific method by which a client registers a new case in the matching support system 1, is not limited to the example given in the embodiment. For example, the client or the partner may input their profile information themselves without the intervention of a consultant in at least one of the user registration and the registration of a new case by the client or the partner.

[0096] (5) Involvement of consultants The degree of involvement of the consultant with the client or partner is not limited to that exemplified in the embodiment. In at least a part of the processing described in the embodiment as involving a consultant, the client or partner may perform the processing by themselves without the involvement of a consultant.

[0097] (6) UI, data format, and processing The UI and data format illustrated in the embodiments are merely exemplary, and the UI and data format according to the present invention are not limited to these. Furthermore, the processes described using sequence charts in the embodiments are merely exemplary, and the order of the processes may be changed or some processes may be omitted. Furthermore, the events that trigger each process are not limited to those illustrated in the embodiments. For example, the processes described in the embodiments as the server 10 sending a request to another device, such as the AI ​​20, in response to some trigger may be performed in a manner in which the other device, such as the AI ​​20, notifies the server 10 in response to some trigger, and the server 10, upon receiving this notification, transmits the necessary data to the other device.

[0098] (7) Target of tagging The data to be tagged is not limited to those exemplified in the embodiment. For example, interview information (especially freely entered items) recorded in the partner database DB3 may be tagged and recorded as tag information in the partner database DB3. This tag information is referenced when matching subsequent cases.

[0099] (8) Functional configuration The functional configuration of the matching support system 1 described in Fig. 2 is an example, and the functions of the matching support system 1 are not limited to this example. Some of the functional elements described in Fig. 2 may be omitted, and functional elements not shown may be added.

[0100] (9) Hardware configuration The hardware configuration of the matching support system 1 is not limited to the examples described in the embodiments. Some of the functions of the server 10 described in the embodiments may be implemented in other hardware such as the AI ​​20, the user terminal 30, the user terminal 40, or the user terminal 50. The server 10 may be a physical server or a virtual server (including a so-called cloud).

[0101] (10) Program The program executed by a processing device such as CPU 101 may be distributed in a form recorded on a computer-readable non-transitory recording medium (e.g., a DVD-ROM), or may be provided in a form that can be downloaded from a server on a network. [Explanation of symbols]

[0102] 1...Matching support system, 10...Server, 101...CPU, 102...Memory, 103...Storage, 104...Communication IF, 11...Storage means, 12...First acquisition means, 13...Second acquisition means, 14...Output means, 15...Display control means, 16...Writing means, 17...Tagging request means, 18...Acceptance means, 19...Learning means, 20...AI, 204...Communication IF, 21...Machine learning model, 22...Machine learning model, 23...Machine learning model, 24...Machine learning model, 30...User terminal, 301...CPU, 302...Memory, 303...Storage, 304...Communication IF, 305...Input device, 306...Output device, 31...Display means, 32...Job input screen, 321...Area area, 322...area, 323...area, 40...user terminal, 401...CPU, 403...storage, 406...output device, 41...display means, 42...partner information input screen, 50...user terminal, 501...case identification information, 502...case information, 503...client interview history, 504...partner interview history, 505...responsible consultant ID, 506...output device, 51...display means, 52...partner information input screen, 551...partner identification information, 552...partner information, 821...area, 822...area, 823...area, 831...registration button, 99...processing means, DB1...case database, DB2...client database, DB3...partner database

Claims

1. a first acquisition means for acquiring, from AI, first tag information indicating any one of a plurality of predefined items, the first tag information corresponding to a new case for which a business operator has input case information for requesting; a second acquiring means for acquiring, from the AI, second tag information corresponding to each of a plurality of partner candidates who have input partner information indicating their own attributes in order to receive the new order, the tag information; an output means for inputting the first tag information and the second tag information for the plurality of partner candidates into the AI, and outputting partner candidates selected using information output from the AI ​​as partner candidates suitable for the new case; and The AI ​​is a first machine learning model that has been trained to output tag information including defined tagging items and their values ​​corresponding to a freely input character string that is input as the case information; a second machine learning model that has been trained to output tag information including defined tagging items and their values ​​corresponding to a freely input character string that is input as the partner information; a third machine learning model that has been trained to output a degree of match between the first tag information and the second tag information when the first tag information and the second tag information are input; and the first acquisition means acquires the first tag information from the first machine learning model by inputting case information of the new case into the first machine learning model; the second acquisition means acquires the second tag information from the second machine learning model by inputting partner information of the plurality of partner candidates into the second machine learning model; The output means outputs partner information selected using the degree of match output from the third machine learning model by inputting the first tag information and the second tag information to the third machine learning model. Matching support device.

2. a display control means for displaying on a display means a case input screen for inputting case information relating to the new case, the case information including free input items; The first acquisition means acquires, as the first tag information, tag information obtained by inputting the free input items of the case information input via the case input screen into the first machine learning model. The matching support device according to claim 1 .

3. a writing means for writing the case information input via the case input screen into a case database; a tagging request means for requesting the AI ​​to generate tags from untagged case information among the case information recorded in the case database, the writing means writes the tag information obtained in response to the request into the case database as tag information corresponding to the case information; The first acquisition means acquires a tag corresponding to the new case from the case database as the first tag information. The matching support device according to claim 2.

4. a display control means for displaying on a display means a partner information input screen for inputting partner information relating to the plurality of applicants who wish to become partner candidates, the partner information including free input items; The second acquisition means acquires, as the second tag information, tag information obtained by inputting the free input items of the partner information input screen input via the partner information input screen into the second machine learning model. The matching support device according to claim 1 .

5. a writing means for writing the partner information inputted via the partner information input screen into a partner database; a tagging request means for requesting the AI ​​to generate tags from the partner information that has not yet been tagged among the partner information recorded in the partner database; the writing means writes the tag information obtained in response to the request into the partner database as tag information corresponding to the partner information; The second acquisition means acquires the second tag information corresponding to the plurality of partner candidates from the partner database. The matching support device according to claim 4.

6. the first tag information includes a business tag related to a business skill related to the execution of the new project, the second tag information includes a business tag related to the current occupation or past work history of the plurality of partner candidates; the third machine learning model has been trained to output a degree of match between the first tag information and the second tag information when a business-related tag as the first tag information and a business-related tag as the second tag information are input; The AI ​​performs matching using the business-related tag included in the first tag information and the business-related tag included in the second tag information. The matching support device according to claim 1 .

7. the first tag information includes a non-business tag related to matters other than the business skill, the second tag information includes non-work-related tags related to matters other than the occupations and past work histories of the plurality of partner candidates; The AI ​​has a fourth machine learning model (24) that has been trained to output a degree of match between the first tag information and the second tag information when a non-business tag as the first tag information and a non-business tag as the second tag information are input, The AI ​​performs matching using the non-business tag included in the first tag information and the non-business tag included in the second tag information. The matching support device according to claim 6.

8. a receiving means for receiving an evaluation of the partner candidate from the client of the new project during or after the new project is completed by the partner candidate who has received the order for the new project; a display control means for displaying on a display means a case input screen for inputting case information relating to the new case, the case information including free input fields; a writing means for writing the case information input via the case input screen into a case database; and The writing means writes the evaluation received by the receiving means into the case database. The matching support device according to claim 1 .

9. The output means inputs the first tag information, the second tag information, and the evaluation for the plurality of partner candidates to an AI, and outputs a partner candidate selected using the information output from the AI ​​as a partner candidate suitable for the new project. The matching support device according to claim 8.

10. A learning means for learning the AI ​​using the evaluations recorded in the case database as training data. The matching support device according to claim 8.

11. A computer comprising: acquiring, from AI, first tag information indicating any one of a plurality of predefined items, the first tag information corresponding to a new case for which the business operator has input case information for requesting; acquiring, from the AI, second tag information corresponding to each of a plurality of partner candidates who have input partner information indicating their own attributes for receiving the new project, as the tag information; inputting the first tag information and the second tag information for the plurality of partner candidates into the AI, and outputting partner candidates selected using information output from the AI ​​as partner candidates suitable for the new case; and The AI ​​is a first machine learning model that has been trained to output tag information including defined tagging items and their values ​​corresponding to a freely input character string that is input as the case information; a second machine learning model that has been trained to output tag information including defined tagging items and their values ​​corresponding to a freely input character string that is input as the partner information; a third machine learning model that has been trained to output a degree of match between the first tag information and the second tag information when the first tag information and the second tag information are input; and In the step of acquiring the first tag information, the first tag information is acquired from the first machine learning model by inputting case information of the new case into the first machine learning model; In the step of acquiring the second tag information, the second tag information is acquired from the second machine learning model by inputting partner information of the plurality of partner candidates into the second machine learning model; In the outputting step, the first tag information and the second tag information are input to the third machine learning model, and partner information selected using the degree of match output from the third machine learning model is output. Matching support methods.

Images