A talent market system using generative AI and a talent market program using generative AI

Abstract

By providing more comprehensive and diversified talent assessment and recommendation based on diverse input data, this system assists recruiters in making decisions. A TMP system 1, designed to assist recruiters and applicants in matching, comprises: an unstructured information processing unit 20, which uses generative AI 2 to output resume / performance data 22 for each talent from unstructured talent data 21; an assessment item setting unit 30, which uses generative AI 2 to extract assessment items for each talent based on recruitment data 31 and output them as skill definitions 32; and a recommendation processing unit 40, which uses generative AI 2 to select recommended items from the assessment items in skill definitions 32, assesses each talent according to separately set assessment criteria based on resume / performance data 22, and generates recommendation reasons for recommended talents using generative AI 2, outputting them along with the assessment results.

Description

Technical Field

[0001] This invention relates to a technology that helps manage corporate human capital, and in particular to an effective technology for implementing a talent market system and procedures. Background Technology

[0002] In Japanese companies facing a shrinking labor force, maximizing the efficient and flexible use of human capital is essential. This means that management must take a company-wide perspective and link the talent strategy—which focuses on securing and developing human resources to execute business strategies—with the business strategies developed by each business unit to leverage their human resource advantages (i.e., management strategy). This maximizes the value of human capital and connects it to the enhancement of medium- to long-term corporate value.

[0003] However, in today's VUCA (Volatility, Uncertainty, Complexity, Ambiguity) world, where the future is difficult to predict, business strategies are constantly changing, and the mobility and diversity of talent are also increasing. Under the centralized decision-making structure that was previously dominated by management and human resources departments, it is extremely difficult to cope with both of these changes simultaneously.

[0004] Therefore, we are currently exploring the formation of an "internal talent market" mechanism (i.e., Talent Marketplace), which uses market principles to match the business required to execute the business strategy with the abilities, performance, and aspirations of individual talents. Through this mechanism, we can flexibly link talent strategy with business strategy, thereby improving employee engagement and productivity on-site.

[0005] As a technology for determining job assignments by matching recruiters (employers) and job seekers (job seekers), for example, Japanese Patent Application Publication No. 2004-62270 (Patent Document 1) describes the following: For multiple categories common to both employers and job seekers, among multiple options set for each, the employer's answer information and the job seeker's answer information are compared, and an evaluation score from the employer's perspective on the job seeker and an evaluation score from the job seeker's perspective on the employer are calculated respectively. Based on these scores, a match is made between the employer and the job seeker.

[0006] Existing technical documents

[0007] Patent documents

[0008] Patent Document 1: Japanese Patent Application Publication No. 2004-62270. Summary of the Invention

[0009] The problem that the invention aims to solve

[0010] According to the prior art described in Patent Document 1, although it is possible to formulate an allocation plan by matching applicants with recruiters, this is only a simple matching method that provides assistance when the human resources department formulates a preliminary plan for job allocation. It has limitations in pursuing the appropriateness and effectiveness of the matching results.

[0011] To achieve more appropriate and effective matching in the talent market, recruiters need more comprehensive and diversified talent assessment and recommendation to assist them in making decisions. To achieve this goal, a mechanism is needed that diversifies the data input into the talent market, diversifies the assessment items, and then conducts standardized assessments and recommendations based on these diversifications.

[0012] Therefore, the purpose of this invention is to provide a talent market system and talent market program that assists recruiters in making decisions through more comprehensive and diversified talent assessment and recommendation based on diverse input data.

[0013] The present invention and other objects and novel features will become clear from the description and drawings herein.

[0014] Solution for solving the problem

[0015] The following is a brief description of the representative inventions disclosed in this application.

[0016] A representative embodiment of the talent market system of the present invention is a talent market system that assists recruiters and job seekers in matching. It comprises: an unstructured information processing unit, which extracts the resume and performance information of each candidate from unstructured information related to each candidate using generative AI, and outputs it as resume / performance data; an evaluation item setting unit, which extracts skill information as evaluation items for each candidate based on recruitment information using generative AI, and outputs it as skill definitions; and a recommendation processing unit, which selects recommended items for focused evaluation in relevant cases of the recruiter from the evaluation items of the skill definitions using generative AI, and evaluates each candidate based on the resume / performance data according to separately set evaluation criteria for each selected recommended item, and generates recommendation reasons for the recommended candidates using generative AI, which are output together with the evaluation results.

[0017] Furthermore, the present invention can also be applied to programs that enable a computer to operate as the aforementioned talent market system.

[0018] Invention Effects

[0019] The effects obtained by representative inventions in the inventions disclosed in this application are briefly described below.

[0020] That is, according to a representative embodiment of the present invention, a more comprehensive and diversified talent assessment and recommendation based on diverse input data can assist recruiters in making decisions. Attached Figure Description

[0021] Figure 1 This is a schematic diagram illustrating a structural example of a talent market system according to one embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram illustrating an example of talent assessment and recommendation in a talent market system according to one embodiment of the present invention.

[0023] Figure 3 This is a summary flowchart illustrating an example of a talent recommendation-related processing flow in one embodiment of the present invention. Detailed Implementation

[0024] Hereinafter, embodiments of the present invention will be described in detail based on the accompanying drawings. Furthermore, in all the drawings describing the embodiments, the same reference numerals are generally used for the same parts, and repeated descriptions are omitted. On the other hand, there are cases where parts described with reference numerals in a certain drawing may not be shown again in the description of other drawings, but are mentioned with the same reference numerals.

[0025] <Summary>

[0026] One embodiment of the present invention is a talent market (hereinafter referred to as "TMP") system, which is an information processing system, for example, used to match multiple internal open recruitments within an enterprise (including departments, units, groups, etc., which are recruiters, hereinafter collectively referred to as "recruiters") with applicants for these recruitments.

[0027] In this embodiment, during matching, in addition to structured information such as the applicant's position, qualifications, and skills, various unstructured information existing within the company that has not been previously utilized (including text, images, audio, and other information; in this embodiment, text information is used as an example) is also included as input data, thereby enabling flexible use of multiple types of input data. Unstructured information includes various types of text information, such as the applicant's resume, self-introduction text, and recruitment requirements related to matching and personnel, as well as business plans, proposals, reports, and chat logs written by the applicant in the course of business.

[0028] To handle this type of unstructured information (text information) input, this implementation utilizes generative AI (Artificial Intelligence), such as GhatGPT (registered trademark), which has become increasingly widely used in recent years. Based on unstructured information, it evaluates and screens potential talents who meet recruitment requirements through diverse projects. Furthermore, in addition to outputting quantitative assessments, it also outputs qualitative assessments such as reasons for recommendation in the form of articles. This more comprehensive and diversified talent evaluation and recommendation assists recruiters in making decisions.

[0029] Figure 2 This is a schematic diagram illustrating an example of talent assessment and recommendation in a talent marketplace (TMP) system according to one embodiment of the present invention. In this diagram, for the recruitment requirements of a specific recruiter, the applicant (talent) on the left side of the diagram first mechanically filters the candidate list of talents using structured information such as job title, qualifications, and skills, narrowing it down to a certain range.

[0030] Next, input the aforementioned talent's relevant resume and supervisor's evaluation, as well as various structured and unstructured information such as business proposals and reports written in the course of business, and structured and unstructured information such as the recruiter's recruitment requirements and recruitment data, to set up evaluation items for diversified talent assessment. The setting of relevant evaluation items can, for example, be manually completed by users such as the recruiter's manager or HR manager, but in this embodiment, it is automatically set by the generative AI2 using unstructured information. Its practicality and effectiveness can also be further improved by manually revising the evaluation items proposed by the generative AI2.

[0031] Subsequently, based on the content recorded in the structured and unstructured information, generative AI2 performs a 5-level assessment on the set evaluation items. Based on this assessment, promising talents are selected as candidates. For each candidate, in addition to the assessment itself, the rationale and basis for the assessment are also provided in the form of an article. For example, in the form of an "Assessment / Recommendation Explanation" on the right side of the image, promising talents are recommended to the recruiter.

[0032] <System Composition>

[0033] Figure 1 This is a schematic diagram illustrating an example structure of a talent marketplace (TMP) system according to one embodiment of the present invention. The TMP system 1 can be implemented, for example, via an information processing terminal such as a PC (personal computer) or a tablet terminal, or via server equipment or a virtual server built on cloud computing services.

[0034] Although it is depicted as a logical information processing system in the diagram, its physical functions can be distributed across multiple information processing terminals or servers.

[0035] TMP system 1, for example, executes middleware such as OS (operating system), DBMS (database management system), and web server programs loaded from recording devices such as HDD (hard disk drive) or SSD (solid-state drive) onto memory, as well as software running on these middleware, through a CPU (central processing unit) not shown, thereby realizing various functions related to TMP construction.

[0036] The TMP system 1, for example, has various parts implemented as software, such as a screening unit 10, an unstructured information processing unit 20, an evaluation item setting unit 30, and a recommendation processing unit 40. These parts can be implemented as a whole as an application, or they can be implemented as independent programs that cooperate with each other.

[0037] The screening unit 10 has the following functions: for the recruitment requirements of the target case, it inputs the talent list 11 of applicants obtained from personnel information and application information, performs pre-screening to narrow down the scope based on structured information such as position and qualifications, and outputs a list of candidates 12.

[0038] The Unstructured Information Processing Department 20 has the following functions: It inputs unstructured talent data 21, such as project proposals, reports, and other unstructured information (text) related to the candidate 12; automatically extracts resumes, achievements, and other content through generative AI 2; and organizes it into a specified format, outputting it as resume / achievement data 22. Automatically writing the resume / achievement data 22 through generative AI 2 not only saves applicants' time but also structures unstructured information to a certain extent, making it easier to utilize.

[0039] Furthermore, by pre-defining the skill definition 32 described below, and automatically extracting resume and performance information accordingly, it is possible to extract information in a format that meets the recruiter's needs. This helps solve the problem of applicants "not knowing what to write" when presenting their resume and performance during the application process, and the problem of recruiters being unable to make judgments based on "unfilled information" when evaluating applicants' resumes and performance.

[0040] The evaluation project setting department 30 has the following functions: it takes in recruitment data 31 from multiple recruiters, dynamically creates and outputs skill definitions 32, which cover the skills (evaluation items) required in each recruitment data 31. Skills required in a large amount of recruitment data 31 can be considered high-demand skills. For example, skill definitions 32 can be automatically created by generative AI2 based on recruitment objectives and content, by inputting recruitment data 31, and then manually adjusted and customized by the human resources department. Characteristics and positioning such as "skills that can be learned quickly" and "skills that require a longer time to learn" can also be set, or weights such as importance can be assigned. When creating skill definitions 32, existing skill definitions can also be used as initial values ​​to reflect existing measures or business policies.

[0041] The recommendation processing unit 40 has the following functions: It takes input skill definitions 32 and resume / performance data 22 related to each candidate 12, etc., and uses generative AI 2 to select an evaluation axis composed of skills that need to be focused on for analysis from the skills (evaluation items) in skill definitions 32, targeting the recruitment requirements relevant to the target case. Based on the selected evaluation axis, the generative AI 2 evaluates each candidate 12, and outputs promising talents as recommended talents 43 according to the evaluation results. When outputting recommended talents 43, for example, the generative AI 2 can generate the target talent's strengths / weaknesses, and suggestions on the support that the relevant department can provide to the talent, etc., and present them in a format such as... Figure 2 The "Evaluation / Recommendation Instructions" and other text formats shown are output to help recruiters understand and facilitate a smooth matching process.

[0042] When selecting evaluation axes, such as those objectively selected by generative AI2, recruiters can subjectively adjust and revise them based on business or case requirements. For each selected evaluation axis, recruiters can set evaluation criteria (i.e., what criteria to use to interpret talent data) based on the specific circumstances of the business or case. For example, a 5-level evaluation can be used, with the following criteria: "5: Deep knowledge and experience sufficient to become a leader; 4: Able to be immediately engaged in work; 3: Experienced and can be put into practice after training; 2: Possesses relevant knowledge; 1: Not very suitable."

[0043] When evaluating each candidate (talent 12), the generative AI2 objectively and meticulously reads a large amount of structured and unstructured information related to each talent, based on the evaluation axis and evaluation criteria. The evaluation results for each candidate (talent 12) yield a skills map (talent 42), which displays the distribution of the number of talents possessing each level of expertise across the various skills (evaluation items) defined in skills definition (32). This enables skills-based talent assessment.

[0044] When outputting recommended talent 43 based on the evaluation results of each candidate talent 12, not only are quantitative values ​​provided, but also an article with the reasons for recommendation (i.e. a summary of the candidate talent 12’s resume based on the evaluation axis) is added. This can help recruiters form a qualitative understanding when evaluating and judging recommended talent 43, and facilitate explanation to relevant parties (without having to obtain the original data again).

[0045] <Processing Flow>

[0046] Figure 3 The following is a summary flowchart illustrating an example of a talent recommendation process in one embodiment of the present invention. First, the screening unit 10 inputs information from a list of applicants 11 based on the recruitment requirements related to the target case, filters the applicants based on structured information such as job title and qualifications to narrow down the pool, and outputs candidate candidates 12 (step S01). Next, the unstructured information processing unit 20 takes project proposals, reports, and other unstructured talent data 21 related to candidate candidates 12 as input, automatically extracts resumes, achievements, etc., using generative AI2, and organizes them into a specified format, outputting them as resume / achievement data 22 (step S02).

[0047] Subsequently, the evaluation project setting unit 30 takes recruitment data 31 from multiple recruiters as input, and dynamically creates or updates skill definitions 32 through generative AI2 combined with manual operation and outputs them. The skill definitions 32 cover the skills required in each recruitment data 31 (evaluation project) (step S03).

[0048] Next, the recommendation processing unit 40, using skill definition 32 and relevant resume / performance data 22 of each candidate 12 as input, selects the key evaluation axes (recommended items) from the evaluation items of skill definition 32 that require in-depth analysis of the recruitment requirements related to the target case through generative AI2 combined with manual operation (step S04). Then, for each selected evaluation axis, the evaluation criteria for scoring are set (step S05), and based on each evaluation axis (recommended item), each candidate 12 is scored and evaluated through generative AI2 (step S06). Furthermore, for potential talents, landing assistance (comprehensive review) is provided (step S07), that is, generating recommendation reasons, the talent's strengths / weaknesses, and suggestions on the support that the relevant department can provide to the talent, and presenting this content in a format such as... Figure 2 The example outputs "Evaluation / Recommendation Description" in the form of "Recommended Talent 43", indicating the end of the process.

[0049] As described above, the TMP system 1 according to one embodiment of the present invention, by flexibly utilizing generative AI 2, can evaluate and select potential talents who meet recruitment requirements from diverse projects based on various unstructured information (text information). Furthermore, in addition to outputting quantitative assessments such as scores, it also outputs qualitative assessments such as reasons for recommendation in the form of articles. This more comprehensive and diversified talent assessment and recommendation assists recruiters in making decisions.

[0050] The invention described above is based on specific embodiments, but the invention is not limited to the above embodiments, and various modifications can be made without departing from its spirit. Furthermore, the above embodiments are detailed to facilitate understanding of the invention, but are not limited to having all the described components.

[0051] Furthermore, other components can be added, deleted, or replaced regarding a portion of the above-described embodiments.

[0052] Furthermore, some or all of the aforementioned components, functions, processing units, and processing methods can be implemented in hardware, such as through integrated circuit design. Alternatively, the aforementioned components and functions can also be implemented in software by a processor interpreting and executing programs that perform their respective functions. The programs, tables, files, and other information that implement these functions can be stored in storage devices such as memory, hard disks, and SSDs, or in recording media such as IC cards, SD cards, and DVDs.

[0053] Furthermore, the control lines and information lines shown in the above diagrams are for illustrative purposes only and may not represent all the control lines and information lines used in actual installation. In practice, it can be assumed that almost all components are interconnected.

[0054] Industrial availability

[0055] This invention can be applied to talent market systems and procedures for realizing talent markets.

[0056] Symbol Explanation

[0057] 1. Talent Market (TMP) System

[0058] 2 Generative AI

[0059] 10 Screening Department

[0060] 11 Talent List

[0061] 12 candidates

[0062] 20 Unstructured Information Processing Department

[0063] 21 Unstructured Talent Data

[0064] 22. Resume / Achievement Data

[0065] 30 Evaluation Project Setting Department

[0066] 31 Recruitment Data

[0067] 32 Skill Definitions

[0068] 40 Recommendation Processing Department

[0069] 42 Skill Chart

[0070] 43 Recommended Talents

Claims

1. A talent market system for assisting recruiters and job seekers in matching, comprising: The Unstructured Information Processing Department uses generative AI to extract the resumes and performance information of each candidate from the unstructured information related to each candidate, and outputs it as resume / performance data. The assessment project setting department, based on recruitment information, uses generative AI to extract skill information for the assessment projects of each talent and outputs it as a skill definition. as well as The recommendation processing department uses generative AI to select recommended items for focused evaluation in relevant cases of the recruiter from the evaluation items defined by the skills. For each selected recommended item, the department evaluates the candidate based on the resume / performance data according to the separately set evaluation criteria. Furthermore, for the recommended candidate, the department generates recommendation reasons using generative AI and outputs them together with the evaluation results.

2. The talent market system according to claim 1, wherein, The assessment project setting department accepts revised inputs for the skill definitions extracted through generative AI.

3. The talent market system according to claim 1, wherein, The recommendation processing unit accepts revision inputs for the recommended items selected by generative AI.

4. The talent market system according to claim 1, wherein, The recommendation processing unit generates information for the recommended talent using generative AI, including the talent's strengths / weaknesses and suggestions for the recruiter, and outputs this information along with the evaluation results.

5. The talent market system according to claim 1, wherein, It also has: The screening department pre-screens each applicant based on their structured information.

6. A talent market program that enables computers to operate as a talent market system to assist recruiters in matching job seekers with job applicants. The computer performs the following steps: Generative AI is used to extract the resumes and performance information of each candidate from unstructured information related to each candidate, and output them as resume / performance data. Based on recruitment information, skill information for the evaluation items of each talent is extracted using generative AI and output as a skill definition; Through generative AI, select recommended items for focused evaluation from the assessment items defined by the skills, based on cases relevant to the recruiter. For each of the selected recommended projects, the talents are evaluated using generative AI based on the resume / performance data according to the established evaluation criteria. For the recommended talents, generative AI is used to generate recommendation reasons, which are then output along with the evaluation results.

Images