Information processing device, information processing method, and program

Abstract

Regarding an information processing device, information processing method, and program for matching job postings with the most suitable job seeker information, the invention enables the search for the most suitable job seeker information for each recruitment medium and further enables the creation of job postings optimized for each recruitment medium. [Solution] The job seeker information search unit 104 performs a search on the formatted job seeker information for each recruitment medium obtained in the formatted job seeker information database 111 via the job seeker information acquisition unit 101, the job seeker information source database 110, and the job seeker information formatting unit 102, according to the search conditions acquired by the search condition acquisition unit 103 for the specified recruitment medium, thereby extracting job seeker information that appropriately matches the input job posting data.

Description

Technical Field

[0005]

[0001] The present invention relates to an information processing apparatus, an information processing method, and a program for matching optimal job hunting information to recruitment information.

Background Art

[0002] Conventionally, a technique for generating a recruitment ticket by referring to organizational information summary and further generating a scout sentence from the recruitment ticket is known (for example, Patent Document 1). Also conventionally, a technique for generating a recruitment ticket by referring to position information and further generating a scout sentence from the recruitment ticket is known (for example, Patent Document 2). Furthermore, conventionally, a technique for generating a scout sentence by receiving an input of a source sentence including profile information related to at least one of an organization (recruitment ticket) and an individual (job hunter) is known (for example, Patent Document 3).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Patent Document 2

Patent Document 3

Summary of the Invention

Problems to be Solved by the Invention

[0004] In an actual recruitment site, it is a very difficult task to search for job hunter information that matches a recruitment ticket, and it is also difficult to generate an optimal scout sentence if job hunter information that matches the recruitment ticket cannot be searched. However, with the above-mentioned conventional techniques, it has been difficult to search for job hunter information that matches a recruitment ticket. <​​Furthermore, when a company posting job openings on multiple recruitment platforms, it has traditionally been difficult to create and provide the most suitable job posting for each platform, and even more difficult to optimally match job postings with job seeker information for each platform.

[0006] Therefore, the present invention aims to enable the search for optimal job seeker information for each recruitment medium, and further to enable the creation of job postings optimized for each recruitment medium. [Means for solving the problem]

[0007] An example of an information processing device includes: a job seeker information acquisition unit that, for each job posting medium, acquires raw data of job seeker information from the website of that job posting medium via network or manual access and stores it in a job seeker information raw data database; a job seeker information formatting unit that, for each raw data of job seeker information stored in the job seeker information raw data database for each job posting medium, generates first query data including one or more first extraction instruction prompts that instruct the first extraction of one or more first metadata, and a first input prompt containing the text data of the raw data of the job seeker information, queries a first generative artificial intelligence system using the first query data to acquire first response text data in which the text data of the raw data of the job seeker information has been classified and formatted for each of the one or more first metadata, stores a new record of formatted job seeker information in a formatted job seeker information database, which has the classified and formatted text data for each of the first metadata in the first response text data as values ​​for each database field corresponding to each of the first metadata; and a job seeker information formatting unit that, for each job posting medium, acquires first response text data in which the text data of the raw data of the job seeker information has been classified and formatted for each of the first metadata, and stores a new record of formatted job seeker information in a formatted job seeker information database, which has the classified and formatted text data for each of the first metadata in the first response text data as values ​​for each database field corresponding to each of the first metadata; and one or more second The system includes: a search condition acquisition unit that generates second query data including one or more second extraction instruction prompts that instruct the second extraction of metadata, and a second input prompt containing the text data of the job posting data to be input, and queries a second generative artificial intelligence system with the second query data to acquire second response text data classified into one or more second metadata corresponding to one or more database fields in the formatted job seeker information database for each job posting medium, and a job seeker information retrieval unit that generates a search condition query for each job posting medium that specifies the text data corresponding to one or more second metadata in the second response text data acquired by the search condition acquisition unit as search keywords for the database fields in the formatted job seeker information database corresponding to the second metadata, extracts records of formatted job seeker information that match the search condition query from the formatted job seeker information database, and extracts the original data of job seeker information corresponding to the formatted job seeker information from the original job seeker information database.

[0008] Furthermore, an information processing device of another embodiment further comprises: a basic job posting data generation unit that generates a third query data including a third extraction instruction prompt in which a user belonging to the company specifies at least the job titles and company information of the company for which the company wishes to recruit, and queries a third generative artificial intelligence system using the third query data to generate basic job posting data for the company as a third response text data; a job posting data optimization unit that generates a fourth query data for each recruitment medium including a fourth extraction instruction prompt containing information indicating the recruitment medium and a fourth input prompt containing text data of the basic job posting data generated by the basic job posting data generation unit, and queries a fourth generative artificial intelligence system using the fourth query data to generate optimized job posting data optimized for the recruitment medium as a fourth response text data and stores it in an optimized job posting database; and a search condition acquisition unit makes the text data of the input job posting data that includes, for each recruitment medium, the text data of the optimized job posting data corresponding to the recruitment medium stored in the optimized job posting database, as a second input prompt. [Effects of the Invention]

[0009] According to the present invention, it becomes possible to search for the most suitable job seeker information for each recruitment medium, and furthermore, it becomes possible to create job postings optimized for each recruitment medium. [Brief explanation of the drawing]

[0010] [Figure 1] This is a functional block diagram of an embodiment of an information processing device. [Figure 2] This figure shows an example of text data from the raw data of job seeker information. [Figure 3] This figure shows an example of a first extraction instruction prompt. [Figure 4] This figure shows an example of the first response text data. [Figure 5] This figure shows an example of a second extraction instruction prompt. [Figure 6]It is a diagram showing an example of a search condition inquiry. [Figure 7] It is a diagram showing an example of the fourth extraction instruction prompt (first recruitment medium) (part 1). [Figure 8] It is a diagram showing an example of the fourth extraction instruction prompt (first recruitment medium) (continuation of part 1). [Figure 9] It is a diagram showing an example of the fourth extraction instruction prompt (second recruitment medium). [Figure 10] It is a diagram showing an example of an optimized recruitment ticket. [Figure 11] It is a diagram showing an example of the fifth extraction instruction prompt (part 1). [Figure 12] It is a diagram showing an example of the fifth extraction instruction prompt (part 2). [Figure 13] It is a diagram showing an example of the fifth response text data. [Figure 14] It is a diagram showing an example of the sixth extraction instruction prompt. [Figure 15] It is a hardware block diagram showing a configuration example of a computer that executes an embodiment of an information processing apparatus. [Figure 16] It is a flowchart showing an example of generation AI inquiry processing. [Figure 17] It is a flowchart showing an example of generation AI API access processing. [[ID=3l]]

Mode for Carrying Out the Invention

[0011] Hereinafter, embodiments for carrying out the present invention (hereinafter referred to as "the present embodiment") will be described in detail with reference to the drawings. FIG. 1 is a functional block diagram of an embodiment of an information processing apparatus, the information processing apparatus 100.

[0012] In FIG. 1, for example, a recruitment staff member of a company who operates the company terminal device 123 connected via the network 120 in FIG. 1 accesses the applicant information acquisition unit 101 from the company terminal device 123 and instructs the collection of applicant information. As a result, the job seeker information acquisition unit 101 acquires raw data of job seeker information from the job media website (indicated as "job media website" in the figure) 121 corresponding to the job media specified by the recruiting manager, via network 120 or manual access, and stores it in the job seeker information raw database (indicated as "job seeker information raw DB" in the figure) 110 shown in Figure 1. Specifically, as shown in Figure 1, a recruiter accesses job seeker information by manually logging in to one of several job posting websites 121 (#1 to #N, where N is a natural number greater than or equal to 2) via a network 120 such as a local area network or the internet from a corporate terminal device 123, and obtains the text data of the job seeker information from that job posting website 121 as the raw data of the job seeker information. Alternatively, if permitted, the job seeker information acquisition unit 101 may automatically access multiple job posting websites 121 numbered #1 to #N and perform a so-called web scraping process to acquire raw job seeker information data from each job posting website 121.

[0013] Figure 2 shows an example of a portion of the raw job seeker information data acquired by the job seeker information acquisition unit 101. The format of the raw job seeker information data is not standardized for each recruitment medium (recruitment medium website 121) and may vary.

[0014] The job seeker information acquisition unit 101 stores in the job seeker information source database 110 records in which the job seeker information acquired from each job posting website 121 is directly used as the value of, for example, one database field.

[0015] In Figure 1, the job seeker information formatting unit 102 generates first query data for each job posting medium and for each piece of raw data of job seeker information stored in the job seeker information raw database 110, including one or more first extraction instruction prompts that instruct the first extraction of one or more first metadata, and a first input prompt containing the text data of the raw data of the job seeker information, and queries the first generative artificial intelligence system, the generative AI system 122, using the first query data.

[0016] Figure 3 shows an example of a first extraction instruction prompt for generating the first query data. In this first prompt for extraction instructions, the "instructions" may include, for example, "Extract the given candidate data strings according to the 'Prisma model definition'," "Ensure the extracted data is in JSON format according to the 'return value type'," or "The candidate data is directly copied from the job posting site. It is not clean data, so determine the appropriate columns from the arrangement of the text and convert it to JSON." Furthermore, the first extraction instruction prompt includes "requests for extraction," such as "Please be sure to maintain data type during extraction," "Please return only JSON type data," "Please double-check to ensure no columns are missing or incorrect," "DateTime type must be ISO-8601," and "Never include null values ​​in array fields." Furthermore, the first extraction prompt specifies the "type of the return value (response)". In addition, the first extraction prompt specifies the format of the "candidate data."

[0017] Here, the generation AI system 122 is connected to, for example, the network 120, and the job seeker information formatting unit 102 accesses the application programming interface (hereinafter referred to as "API") function of the generation AI system 122 via the network 120. The API of the generation AI system 122 can be one or more of the following API services, or a combination thereof: ChatGPT-3.5, ChatGPT-4, or ChatGPT-4o ("GPT" is a trademark of OpenAI Opco LLC, Inc. in the United States), Gemini (a registered trademark of Google LLC, Inc. in the United States), Claude (a registered trademark of Anthropic, PBC, Inc. in the United States). Furthermore, the generation AI system 122 may be locally constructed within the organization where the information processing system 100 is operated and accessed via network 120, which is an internal organizational network such as a local area network.

[0018] Next, the job seeker information formatting unit 102 obtains first response text data from the generating AI system 122, in which the original text data of the job seeker information has been classified and formatted into one or more first metadata items.

[0019] Figure 4 shows an example of the first response text data that the generating AI system 122 responds to in response to a query using first query data including the first extraction instruction prompt illustrated in Figure 3. By specifying the first extraction instruction prompt, the data is formatted into pre-formatted job seeker information with unified columns, for example, in JSON format, for each job posting medium.

[0020] Furthermore, the job seeker information formatting unit 102 stores new records of formatted job seeker information in the formatted job seeker information database (indicated as "formatted job seeker information DB" in the figure) 111, which contains the text data that has been classified and formatted according to each of the first metadata in the first response text data obtained from the generation AI system 122, as values ​​for each database field corresponding to each of the first metadata. In this case, the new record may store information that links it to the corresponding job seeker information source data (Figure 2) record in the job seeker information source database 110 (for example, the record ID (identifier) ​​of the job seeker information source data record in the job seeker information source database 110).

[0021] As described above, in this embodiment, the raw job seeker information data for each recruitment medium obtained from the network 120 can be formatted into formatted job seeker information classified from each perspective corresponding to one or more first metadata specified by the first extraction instruction prompt, and stored in the formatted job seeker information database 111. In this embodiment, by utilizing the excellent text data formatting capabilities of the generation AI system 122, it becomes possible to convert raw job seeker information data into uniformly formatted job seeker information, regardless of the recruitment medium.

[0022] In Figure 1, the recruiter specifies the recruitment medium from the company terminal device 123 and instructs it to search for job seeker information. As a result, the search condition acquisition unit 103 first generates second query data which includes one or more second extraction instruction prompts that instruct the second extraction of one or more second metadata for the specified job posting media, and a second input prompt which includes pre-created and input text data of the job posting data, and then queries the second generative artificial intelligence system, the generative AI system 122, using this second query data.

[0023] Figure 5 shows an example of a second extraction instruction prompt for generating second query data. In this second prompt for extraction instructions, for example, the "thought process when setting search criteria" is specified, such as "read the job posting and identify the required conditions that are explicitly stated," and "consider other conditions that the hiring manager might set." The "strictness of the search criteria" is also specified (for example, the more the population of the occupation and the greater the demand for the job, the stricter the search criteria will be for job postings with high demands or high salary levels, and the looser the search criteria will be for job postings with low demands or low salary levels). Furthermore, the second prompt for extraction instructions specifies "requests" such as "Please include any required conditions such as age or job type in the job posting as search criteria" and "Please be sure to adhere to the commented format in JSON format." Then, in the second extraction instruction prompt, search terms and filtering conditions on the direct recruiting platform are specified, such as "job posting information," "search conditions and specific format," and "response JSON format and example values."

[0024] Next, the search condition acquisition unit 103 acquires from the generation AI system 122, for each job posting medium, second response text data classified into one or more second metadata entries, each corresponding to one or more database fields in the formatted job seeker information database 111, where the text data of the job posting is formatted.

[0025] Next, in Figure 1, the job seeker information search unit 104 generates a search condition query that specifies, for the specified job posting medium, the text data corresponding to one or more second metadata in the second response text data obtained by the search condition acquisition unit 103 as search keywords for the database fields in the formatted job seeker information database 111 that correspond to those second metadata. It then extracts records of formatted job seeker information that match the search condition query from the formatted job seeker information database 111 and extracts the original data of the job seeker information corresponding to that formatted job seeker information from the job seeker information source database 110.

[0026] Figure 6 shows an example of a search condition query generated by the job seeker information retrieval unit 104 based on the second response text data obtained from the AI ​​system 122 for each job posting medium by the search condition acquisition unit 103. For example, if a recruiter instructs a search for job seeker information on job board 1, which is the job board website 121(#1) shown in Figure 1, the search conditions exemplified in Figure 6(a) will be displayed on the display of the corporate terminal device 123 shown in Figure 1, which is operated by the recruiter, for the formatted job seeker information in the formatted job seeker information database 111 for that job board 1. This allows the recruiter to use the search form exemplified in Figure 6(a) to search for job seeker information on job board 1, which has been collected as raw job seeker information in the raw job seeker information database 110, formatted as formatted job seeker information in the formatted job seeker information database 111, and collected. Here, the "keywords," "exclusion keywords," "job type," "region," "new graduate filter," and "age" shown as an example in Figure 6(a) correspond to one or more second metadata entries in the second response text data corresponding to the job posting platform 1, and each database field of the record corresponding to the job posting platform 1 in the formatted job seeker information database 111 also corresponds to each of these second metadata entries. Therefore, recruiters can accurately search for job seeker information corresponding to the job posting platform 1 by making search condition queries using these search keywords. Similar to the case in Figure 6(a), when a recruiter instructs a search for job seeker information on the job media website 121(#2) in Figure 1, the search conditions exemplified in Figure 6(b) are displayed on the display of the corporate terminal device 123 for the formatted job seeker information in the formatted job seeker information database 111 for that job media 2. As a result, the recruiter uses the search form exemplified in Figure 6(b) to search for job seeker information on the job media 2, which has been collected as raw job seeker information in the raw job seeker information database 110, formatted as formatted job seeker information in the formatted job seeker information database 111, and collected. Here, the "job seekers with high response rates," "job seekers of interest to your company," "age," "annual income," "highest level of education," "desired industry," and "desired job type," as exemplified in Figure 6(b), correspond to one or more second metadata entries in the second response text data corresponding to the job posting platform 2. Each database field in the records corresponding to the job posting platform 2 in the formatted job seeker information database 111 also corresponds to each of these second metadata entries. Therefore, recruiters can accurately search for job seeker information corresponding to the job posting platform 2 by making search condition queries using these search keywords.

[0027] As described above, according to this embodiment, even if job seeker information is spread across multiple job posting platforms, it becomes possible to accurately search for job seeker information for each job posting platform.

[0028] Next, the operation of generating job posting data will be explained. In Figure 1, for example, when a recruiter operating a corporate terminal device 123 instructs the creation of job posting data for their company, the basic job posting data generation unit 105 first generates a third query data that includes a third extraction instruction prompt (not shown) which specifies, for example, at least the job type and company information of the company for which the company wishes to recruit.

[0029] Then, the basic job posting data generation unit 105 queries the generation AI system 122 using the third query data, and generates the company's basic job posting data as the third response text data.

[0030] Here, the basic job posting data generation unit 105 may further include a third input prompt in the third query data in which the user specifies job posting data of competitors for their own company, and query the third generative artificial intelligence system, the generative AI system 122, using the third query data. This makes it possible to generate basic job posting data that includes the content of the company's job postings that are superior to those of competitors.

[0031] Next, in Figure 1, the job posting data optimization unit 106 generates a fourth query data which includes a fourth extraction instruction prompt containing information indicating the job posting medium for each job posting medium handled by the company, and a fourth input prompt containing text data of the basic job posting data generated by the basic job posting data generation unit 105.

[0032] Figures 7 and 8 show examples of a fourth extraction instruction prompt, which may be specified for, for example, a first job posting platform. As shown in Figure 7, the content of the prompt is specified in detail for each item. In Figure 8, the specific instruction content is omitted from the illustration. On the other hand, Figure 9 shows an example of a fourth extraction instruction prompt specified for, for example, a second job posting platform. As can be seen by comparing Figure 9 with Figures 7 and 8, the prompt may be specified with different items for each job posting platform.

[0033] Next, the job posting data optimization unit 106 queries the generation AI system 122, which is the fourth generative artificial intelligence system, using a fourth query data that includes a fourth extraction instruction prompt, which includes a fourth extraction instruction prompt generated for each recruitment medium as described above, and a fourth input prompt, which includes basic job posting data generated by the basic job posting data generation unit 105. As a result, it generates optimized job posting data optimized for each recruitment medium as the fourth response text data, and stores it in the optimized job posting database (indicated as "Optimized Job Posting DB" in the figure) 112 shown in Figure 1.

[0034] In this way, by individually specifying a fourth extraction instruction prompt for each recruitment medium, the recruitment data optimization unit 106 optimizes the basic recruitment data generated by the basic recruitment data generation unit 105 to match each recruitment medium, and the optimized recruitment data is then stored in the optimized recruitment database 112.

[0035] In the above-described process, the job posting data optimization unit 106 displays, for example, the information shown in Figure 10 on the display of the corporate terminal device 123 in Figure 1. In this display, if a recruiter checks the box for the first recruitment medium, as shown in Figure 10(a), the screen can display job postings optimized for that first recruitment medium. Also, if a recruiter checks the box for the second recruitment medium, as shown in Figure 10(b), the screen can display job postings optimized for that second recruitment medium.

[0036] The aforementioned search condition acquisition unit 103 incorporates the text data of the optimized job posting data for each job posting medium obtained from the optimized job posting database 112 into the second input prompt, which is then used as the text data of the input job posting data. This allows the search condition acquisition unit 103 to specify to the generation AI system 122 that it should generate search conditions optimized for each job posting medium.

[0037] In Figure 1, the job seeker information score calculation unit 107 generates a fifth query data for each job posting medium, which includes a fifth extraction instruction prompt, a fifth input prompt including the input job posting data or second response text data acquired by the search condition acquisition unit 103, and the raw data or formatted job seeker information text data extracted by the job seeker information search unit 104. By querying the fifth generative artificial intelligence system, which is the generative AI system 122, with this fifth query data, the system may calculate score information as the fifth response text data, which includes the degree of match between the input job posting data or second response text data acquired by the search condition acquisition unit 103 and the raw data or formatted job seeker information text data extracted by the job seeker information search unit 104, as well as text data indicating the reason for calculating that degree of match.

[0038] Figures 11 and 12 show examples of the fifth extraction instruction prompt. In this fifth extraction instruction prompt, for example, for each job posting medium, as illustrated in Figure 11, the text data of the optimized job posting obtained in the optimized job posting database 112 corresponding to that job posting medium is specified as the "Job Posting" item, and the text data of the work history extracted by the job seeker information search unit 104 from the formatted job seeker information is specified as the "Work History" item. Furthermore, as illustrated in Figure 11, the fifth prompt for extraction instructions specifies a specific response format, along with instructions such as, "Compare the above job posting with multiple work histories, determine how well each work history matches the job posting, and return a response in the following format," specifying the display of a score indicating the degree of matching between the job posting and the work history, and the reason for that matching score (approximately 30 characters). Furthermore, as shown in Figure 11, the fifth extraction instruction prompt includes instructions regarding "[Criteria for determining the degree of match]," such as "Job type / specialization (importance: 5)," "Skills (importance: 5)," and "Education / major (importance: 5)." In addition, as shown in Figure 12, the fifth extraction prompt includes instructions such as "How to write the reason for the match" and "Things that must be followed."

[0039] Figure 13 shows an example of the fifth response text data obtained from the generating AI system 122 based on the fifth extraction instruction prompt illustrated in Figures 11 and 12. For each work history where the "resume id" is indicated by numbers from 1 to 5, the match score is indicated as "match_score" with a maximum score of 100, and the reason for the match is obtained in Japanese text in the "reasons" field.

[0040] Along with the search results for job seeker information for each recruitment medium by the job seeker information search unit 104, the aforementioned score (match degree) and reason for matching are displayed on the display of, for example, the corporate terminal device 123, enabling recruiters to make accurate hiring decisions for each job seeker.

[0041] In Figure 1, the scout message generation unit 109 generates a sixth query data for each recruitment medium, which includes a sixth extraction instruction prompt, a sixth input prompt that includes the input job posting data or second response text data acquired by the search condition acquisition unit 103, and the raw data of job seeker information or formatted job seeker information text data extracted by the job seeker information search unit 104. The unit then queries the sixth generative artificial intelligence system, the generative AI system 122, using this sixth query data, thereby generating a scout message as the sixth response text data.

[0042] Figure 14 shows an example of a sixth extraction instruction prompt. This sixth prompt for extraction instructions allows you to specify, for example, for each searched job seeker information, the instructions shown in the "Instructions" field, the instructions shown in the "Thought Process" field, as well as "Notes," "Requests when creating," "Information to include in the scout message to make it more appealing," "Reasons for sending the scout message," "About the email title," and "Output format" (output is JSON). The format, etc., is specified, along with "information to be used for scouting" (such as the job seeker's resume and job posting information).

[0043] By accessing the generation AI system 122 via a sixth extraction instruction prompt as shown in Figure 14, attractive recruitment messages can be generated for each job seeker.

[0044] In Figure 1, for example, based on instructions from a recruiter operating a corporate terminal device 123, the scout message distribution unit 108 distributes scout messages, for example, those generated by the scout message generation unit 109 in Figure 1, to job seekers determined by the recruiter based on the search results of the job seeker information search unit 104 or the score information calculated by the job seeker information score calculation unit 107, for each recruitment medium. In this way, it becomes possible to deliver the right recruitment message to the right job seeker across multiple job boards.

[0045] In the embodiments described above, the generative AI system 122 in Figure 1 is commonly used as the first to sixth generative artificial intelligence systems mentioned above. However, these may be different generative artificial intelligence systems such as ChatGPT-3.5, ChatGPT-4, ChatGPT-4o, Gemini, or Claude, or these systems may be applied in combination.

[0046] Figure 15 is a hardware block diagram showing an example of a computer configuration that performs the functions of the information processing device 100 in Figure 1. This hardware comprises a server computer configuration in which a CPU (Central Processing Unit) 1501, a memory 1502 where programs are loaded and executed, an external storage device 1503 that stores programs, databases such as the job seeker information source database 110, the formatted job seeker information database 111, and the optimized job posting database 112 in Figure 1, and other data, and a network interface circuit 1504 that controls access to the network 120 in Figure 1 are interconnected by a system bus 1505. Although not specifically shown, input devices such as a keyboard and mouse, and output devices such as a display may also be connected.

[0047] Figure 16 is a flowchart illustrating an example of a generation AI query processing that controls queries to the generation AI system 122 in the job seeker information formatting unit 102, search condition acquisition unit 103, basic job posting data generation unit 105, job posting data optimization unit 106, job seeker information score calculation unit 107, and scout message generation unit 109 of Figure 1. This processing is initiated from the processing of each function of the job seeker information formatting unit 102, search condition acquisition unit 103, basic job posting data generation unit 105, job posting data optimization unit 106, job seeker information score calculation unit 107, and scout message generation unit 109 of Figure 1. This flowchart shows the process of executing a generation AI query processing program written in a computer programming language such as Python ("Python" is a registered trademark of the Python Software Foundation of the United States) that the CPU 1501 of Figure 10 has loaded from the external storage device 1503 into memory 1502.

[0048] First, CPU1501 configures the library and API authentication information for accessing the generated AI API (step S1601). This library is, for example, the ChatGPT API library provided by Python. This library is pre-installed on, for example, the external storage device 1503 shown in Figure 15. In step S1601, the CPU 1501 executes a process to import the above library from the external storage device 1503 shown in Figure 15 to memory 1502, for example, by executing the "import openai" command. Furthermore, in step S1601, the CPU 1501 executes an instruction at the beginning of the program to set API authentication information data that has been previously registered with, for example, the ChatGPT service.

[0049] Next, the CPU 1501 sets the text data for one of the first to sixth extraction instruction prompts mentioned above (step S1602).

[0050] Next, the CPU 1501 sets the text data for one of the first to sixth input prompts mentioned above (step S1603).

[0051] Then, the CPU 901 generates query data that includes the extraction instruction prompt set in step S1602, the input prompt set in step S1603, and other necessary text data (step S1604).

[0052] Then, CPU 1501 calls the program for generating AI API access processing, using the query data generated in step S1604 as an argument (step S1605). This process will be described later using the flowchart in Figure 17.

[0053] Finally, CPU 1501 processes the response text data returned from the generation AI API processing and terminates the generation AI query processing shown in the flowchart of Figure 16 (step S1606).

[0054] Figure 17 is a flowchart showing an example of the processing of a generative AI API access program, which is called as a subroutine from step S1605 in Figure 16. Similar to Figure 16, this flowchart also shows the process of executing a generative AI API access program, created, for example, by the Python computer programming language, which is loaded from the external storage device 1503 to memory 1502 by the CPU 1501 in Figure 15.

[0055] In Figure 17, the CPU 1501 first sets the generative AI model (step S1701). This is an instruction to specify, for example, the aforementioned predetermined version of ChatGPT or another generative artificial intelligence system as the generative AI model.

[0056] Next, the CPU 1501 sets the query data passed by the subroutine call from the processing in step S1605 of Figure 16 (see steps S1602 to S1604 of Figure 16) into a predetermined variable in memory 1502 (step S1702).

[0057] Then, the CPU 1501 sets other control parameters for the ChatGPT API, for example, to other predetermined variables in memory 1502 (step S1703).

[0058] After the processing in steps S1701 to S1703 described above, the CPU 1501 calls the response setting function in the API library read in step S1601 of Figure 16 (step S1704), using the text data or parameter set in the variables in steps S1701 to S1703 as arguments.

[0059] Next, CPU1501 calls the response retrieval function of the API library (step S1705).

[0060] Finally, the CPU 1501 returns the response from the generating AI system 122 (Figure 1, e.g., the ChatGPT system) obtained by the call in step S1705 as the first response text data or the second response text data mentioned above, and returns to the processing of the flowchart in Figure 11 or Figure 13 described later (step S1706).

[0061] As described above, according to this embodiment, not only is formatting processing performed on job seeker information data possible, but appropriate search conditions are also extracted for each recruitment medium for job posting data, and search processing is performed on the original job seeker information database 110. In this case, the advanced formatting function and search condition extraction function of the generation AI system 122 can be utilized. High-precision matching of job posting data and job seeker information can be performed across multiple recruitment media, making it possible to provide job seeker information desired by company recruiters and others. [Explanation of symbols]

[0062] 100 Information Processing Devices 101 Job applicant information acquisition department 102 Job Applicant Information Department 103 Search Criteria Acquisition Unit 104 Job Seeker Information Search Department 105 Basic Job Posting Data Generation Department 106 Job Posting Data Optimization Department 107 Job Seeker Information Score Calculation Unit 108 Scout Message Distribution Department 109 Scout Message Generation Unit 110 Job Seeker Information Database 111 Processed Job Seeker Information Database 112 Optimized Job Posting Database 120 Internet 121 Job board websites 122 Generating AI System 123 Enterprise terminal equipment

Claims

1. For each job posting platform, a job seeker information acquisition unit acquires raw data of job seeker information from the website of the job posting platform via network or manual access and stores it in a job seeker information database. A job seeker information formatting unit generates first query data for each job seeker information source data stored in the job seeker information source database, which includes one or more first extraction instruction prompts that instruct the first extraction of one or more first metadata for each job seeker information source data stored in the job seeker information source database, and a first input prompt that includes the text data of the job seeker information source data, and queries a first generative artificial intelligence system with the first query data to obtain first response text data in which the text data of the job seeker information source data has been classified and formatted for each of the one or more first metadata, and stores a new record of formatted job seeker information in the formatted job seeker information database, in which the text data classified and formatted for each of the first metadata in the first response text data is the value of each database field corresponding to each of the first metadata, A search condition acquisition unit generates second query data for each of the job posting media, which includes one or more second extraction instruction prompts that instruct the second extraction of one or more second metadata, and a second input prompt that includes text data of the job posting data to be input, and queries a second generative artificial intelligence system using the second query data to acquire second response text data, in which the text data of the job posting data is classified into one or more of the second metadata corresponding to one or more of the database fields of the formatted job seeker information database. A job seeker information search unit generates a search condition query for each of the job posting media, specifying the text data corresponding to one or more of the second metadata in the second response text data obtained by the search condition acquisition unit as a search keyword for the database field in the formatted job seeker information database corresponding to the second metadata, extracts records of the formatted job seeker information that match the search condition query from the formatted job seeker information database, and extracts the original data of the job seeker information corresponding to the formatted job seeker information from the original job seeker information database. An information processing device equipped with the following features.

2. A basic job posting data generation unit generates a third query data that includes a third extraction instruction prompt specifying at least the job titles and company information of the company in which the company wishes to recruit, and generates the company's basic job posting data as a third response text data by querying a third generative artificial intelligence system using the third query data. A job posting data optimization unit generates a fourth query data for each of the aforementioned job posting media, which includes a fourth extraction instruction prompt containing information indicating the job posting media, and a fourth input prompt containing text data of the basic job posting data generated by the basic job posting data generation unit, and queries a fourth generative artificial intelligence system using the fourth query data to generate optimized job posting data optimized for the job posting media as a fourth response text data, and stores it in an optimized job posting database. Furthermore, The search condition acquisition unit, for each job posting medium, uses the text data of the optimized job posting data corresponding to that job posting medium, stored in the optimized job posting database, as the text data of the job posting data to be input, which is included as the second input prompt. The information processing apparatus according to claim 1.

3. The information processing apparatus according to claim 2, wherein the basic job posting data generation unit further includes in the third query data a third input prompt in which the user specifies job posting data of a competitor to the company, and generates the basic job posting data which includes the content of the company's job postings that are superior to those of the competitors, by querying the third generative artificial intelligence system with the third query data.

4. The information processing apparatus according to claim 1, further comprising: a job seeker information score calculation unit that generates a fifth query data for each of the job posting media, including a fifth extraction instruction prompt, a fifth input prompt including the input job posting data or the second response text data acquired by the search condition acquisition unit, and the raw data of the job seeker information or the formatted job seeker information text data extracted by the job seeker information search unit; and a job seeker information score calculation unit that calculates score information as a fifth response text data, including the degree of match between the input job posting data or the second response text data acquired by the search condition acquisition unit and the raw data of the job seeker information or the formatted job seeker information text data extracted by the job seeker information search unit, and text data indicating the reason for calculating the degree of match.

5. The information processing apparatus according to claim 1 or 4, further comprising a scout message distribution unit that distributes scout messages to each of the aforementioned job posting media to job seeker information determined based on the search results of the job seeker information search unit described in claim 1, or to job seeker information determined based on the score information calculated by the job seeker information score calculation unit described in claim 4.

6. The information processing apparatus according to claim 5, further comprising: a sixth query data generation unit that generates a sixth query data for each of the aforementioned job posting media, the sixth query data generation unit that generates the scout text as the sixth response text data by querying a sixth generative artificial intelligence system using the sixth query data. The scout text generation unit further comprises a sixth query data generation unit that generates the scout text as the sixth response text data.

7. The information processing apparatus according to claim 1, 2, 3, 4, or 6, wherein the first generative artificial intelligence system and the second generative artificial intelligence system according to claim 1, the third generative artificial intelligence system and the fourth generative artificial intelligence system according to claim 2 or 3, the fifth generative artificial intelligence system according to claim 4, and the sixth generative artificial intelligence system according to claim 6 are identical or different generative artificial intelligence systems.

8. For each job posting platform, the process involves obtaining raw data of job seekers from the website of the job posting platform via network or manual access and storing it in a job seeker information database. For each of the job posting media, for each of the raw data of the job seeker information stored in the raw data database of the job seeker information, a first query data is generated which includes one or more first extraction instruction prompts that instruct the first extraction of one or more first metadata, and a first input prompt which includes the text data of the raw data of the job seeker information; a query is made to a first generative artificial intelligence system using the first query data to obtain first response text data in which the text data of the raw data of the job seeker information has been classified and formatted according to one or more of the first metadata; and a new record of formatted job seeker information is stored in the formatted job seeker information database which has the classified and formatted text data of the first metadata in the first response text data as values ​​for each database field corresponding to each of the first metadata; A search condition acquisition process that generates second query data for each job posting medium, including one or more second extraction instruction prompts that instruct the second extraction of one or more second metadata, and a second input prompt containing the text data of the job posting data to be input, and queries a second generative artificial intelligence system using the second query data to acquire second response text data classified into one or more of the second metadata corresponding to one or more of the database fields of the formatted job seeker information database, For each of the aforementioned job posting media, a job seeker information search process is performed which generates a search condition query that specifies, for each of the second metadata entries in the second response text data obtained in the search condition acquisition process, the text data corresponding to each of the second metadata entries in the formatted job seeker information database as a search keyword for the database field in the formatted job seeker information database corresponding to the second metadata entry, extracts records of the formatted job seeker information that match the search condition query from the formatted job seeker information database, and extracts the original data of the job seeker information corresponding to the formatted job seeker information from the original job seeker information database. An information processing method for executing this.

9. A basic job posting data generation process generates basic job posting data for a company by having a user belonging to the company generate a third query data that includes a third extraction instruction prompt specifying at least the job titles and company information of the company for which the company wishes to recruit, and by querying a third generative artificial intelligence system with the third query data, thereby generating basic job posting data for the company as third response text data. For each of the aforementioned job posting media, a fourth query data is generated, which includes a fourth extraction instruction prompt containing information indicating the job posting media, and a fourth input prompt containing text data of the basic job posting data generated by the basic job posting data generation unit. By querying a fourth generative artificial intelligence system using the fourth query data, optimized job posting data optimized for the job posting media is generated as fourth response text data, and stored in the optimized job posting database. Further execution, For each of the aforementioned job posting media, the text data of the optimized job posting data corresponding to that job posting media, which is stored in the optimized job posting database, is included as the text data of the input job posting data to be included as the second input prompt, and the search condition acquisition process is executed accordingly. The information processing method according to claim 8.

10. For each job posting platform, the process involves obtaining raw data of job seekers from the website of the job posting platform via network or manual access and storing it in a job seeker information database. For each of the job posting media, for each of the raw data of the job seeker information stored in the raw data database of the job seeker information, a first query data is generated which includes one or more first extraction instruction prompts that instruct the first extraction of one or more first metadata, and a first input prompt which includes the text data of the raw data of the job seeker information; a query is made to a first generative artificial intelligence system using the first query data to obtain first response text data in which the text data of the raw data of the job seeker information has been classified and formatted according to one or more of the first metadata; and a new record of formatted job seeker information is stored in the formatted job seeker information database which has the classified and formatted text data of the first metadata in the first response text data as values ​​for each database field corresponding to each of the first metadata; A search condition acquisition process that generates second query data for each job posting medium, including one or more second extraction instruction prompts that instruct the second extraction of one or more second metadata, and a second input prompt containing the text data of the job posting data to be input, and queries a second generative artificial intelligence system using the second query data to acquire second response text data classified into one or more of the second metadata corresponding to one or more of the database fields of the formatted job seeker information database, For each of the aforementioned job posting media, a job seeker information search process is performed which generates a search condition query that specifies, for each of the second metadata entries in the second response text data obtained in the search condition acquisition process, the text data corresponding to each of the second metadata entries in the formatted job seeker information database as a search keyword for the database field in the formatted job seeker information database corresponding to the second metadata entry, extracts records of the formatted job seeker information that match the search condition query from the formatted job seeker information database, and extracts the original data of the job seeker information corresponding to the formatted job seeker information from the original job seeker information database. A program that causes a computer to execute something.

11. A basic job posting data generation process generates basic job posting data for a company by having a user belonging to the company generate a third query data that includes a third extraction instruction prompt specifying at least the job titles and company information of the company for which the company wishes to recruit, and by querying a third generative artificial intelligence system with the third query data, thereby generating basic job posting data for the company as third response text data. For each of the aforementioned job posting media, a fourth query data is generated, which includes a fourth extraction instruction prompt containing information indicating the job posting media, and a fourth input prompt containing text data of the basic job posting data generated by the basic job posting data generation unit. By querying a fourth generative artificial intelligence system using the fourth query data, optimized job posting data optimized for the job posting media is generated as fourth response text data, and stored in the optimized job posting database. Further execution, For each of the aforementioned job posting media, the text data of the optimized job posting data corresponding to that job posting media, which is stored in the optimized job posting database, is included as the text data of the input job posting data to be included as the second input prompt, and the search condition acquisition process is executed accordingly. The program according to claim 10.

Images