Analysis system and analysis apparatus

Abstract

To provide an analysis system and an analysis apparatus, based on the analysis results of access status to a website, which can obtain proposal information on effective measures to improve the access status to the website.SOLUTION: An analysis apparatus comprises: a prompt generation unit 11 for generating a prompt including intrusions for suggesting ways to improve access status and reference information using information describing the contents of a website as an analysis target, report information of analysis results related to access status to the target site, information describing the contents of one or more services utilizable for the improvement of the access status and a template for prompt generation; and an analysis unit 12 for acquiring proposal information on measures to improve the access status to the target site from an LLM 30 utilizing services by inputting the generated prompt into the LLM 30. With this configuration, the proposal information of the measures in consideration of the site content information and the service content information can be obtained.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an analysis system and an analysis device, and is particularly suitable for use in a system and device that generates proposal information for measures effective in improving access to a website. [Background technology]

[0002] Conventionally, there is known a technology for analyzing access history to a website and proposing effective measures based on the analysis results (see, for example, Patent Document 1). In the information processing device described in Patent Document 1, access log data of a user's webpage browsing to a machine-learned cluster estimation model is inputted to obtain an estimated value of characteristics related to the user's consumption behavior, and the estimated value is inputted to a machine-learned measure estimation model to output actions for sales promotion and the like that are estimated to be effective for the user.

[0003] Actions for sales promotion include, for example, determining products to be proposed, a method of sales promotion, determining a proposed purchasing method, and determining a proposed mode of use. Methods of sales promotion include, for example, displaying advertisements on a web page, telephone, direct mail, e-mail, visiting, etc. Purchasing methods include, for example, lump-sum cash payment, loan, lease, etc. Modes of use include, for example, purchase, lease, share, etc.

[0004] The cluster estimation model is trained using integrated data created from the web browsing access log data of the target user, the site classification master, and the survey response data on the values ​​and consumption behavior of the target user as training data.The policy estimation model is trained using integrated data created from the survey response data on the values ​​and consumption behavior and the survey response data on policy evaluation as training data.

[0005] According to the technology described in Patent Document 1, it is possible to propose effective actions according to the access status of each user to a website. However, the technology described in Patent Document 1 proposes actions related to sales promotion to be implemented for each individual user, and cannot propose effective measures for improving the access status of a website.

[0006] There is also a technology that evaluates the effectiveness of a certain measure by comparing the access history before and after the measure is implemented on a website. However, this type of technology can only verify the effectiveness of the measure after it is implemented, and cannot suggest effective measures in advance based on the access status of the website. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] Patent Publication No. 2022-96488 Summary of the Invention [Problem to be solved by the invention]

[0008] The present invention has been made to solve such problems, and has an object to make it possible to obtain, based on the results of an analysis of access conditions to a website, proposal information for measures that are effective in improving access conditions to the website. [Means for solving the problem]

[0009] In order to solve the above-mentioned problems, in the present invention, a prompt is generated using site content information representing the contents of a target site, which is a website to be analyzed, report information of the analysis results regarding the access status of the target site, service content information representing the contents of one or more services that can be used to improve the access status, and a template for generating a prompt, the prompt including an instruction statement instructing the user to propose measures to improve the access status and reference information generated from the site content information, report information, and service content information, and the generated prompt is input into a large-scale language model, so that proposed information for measures that utilize services to improve the access status of the target site is obtained from the large-scale language model. Effect of the Invention

[0010] According to the present invention configured as described above, by providing the site content information and the service content information as reference information to the large-scale language model together with the report information of the analysis result regarding the access situation of the target site, it is possible to generate proposal information of measures analyzed using the site content information and the service content information as information sources as measures for improving the access situation indicated in the report information. In this way, proposal information of measures effective for improving the access situation of the website can be obtained based on the analysis result of the access situation of the website. [Brief description of the drawings]

[0011] [Figure 1] 1 is a diagram illustrating an example of the overall configuration of an analysis system according to an embodiment of the present invention. [Diagram 2] 2 is a block diagram showing an example of the configuration of an analysis device and an information storage unit according to the present embodiment. FIG. [Diagram 3] 1 is a block diagram showing an example of a functional configuration of an analysis device according to an embodiment of the present invention, which is related to automatic generation of site content information; [Figure 4] FIG. 10 is a diagram illustrating an example of association information between types of access situations and available services. [Diagram 5]FIG. 2 is a diagram illustrating an example of a generating template according to the present embodiment. [Figure 6] FIG. 13 is a block diagram showing an example of the configuration of an analysis device and an information storage unit according to a first modified example. [Figure 7] FIG. 10 is a diagram illustrating an example of business-specific service definition information. [Figure 8] FIG. 10 is a diagram illustrating another example of business-specific service definition information. [Figure 9] FIG. 13 is a block diagram showing an example of the configuration of an analysis device and an information storage unit according to a second modified example. [Figure 10] FIG. 13 is a block diagram showing an example of the configuration of an analysis device and an information storage unit according to a third modified example. [Figure 11A] FIG. 13 is a diagram illustrating an example of a generating template according to a third modified example. [Figure 11B] FIG. 13 is a diagram illustrating an example of a generating template according to a third modified example. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing an example of the overall configuration of an analysis system according to this embodiment. As shown in FIG. 1, the analysis system according to this embodiment is configured to include an analysis device 10, an information storage unit 20, and large-scale language models (LLM: Large Language Models) 30. The analysis device 10 and the LLM 30 are connected via a communication network 100 such as the Internet or a mobile phone network. The analysis device 10 and the information storage unit 20 are connected by, for example, a LAN (Local Area Network). The analysis device 10 and the information storage unit 20 may be connected by the communication network 100.

[0013] The analysis device 10 uses the information stored in the information storage unit 20 to cooperate with the LLM 30 to generate and output proposed information for measures to improve the access status of the website being analyzed (hereinafter referred to as the target site).

[0014] The LLM30 is a natural language processing model trained using a large amount of text data, and takes a sentence as input and outputs a sentence. The LLM30 of this embodiment inputs an instruction sentence (which may be in the form of a question sentence) that instructs the creation of an answer sentence, and generates and outputs an answer sentence according to the instructions indicated by the instruction sentence. In the LLM30, the sentence that is input is called a "prompt."

[0015] Generally, compared to inputting only a simple question sentence into the LLM 30, adding reference information to a question sentence and inputting the question sentence into the LLM 30 makes it easier for the LLM 30 to obtain a more appropriate answer sentence. This is because, when there is additional information, the LLM 30 generates an answer sentence to the question sentence by referring to the additional information. In order to make use of this feature of the LLM 30 in the analysis device 10 of this embodiment, a prompt that adds reference information to an instruction sentence is generated and input to the LLM 30.

[0016] Fig. 2 is a block diagram showing an example of the configuration of the analysis device 10 and the information storage unit 20 according to this embodiment. As shown in Fig. 2, the analysis device 10 includes a prompt generation unit 11 and an analysis unit 12 as functional components for generating proposal information for measures. The information storage unit 20 includes a site information storage unit 21, a report information storage unit 22, a service information storage unit 23, and a template storage unit 24.

[0017] The functional blocks 11-12 of the analysis device 10 execute the processes described below through cooperation between hardware and software. For example, the processes of the functional blocks 11-12 are executed by programs stored in a storage medium such as a RAM, a ROM, a hard disk, or a semiconductor memory, under the control of a microcomputer including a CPU, a RAM, a ROM, etc. In addition to the microcomputer, a DSP (Digital Signal Processor), an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), etc. may be included.

[0018] Each of the storage units 21 to 24 of the information storage unit 20 is configured with a storage medium such as a magnetic disk (HDD: Hard Disk Drive, etc.), an optical disk (CD: Compact Disc, DVD: Digital Versatile Disc, BD: Blu-ray Disc, etc.), or a semiconductor memory (SSD: Solid State Drive, etc.).

[0019] The site information storage unit 21 stores site content information that indicates the contents of the target site to be analyzed. The site content information may be information obtained by copying text written on a web page of the target site, or may be text information that summarizes the contents of the web page. The site content information may be information manually created by a person in charge of the business entity that operates the analysis device 10, or may be information automatically generated using a scraping (crawling) technique.

[0020] Fig. 3 is a block diagram showing an example of a functional configuration of the analysis device 10 when the site content information is automatically generated using a scraping technique. As shown in Fig. 3, the analysis device 10 includes a site information acquisition unit 13 and a summary generation unit 14 as functional configurations for generating site content information. These functional blocks 13 to 14 also execute processing by cooperation between hardware such as a microcomputer and software (programs).

[0021] The site information acquisition unit 13 acquires text information of a web page from the target site 200 using a scraping technique. The URL (Uniform Resource Locator) of the target site 200 is acquired in advance from, for example, a company that operates the target site 200 and stored in the analysis device 10. The site information acquisition unit 13 accesses a web page of the target site 200 according to the URL stored in advance, and extracts and acquires text information from the web page. The target web page from which text information is to be acquired may be one web page specified from the target site 200, or may be multiple web pages included in the target site 200.

[0022] The summary generation unit 14 uses the text information acquired by the site information acquisition unit 13 to generate a sentence summarizing the contents posted on the webpage, and stores the generated summary as site content information in the site information storage unit 21. For example, the summary generation unit 14 generates the summary by utilizing the LLM 30. That is, the summary generation unit 14 generates a prompt including the text information acquired by the site information acquisition unit 13, followed by an instruction to create a summary, and inputs the prompt to the LLM 30 to acquire the summary from the LLM 30.

[0023] The method of generating the summary is not limited to the method of utilizing the LLM 30. For example, the summary generation unit 14 may use a technique such as natural language processing to extract important parts from the text information acquired by the site information acquisition unit 13 and generate the summary.

[0024] Returning to FIG. 2, the explanation will be given. The report information storage unit 22 stores report information of the analysis result regarding the access situation to the target site. The analysis of the access situation is executed, for example, using an access analysis device not shown, and report information indicating the execution result is stored in the report information storage unit 22. The analysis device 10 may have a function of analyzing the access situation of the website. The report information of the access situation is, for example, information indicating one or more of the number of accesses, the number of unique users, the number of organizational accesses, the number of unique organizations, the number of first-time visiting organizations, etc. The information listed here is an example, and the report information indicating other access situations may be used. For example, the number of clicks, the time spent on a page, the page exit rate, the page bounce rate, etc. may be used.

[0025] The service information storage unit 23 stores service content information that indicates the contents of a plurality of services that can be used to improve the access situation. The service content information is created, for example, by a person in charge of the business entity that operates the analysis device 10, and is stored in advance in the service information storage unit 23. Alternatively, text information described on a web page that introduces the service content may be acquired using a scraping technique, and the text information may be stored as the service content information in the service information storage unit 23. Also, a summary of the service content may be generated from the text information acquired using the scraping technique, and the summary may be stored as the service content information in the service information storage unit 23.

[0026] Here, the service information storage unit 23 may store a plurality of available services in association with each type of access status to be improved. Fig. 4 is a diagram that shows an example of association information between the type of access status and the available services. In Fig. 4, it is assumed that there are a total of 10 services available for improving the access status, and service content information indicating the contents of the 10 services is stored in the service information storage unit 23. The association information is information that indicates, for example, which of the 10 services is being used when the improvement target is the number of accesses, or which of the 10 services is being used when the improvement target is the number of unique users.

[0027] The template storage unit 24 stores a generation template for a prompt to be provided to the LLM 30. Fig. 5 is a diagram showing an example of a generation template. As shown in Fig. 5, the generation template includes an instruction statement 41 for instructing a proposal for an improvement measure for the access situation, and an input field 42 for reference information to be added to the instruction statement 41. The reference information input field 42 includes an input field for information representing the contents of the target site, an input field for report information representing the access situation to the target site, and an input field for information representing the contents of available services.

[0028] A generation template is prepared for each type of access status to be improved, for example. The generation templates for each type of access status only differ in the content of the instruction statement 41, and the reference information input field 42 is common. For example, a generation template for when the number of accesses is to be improved includes an instruction statement 41 that instructs proposing improvement measures for the number of accesses. A generation template for when the number of unique users is to be improved includes an instruction statement 41 that instructs proposing improvement measures for the number of unique users. A generation template for when multiple types of access status are to be improved includes an instruction statement 41 that instructs proposing improvement measures for the multiple types of access status. Note that a shared generation template including a common instruction statement 41 may be used regardless of the type of access status to be improved.

[0029] The prompt generation unit 11 of the analysis device 10 uses site content information of the target site stored in the site information storage unit 21, report information of the target site stored in the report information storage unit 22, service content information stored in the service information storage unit 23, and a generation template stored in the template storage unit 24 to generate a prompt that includes an instruction statement instructing the user to propose measures to improve the access situation and reference information generated from the site content information, report information, and service content information.

[0030] Here, the prompt generation unit 11 reads out from the template storage unit 24 a generation template corresponding to the type of access status to be improved, and uses it. Then, in the read generation template, the prompt generation unit 11 generates reference information by inputting site content information, report information, and service content information into a reference information input field 42 shown in Fig. 5. At this time, the prompt generation unit 11 acquires from the service information storage unit 23 a plurality of pieces of service content information stored in the service information storage unit 23, and inputs them into the reference information input field 42.

[0031] 4, when association information between the type of access situation to be improved and available services is stored in the service information storage unit 23, the prompt generation unit 11 acquires, from the service information storage unit 23, one or more pieces of service content information that are available according to the type of access situation to be improved, among the 10 pieces of service content information stored in the service information storage unit 23, and inputs them into the reference information input field 42. When such association information is not stored in the service information storage unit 23, the prompt generation unit 11 acquires, from the service information storage unit 23, the 10 pieces of service content information stored in the service information storage unit 23, and inputs them into the reference information input field 42.

[0032] 5, the example of the generation template shows a configuration in which the reference information input field 42 has an input field for site content information, an input field for report information, and an input field for service content information, but is not limited to this. The reference information input field 42 may have only one input field, and the site content information, report information, and service content information may be input in this order into the single input field.

[0033] The analysis unit 12 inputs the prompt generated by the prompt generation unit 11 to the LLM 30, and acquires from the LLM 30 proposal information for a measure utilizing a service whose service content information is described in the prompt as a measure that contributes to improving the access situation of the target site. The proposal information acquired from the LLM 30 includes, for example, information on the service to be utilized and an implementation plan for the measure utilizing the service. The service to be utilized is any one or more of the multiple services whose service content information is described in the prompt.

[0034] Thus, according to this embodiment, by providing the site content information and service content information as reference information to the LLM 30 together with the report information of the analysis result regarding the access situation to the target site, it becomes possible to generate proposal information for measures that are analyzed using the site content information and service content information as information sources as measures for improving the access situation indicated in the report information. As a result, according to this embodiment, proposal information for measures that are effective for improving the access situation to a website can be obtained based on the analysis result of the access situation to the website.

[0035] In order to obtain proposal information for measures utilizing more appropriate services, the analysis device 10 and the information storage unit 20 of this embodiment may be configured as follows. Three variations of the above embodiment are described below.

[0036] <First Modification> Fig. 6 is a block diagram showing an example of the configuration of an analysis device 10A and an information storage unit 20A according to a first modified example. In Fig. 6, the same reference numerals as those in Fig. 2 have the same functions, and therefore repeated explanations will be omitted here.

[0037] As shown in Fig. 6, analysis device 10A according to the first modification includes a prompt generation unit 11A instead of prompt generation unit 11 in the configuration shown in Fig. 2. Also, information storage unit 20A includes a site information storage unit 21A instead of site information storage unit 21 in the configuration shown in Fig. 2, and further includes a definition information storage unit 25.

[0038] The site information storage unit 21A stores site content information that indicates the contents of the target site to be analyzed, together with information indicating the type of business related to the target site. In the first modification, the site content information stored in the site information storage unit 21A may be manually created by a person in charge of the business entity that operates the analysis device 10A, or may be automatically generated using a scraping technique. In addition, the type of business related to the target site is specified by the person in charge, and the business type information is stored in the site information storage unit 21A together with the site content information.

[0039] In addition, here, an example has been described in which the business type related to the target site is specified in advance and stored in the site information storage unit 21A, but the present invention is not limited to this. For example, when the analysis device 10A generates the proposal information for the measures related to the target site, the person in charge may specify the business type.

[0040] The definition information storage unit 25 stores industry-specific service definition information that defines, for each industry, one or more available services among the multiple services. The multiple services referred to here means multiple services corresponding to multiple pieces of service content information stored in the service information storage unit 23. Note that, in the case where association information between types of access situations and available services is stored in the service information storage unit 23 as shown in Fig. 4, the multiple services mean multiple services associated with each type of access situation to be improved.

[0041] Fig. 7 is a diagram showing an example of business-specific service definition information. In the example shown in Fig. 7, it is assumed that there are a total of 10 services that can be used to improve the access situation, and that the 10 pieces of service content information are stored in the service information storage unit 23. Under this assumption, the business-specific service definition information shown in Fig. 7 is information indicating which of the 10 services can be used for each of a plurality of business types, from business type A to business type Z.

[0042] Prompt generation unit 11A refers to definition information storage unit 25 based on the business type specified for the target site, thereby identifying one or more services available in that business type, and acquires service content information indicating the contents of the identified one or more services from service information storage unit 23. Prompt generation unit 11A then generates a prompt including an instruction statement and reference information, using site content information, report information, and generation templates acquired from site information storage unit 21A, report information storage unit 22, and template storage unit 24, in addition to the above.

[0043] According to the first variant configured in this manner, prompt reference information is generated using only one or more pieces of service content information that can be used depending on the business type related to the target site, so that it is possible to obtain proposal information for measures that utilize more appropriate services depending on the business type related to the target site.

[0044] As shown in FIG. 8, priorities may be set for multiple services in the industry-specific service definition information stored in the service information storage unit 23, and the prompt generation unit 11 may generate a prompt that describes service content information representing the contents of one or more available services identified based on the industry related to the target site in order of priority.

[0045] The LLM 30 has a property of interpreting the contents described higher in the prompt as having a higher priority and executes the analysis process accordingly. By generating a prompt using this property of the LLM 30, it is possible to obtain proposal information for measures that utilize services with higher priority. This configuration using priority can also be applied to the above-mentioned embodiment or the second and third modified examples described below.

[0046] <Second Modification> The second modified example generates a prompt using one or more pieces of service content information corresponding to the business type related to the target site, similar to the first modified example. The difference from the first modified example is that in the first modified example, the person in charge specifies the business type related to the target site, whereas in the second modified example, the business type related to the target site is automatically specified.

[0047] Fig. 9 is a block diagram showing an example of the configuration of an analysis device 10B and an information storage unit 20B according to the second modified example. In Fig. 9, the same reference numerals as those shown in Fig. 2, Fig. 3, and Fig. 6 have the same functions, and therefore repeated explanations will be omitted here.

[0048] 9, analysis device 10B according to the second modification includes a prompt generation unit 11B instead of prompt generation unit 11, and further includes a site information acquisition unit 13 and a site analysis unit 15, in the configuration shown in FIG. 2. Also, information storage unit 20B includes a site information storage unit 21A instead of site information storage unit 21, and further includes a definition information storage unit 25, in the configuration shown in FIG.

[0049] The site analysis unit 15 estimates the business type related to the target site 200 by analyzing the contents of the target site 200 based on the text information of the target site 200 acquired by the site information acquisition unit 13. Then, the site analysis unit 15 stores business type information indicating the estimated business type in the site information storage unit 21A in association with the site content information of the target site.

[0050] For example, the site analysis unit 15 estimates the industry by utilizing the LLM 30. That is, the site analysis unit 15 generates a prompt including an instruction statement for instructing to estimate the industry, followed by text information of the target site acquired by the site information acquisition unit 13, and inputs the prompt to the LLM 30 to acquire industry information from the LLM 30.

[0051] The method of estimating the industry related to the target site is not limited to the method of utilizing the LLM 30. For example, the site analysis unit 15 may use a technique such as natural language processing to extract words from the text information acquired by the site information acquisition unit 13, and estimate the industry by referring to dictionary data that associates words with industries and stores the extracted words as keys.

[0052] Furthermore, in addition to estimating the type of business related to the target site, the site analysis unit 15 may generate site content information representing the contents of the target site and store this in the site information storage unit 21A. For example, the service content information may be generated by copying text information (sentences) of the target site acquired by the site information acquisition unit 13, or a sentence summarizing the contents posted on the web page may be generated as the service content information. When generating a summary sentence, it is possible to generate the service content information of the summary sentence by processing similar to that of the summary generation unit 14 shown in FIG. 3.

[0053] Here, an example has been described in which the business type related to the target site is estimated in advance and stored in the site information storage unit 21A, but the present invention is not limited to this. For example, as shown by the dotted arrow in FIG. 9, when the analysis device 10B generates the proposal information for measures related to the target site, the site analysis unit 15 may execute a process to estimate the business type, and the estimated business type information may be supplied to the prompt generation unit 11B. Also, when the analysis device 10B generates the proposal information for measures related to the target site, the site analysis unit 15 may execute a process to generate site content information and estimate the business type, and the result may be supplied to the prompt generation unit 11B. In the latter case, the site information storage unit 21A is not necessary. Similarly in FIG. 3 described above, the summary text generated by the summary generation unit 14 may be supplied to the prompt generation unit 11.

[0054] Prompt generation unit 11B refers to definition information storage unit 25 based on the business type estimated by site analysis unit 15 (business type information supplied from site analysis unit 15 to site information storage unit 21A and stored therein, or business type information supplied directly from site analysis unit 15), to identify one or more services available in the business type, and acquires service content information indicating the contents of the one or more identified services from service information storage unit 23. Prompt generation unit 11A then generates a prompt including an instruction statement and reference information, using the site content information, report information, and generation template acquired from site information storage unit 21A, report information storage unit 22, and template storage unit 24 in addition to the above.

[0055] In the second modified example configured in this way, the prompt reference information is generated using only one or more pieces of service content information that can be used according to the business type related to the target site, so that it is possible to obtain proposal information for measures that utilize more appropriate services according to the business type related to the target site. Also, according to the second modified example, it is possible to automatically specify the business type by analyzing the contents of the target site.

[0056] <Third Modification> In the above-mentioned first and second modified examples, one or more services available depending on the business type related to the target site are determined in advance. In contrast, in the third modified example, one or more services available depending on the target site are selected from a plurality of services prepared in advance, based on site content information, using the LLM30.

[0057] Fig. 10 is a block diagram showing an example of the configuration of an analysis device 10C and an information storage unit 20C according to a third modified example. In Fig. 10, the same reference numerals as those shown in Fig. 2 and Fig. 6 have the same functions, and therefore a duplicated description will be omitted here.

[0058] As shown in Fig. 10, analysis device 10C according to the third modification includes a first prompt generation unit 16 and a second prompt generation unit 18 instead of prompt generation unit 11, and a first analysis unit 17 and a second analysis unit 19 instead of analysis unit 12, in comparison with the configuration shown in Fig. 2. Second prompt generation unit 18 and second analysis unit 19 have the same functions as prompt generation unit 11 and analysis unit 12 shown in Fig. 2. Information storage unit 20C includes a template storage unit 24C instead of template storage unit 24 in the configuration shown in Fig. 2.

[0059] The template storage unit 24C stores two types of generation templates as generation templates for prompts to be provided to the LLM 30. Fig. 11A and Fig. 11B are diagrams that show examples of the two types of generation templates. As shown in Fig. 11A, the first generation template includes a first instruction statement 111 that instructs the specification of one or more services, and an input field 112 for first reference information to be added to the first instruction statement 111. The first reference information input field 112 includes an input field for information representing the contents of the target site, and an input field for information representing the contents of multiple services prepared in advance.

[0060] 11B, the second generation template is similar to the generation template shown in FIG 5, and includes a second instruction statement 41 that instructs proposing measures to improve the access situation, and an input field 42 for second reference information to be added to the second instruction statement 41. The second reference information input field 42 includes an input field for information representing the content of the target site, an input field for report information representing the access situation to the target site, and an input field for information representing the content of one or more available services selected by LLM 30.

[0061] The first prompt generation unit 16 generates a first prompt including a first instruction statement instructing the identification of one or more services and first reference information, using site content information stored in the site information storage unit 21 (which may be site content information supplied from the site analysis unit 15, as in the second variant example), service content information stored in the service information storage unit 23, and a first generation template (see Figure 11A) stored in the template storage unit 24C.

[0062] Here, first prompt generation unit 16 reads out a first generation template corresponding to the type of access situation to be improved from template storage unit 24C and uses it. Then, first prompt generation unit 16 generates reference information by inputting site content information and a plurality of pieces of prepared service content information into first reference information input field 112 shown in FIG. 11A in the read out first generation template.

[0063] The multiple pieces of service content information to be input at this time refer to service content information related to multiple services that have been prepared in advance as those that can be used to improve the access situation. When association information between the type of access situation to be improved and the available services is stored in the service information storage unit 23 as shown in Fig. 4, the service content information related to multiple services defined as available according to the type of access situation is input.

[0064] The first analysis unit 17 inputs the first prompt generated by the first prompt generation unit 16 to the LLM 30, and thereby acquires from the LLM 30 service specification information representing the specification result of one or more services analyzed as available according to the site content information. Then, by referring to the service information storage unit 23 based on the acquired service specification information, the first analysis unit 17 acquires service content information representing the contents of the one or more services from the service information storage unit 23. Then, the first analysis unit 17 supplies the acquired one or more service content information to the second prompt generation unit 18.

[0065] The second prompt generation unit 18 generates a second prompt including a second instruction statement instructing the generation of proposed information and second reference information, using site content information of the target site stored in the site information storage unit 21, report information of the target site stored in the report information storage unit 22, service content information representing the contents of one or more available services identified by the first analysis unit 17, and a second generation template (see Figure 11B) stored in the template storage unit 24C.

[0066] Like the first prompt generation unit 16, the second prompt generation unit 18 also reads out from the template storage unit 24C a second generation template corresponding to the type of access status to be improved, and uses the read out template. The second prompt generation unit 18 then generates second reference information by inputting site content information, report information, and service content information into the second reference information input field 42 shown in Fig. 11B in the read second generation template. At this time, the second prompt generation unit 18 inputs one or more pieces of service content information provided from the first analysis unit 17 into the second reference information input field 42.

[0067] The second analysis unit 19 inputs the second prompt generated by the second prompt generation unit 18 to the LLM 30, and acquires from the LLM 30 proposal information for a measure utilizing a service whose service content information is described in the prompt as a measure that contributes to improving the access situation of the target site. The proposal information acquired from the LLM 30 includes, for example, information on the service to be utilized and an implementation plan for the measure utilizing the service. The service to be utilized is any one or more of the one or more services whose service content information is described in the prompt (services identified by the first analysis unit 17).

[0068] According to the third modified example configured in this manner, one or more services available depending on the contents of the target site are identified by the LLM 30, and prompt reference information is generated using only the service content information related to the identified services, so that it is possible to obtain proposal information for measures utilizing more appropriate services depending on the contents of the target site. Note that although it is possible that the LLM 30 identifies an appropriate service corresponding to the business type depending on the contents of the target site, the third modified example may be applied in combination with the first modified example or the second modified example.

[0069] It should be noted that the above-described embodiment and each of the modified examples are merely examples of the embodiment of the present invention, and the technical scope of the present invention should not be interpreted as being limited thereby. In other words, the present invention can be embodied in various forms without departing from the gist or main characteristics thereof. [Explanation of symbols]

[0070] 10,10A,10B,10C analysis device 11, 11A, 11B Prompt generation section 12 Analysis Department 13 Site Information Acquisition Department 14 Summary generator 15 Site Analysis Section 16 First prompt generator 17 First Analysis Section 18 Second prompt generator 19 Second Analysis Section 20,20A,20B,20C Information storage section 21, 21A Site information storage section 22 Report information storage unit 23 Service information storage unit 24,24C Template memory section 25 Definition information storage section 30 LLM (Large Scale Language Models)

Claims

1. a prompt generation unit that generates a prompt including an instruction for proposing measures to improve the access status and reference information generated from the site content information, the report information, and the service content information, using site content information representing the content of a target site, which is a website to be analyzed, report information of an analysis result regarding the access status of the target site, service content information representing the content of one or more services that can be used to improve the access status, and a prompt generation template; an analysis unit that inputs the prompt generated by the prompt generation unit into a large-scale language model and acquires, from the large-scale language model, proposal information for a measure that utilizes the service as a measure that contributes to improving the access status of the target site; 1. An analysis system comprising:

2. a service information storage unit that stores service content information representing the contents of a plurality of services; a definition information storage unit configured to store business-specific service definition information that defines one or more of the available services among the plurality of services for each business type, The prompt generation unit refers to the definition information storage unit based on the business type specified for the target site, thereby identifying one or more services available in the business type, acquiring service content information representing the contents of the identified one or more services from the service information storage unit, and generating the prompt using the acquired service content information.

2. The analysis system according to claim 1 .

3. a service information storage unit that stores service content information representing the contents of a plurality of services; a definition information storage unit that stores business-specific service definition information that defines one or more of the available services among the plurality of services for each business type; a site analysis unit that estimates a business type related to the target site by analyzing the content of the target site, The prompt generation unit refers to the definition information storage unit based on the business type estimated by the site analysis unit to identify one or more services available in the business type, obtains service content information representing the contents of the identified one or more services from the service information storage unit, and generates the prompt using the obtained service content information.

2. The analysis system according to claim 1 .

4. A service information storage unit stores service content information representing the contents of a plurality of services, The prompt generation unit a first prompt generation unit that generates a first prompt using site content information representing the content of the target site, service content information representing the contents of a plurality of services, and a first generation template, the first prompt including a first instruction statement instructing the specification of the one or more available services, and first reference information generated from the site content information and the service content information; a second prompt generation unit that generates a second prompt using the site content information representing the content of the target site, the report information of the target site, the service content information representing the content of the one or more available services identified by the analysis unit, and a second generation template, the second prompt including a second instruction statement instructing generation of the proposed information and second reference information generated from the site content information, the report information, and the service content information; The analysis unit is a first analysis unit that inputs the first prompt generated by the first prompt generation unit into the large-scale language model, thereby acquiring service specification information representing a result of identifying the one or more services from the large-scale language model, and acquires the service content information representing contents of the one or more services from the service information storage unit by referring to the service information storage unit based on the acquired service specification information; a second analysis unit that inputs the second prompt generated by the second prompt generation unit into the large-scale language model and acquires the measure proposal information from the large-scale language model.

2. The analysis system according to claim 1 .

5. a priority level is set for the plurality of services whose service content information is stored in the service information storage unit; The prompt generating unit generates a prompt describing service content information indicating the contents of the one or more identified services in the order of priority.

5. The analysis system according to claim 2, wherein the analysis system comprises:

6. a prompt generation unit that generates a prompt including an instruction for proposing measures to improve the access status and reference information generated from the site content information, the report information, and the service content information, using site content information representing the content of a target site, which is a website to be analyzed, report information of an analysis result regarding the access status of the target site, service content information representing the content of one or more services that can be used to improve the access status, and a prompt generation template; an analysis unit that inputs the prompt generated by the prompt generation unit into a large-scale language model and acquires, from the large-scale language model, proposal information for a measure that utilizes the service as a measure that contributes to improving the access status of the target site; An analytical device characterized by:

Images