Request response support device

Abstract

The objective is to provide a request handling support device that improves work efficiency and reduces the burden on staff by allowing them to understand the format and other requirements of submitted documents. [Solution] The request response support device 1 comprises a model generation unit 5 and an estimation unit 6. The model generation unit 5 uses request information transmitted from the requesting terminal 50 as input data and business information corresponding to the request information as output data, and generates a model by machine learning from this training data. The estimation unit 6 uses newly transmitted request information from the requesting terminal 50 as estimation input data and estimates business information corresponding to the estimation input data based on the model. The request information is a request text that includes at least the name of the submission to be submitted by the requesting party, and the business information is format information for the submission corresponding to the submission name and destination information for the submission corresponding to the submission name.

Description

【Technical Field】 【0001】 The present invention relates to a request response support device that extracts and outputs business information corresponding to a request destination from request information requested by a request source to the request destination. 【Background Art】 【0002】 Conventionally, when a request source sends a request letter with common content to each department within an organization that is the request destination, only the submission deadline is different from the previous request, and the content of the request letter itself is often the same each time. Therefore, the same investigation can be conducted each time, and the response to the request content can be made in the same format. However, since the request source requests multiple request destinations simultaneously, the request letter may include request content and submissions that have nothing to do with the department to which the request source belongs. In this case, it is difficult to accurately grasp the investigation target and the submissions. In addition, since the frequency of requests is once every six months or one year, and there is a large time interval, when the person in charge who responded to the previous request transfers, the successor has to read and understand the request letter from the beginning. In this case, since it takes time to respond, the work efficiency is not good. Also, there is a risk that the burden on the successor will become excessive. In order to solve such problems, in recent years, technologies for improving work efficiency have been developed by analyzing and using historical information of past operations, and inventions related thereto have already been disclosed. 【0003】 Patent Document 1 discloses an invention related to a business support system and the like that provides useful information about a business obtained from historical information (log information) of past operations under the name of "Business Support System and Business Support Method". The invention disclosed in Patent Document 1 is characterized by storing the following: a business pattern estimation model which is generated based on information that associates a business flow, which is information that shows the flow of a series of tasks performed in relation to a business, with a business pattern, which is information that shows the type of business flow corresponding to a series of tasks, and identifies a business pattern corresponding to history information, which is information that shows the history of tasks related to a business; pattern-specific series information which includes information that classifies the series constituting the business flow corresponding to each business pattern according to the impact each has on the results of the business execution; and target history information which is history information of the business to be supported; identifying a business pattern corresponding to the target history information using the business pattern estimation model; and identifying the current task in the business to be supported by comparing the target history information with a business flow based on the pattern-specific series information of the identified business pattern. In inventions with these characteristics, log information is acquired separately for each system, such as a sales management system or a travel expense settlement system. Furthermore, by identifying the current task, subsequent tasks such as "quoting" or "proposal creation" can also be identified. [Prior art documents] [Patent Documents] 【0004】 [Patent Document 1] Japanese Patent Publication No. 2023-141668 [Overview of the project] [Problems that the invention aims to solve] 【0005】 According to the invention disclosed in Patent Document 1, the subsequent operations are identified, and therefore their names can be determined. However, even if the name of the task is identified, the specific format to be used and the submission destination are not specified, so the problem of difficulty in accurately understanding the required documents remains unresolved. Consequently, issues such as poor work efficiency and a heavy burden on the successor remain. 【0006】 This invention addresses the aforementioned conventional circumstances and aims to provide a request handling support device that can improve operational efficiency and reduce the burden on the person in charge by allowing the format of the submitted documents to be specifically understood. [Means for solving the problem] 【0007】 To achieve the above objective, the first invention is a request response support device that extracts and outputs business information to be handled by a requesting party from request information submitted by a requesting party to a requesting party, the request response support device is installed at the requesting party and is capable of communicating with the requesting party's requesting terminal via a communication channel, and comprises a model generation unit and an estimation unit, the model generation unit generates a model by machine learning from training data which is request information transmitted from the requesting party's terminal as input data and business information corresponding to the request information as output data, the estimation unit uses newly transmitted request information from the requesting party's terminal as estimation input data and estimates business information corresponding to the estimation input data based on the model, the request information is a request text which includes at least the name of the submission to be submitted by the requesting party, and the business information which includes format information of the submission corresponding to the name of the submission and destination information of the submission corresponding to the name of the submission. 【0008】 In an invention with such a configuration, the request letter may include, in addition to the name of the submitted work, the submission deadline, information about the recipient, etc. Furthermore, the request letter may include links in addition to text. Furthermore, the format information for submitted documents may include not only the name of the document but also a number indicating the format. In addition, the submission destination information may include not only the name of the submission destination but also an identification code assigned to each destination. 【0009】 In the invention with the above configuration, the model generation unit extracts the name of the submitted item from the request information, which is the request text, learns the relationship between the format information of the submitted item and the destination information corresponding to this submitted item name, and generates a model. Using this model, the estimation unit can extract the name of the submitted item from the newly transmitted request information, which is the request text, and estimate the format information of the submitted item and the destination information corresponding to this submitted item name. 【0010】 The second invention is characterized in that, in the first invention, a plurality of request response support devices are provided so as to be able to communicate with the requesting terminal via a communication channel, and in each of the plurality of request response support devices, the model generation unit generates a different model by learning business information with different content for each of the plurality of requesting parties, which is extracted from a single request information transmitted by the requesting terminal. 【0011】 In this type of invention, the number of request handling support devices is equal to or greater than the number of requesting terminals. Furthermore, the number of requesting terminals may be multiple. Specifically, in addition to a configuration in which one requesting terminal transmits one request to multiple request handling support devices, a configuration is conceivable in which one of two requesting terminals transmits a first request to a first group consisting of multiple request handling support devices, and the other of the two requesting terminals transmits a second request to the first group and a second group consisting of multiple request handling support devices other than the first group. 【0012】 In the invention with the above configuration, in addition to the function of the first invention, a different model is generated for each of the multiple request response support devices. Therefore, each estimation unit does not extract the names of submissions unrelated to the departments where each request response support device is installed from the request text of the newly transmitted request information. Thus, each estimation unit estimates the format, etc., only for submissions that each department should handle. 【0013】 The third invention is characterized in that, in the first or second invention, the request information includes attached documents in addition to the request document, and the request document includes the deadline for submission of the submission in addition to the name of the submission. In an invention with such a configuration, the attached documents are provided in file format and may include links in addition to text-based documents. Furthermore, the content of the attached documents may include at least one of the name of the submission and the submission deadline. 【0014】 In the invention with the above-described configuration, in addition to the operations of the first or second invention, the model generation unit extracts the submission name of the submission and the submission deadline from the request sentence, and also extracts at least one of the submission name and the submission deadline from the attached materials. Further, the model generation unit learns the relevance between all the extracted submission names, the format information of the submission corresponding to the submission deadline, and the submission destination information, and generates a model. Therefore, the format information of the submission and the submission destination information estimated by the estimation unit take into account the submission deadline. 【Effect of the Invention】 【0015】 According to the first invention, since the estimation unit can estimate the format information of the submission corresponding to the submission name extracted from the newly transmitted request sentence and the submission destination information, these can be accurately grasped. In addition, since the person in charge does not need to interpret the request sentence from scratch and the time until the response can be shortened, the work efficiency can be significantly improved. 【0016】 According to the second invention, in addition to the effect of the first invention, each estimation unit estimates the format and the like related only to the submissions that each department should handle. Therefore, each department can accurately grasp the format information of the submission and the submission destination information. 【0017】 According to the third invention, in addition to the effect of the first or second invention, the format information of the submission and the submission destination information estimated by the estimation unit take into account the submission deadline. Therefore, these business information becomes more accurate. 【Brief Description of the Drawings】 【0018】 [Figure 1] It is a configuration diagram of a request response support device according to an embodiment. [Figure 2] It is a configuration diagram in the case where a plurality of request response support devices are provided. [Figure 3] It is a process diagram of a model generation method executed by the request response support device. [Figure 4] It is a process diagram of a business information estimation method executed by the request response support device. [Figure 5] This is an example of a request sentence included in the request information. [Figure 6] This is an example of an attached file included in the request information. [Figure 7] This is an example of business information. 【Mode for Carrying Out the Invention】 【Examples】 【0019】 The request response support device according to an embodiment of the present invention will be described in detail with reference to FIGS. 1 to 7. FIG. 1 is a configuration diagram of the request response support device according to the example. As shown in FIG. 1, the request response support device 1 according to the example is a device that extracts and outputs business information corresponding to the request destination from the request information requested by the request source to the request destination. Here, the request source is a department that transmits request information for requesting the creation and submission of submissions to each department within an organization such as a company. And the request destination is a plurality of departments within an organization such as a company that receive the request information. Further, the request response support device 1 is a computer provided at the request destination, and includes an input / output unit 2, a control unit 3, a storage unit 7, and a display unit 12. Furthermore, the request response support device 1 can communicate with a request source terminal �0 provided at the request source via a communication path 60. Note that the communication path 60 is a network. 【0020】 Among these, the input / output unit 2 receives the request information transmitted by the request source terminal fifty. Specifically, this request information is a request sentence by email transmitted by the request source terminal fifty and attached materials attached to this email. However, the attached materials may not be attached. Furthermore, the input / output unit 2 transmits a submission created based on the business information output by the estimation unit 6 described later to a submission destination terminal 51 (see FIG. 2) via the communication path 60. Next, the control unit 3 is a central processing unit that controls the operation of the request response support device 1, and includes a character extraction unit 4, a model generation unit 5, and an estimation unit 6. The storage unit 7 includes an email storage unit 8, a teacher data storage unit 9, a model storage unit 10, and an estimation data storage unit 11. The following describes the various components of the control unit 3. 【0021】 The character extraction unit 4 automatically opens the request text and attached documents of the request information, which are the input data of the training data input via the input / output unit 2, and the request text and attached documents of the request information, which are the input data for estimation. The character extraction unit 4 then extracts specific text characters contained in the opened request information, etc., and the characters before or after them, and outputs character data. Specific text characters include, for example, characters such as "report," "table," "slip," and "form" used in the name of the submitted item, and characters such as "year," "month," "day," "hour," "minute," "AM," and "PM" used in the submission deadline of the submitted item. The opening of the request information and the extraction of characters are performed using known methods. 【0022】 The model generation unit 5 uses request information transmitted from the requesting terminal 50 as input data and business information corresponding to the request information as output data, and generates a model using machine learning with known methods from this training data. The above request information includes a request letter containing the name of the item to be submitted, the deadline for submission, and any attached documents. The request letter also includes the name of the item to be submitted and the deadline for submission. Furthermore, the business information includes format information for the submitted document corresponding to the document name, and information on the recipient of the submitted document corresponding to the document name. Specifically, the input data for the training data is character data extracted by the character extraction unit 4 from the request information, and the output data for the training data is business information corresponding to this character data. 【0023】 The estimation unit 6 uses the newly transmitted request information from the requesting terminal 50 as estimation input data, and estimates the business information corresponding to the estimation input data based on the model generated by the model generation unit 5. Specifically, the estimation input data is the character data extracted by the character extraction unit 4 from the new request information. The output of the estimation unit 6 is the business information corresponding to this character data. 【0024】 Next, the email storage unit 8 stores the email sent from the requesting terminal 50 and its attachments. The training data storage unit 9 stores the input data and output data for training data used by the model generation unit 5 to generate a model. As mentioned above, the input data is the character data output by the character extraction unit 4. Furthermore, the model storage unit 10 stores the models generated by the model generation unit 5. Furthermore, the estimation data storage unit 11 stores the estimation input data input to the estimation unit 6 and the business information corresponding to the estimation input data estimated by the estimation unit 6. As mentioned above, the estimation input data is the character data output by the character extraction unit 4. The display unit 12 is a monitor that displays the business information estimated by the estimation unit 6. 【0025】 Next, the configuration when multiple request response support devices are provided will be explained using Figure 2. Figure 2 is a configuration diagram when multiple request response support devices are provided. Note that the same reference numerals are used for the components shown in Figure 1 in Figure 2, and their explanations are omitted. As shown in Figure 2, multiple request response support devices 1 are provided (indicated by symbols 1-1, 1-2, ..., 1-m, where m is a natural number) that can communicate with the requesting terminal 50 via the communication channel 60. In each of the multiple request handling support devices 1, the model generation unit 5 (see Figure 1) generates a different model by learning business information that differs for each of the multiple request destinations, extracted from a single request information transmitted by the requesting terminal 50. This is because, although the input data of the training data received by each of the multiple request handling support devices 1 is common, the output data of the training data differs for each of the multiple request handling support devices 1. 【0026】 Therefore, in each of the multiple request handling support devices 1, the estimation unit 6 (see Figure 1) uses the common request information newly transmitted from the requesting terminal 50 as estimation input data, and estimates different business information corresponding to the estimation input data based on different models. Then, each person in charge of the multiple request handling support devices 1 creates at least one submission based on different business information. The person in charge then sends the created submission to one of the multiple submission destination terminals 51 (represented by symbols 51-1, 51-2, ..., 51-n, where n is a natural number) that is specified in the submission destination information of the business information. 【0027】 Next, the model generation method performed by the request response support device will be explained using Figure 3. Figure 3 is a process diagram of the model generation method performed by the request response support device. As shown in Figure 3, the model generation method 20 comprises a character extraction step in step S21, a training data creation step in step S22, and a model generation step in step S23. Each step will be described below. Step S21, the character extraction process, is a process in which the character extraction unit 4 extracts pre-set specific characters from the request text and attached documents of the email sent from the requesting terminal 50. The extracted character data is stored in the training data storage unit 9. 【0028】 Step S22, the training data creation process, is a process in which the model generation unit 5 creates training data by associating the character data stored in the training data storage unit 9 with predetermined business information corresponding to this character data. The created training data is stored in the training data storage unit 9. 【0029】 Step S23, the model generation process, is the process in which the model generation unit 5 generates a model from the training data stored in the training data storage unit 9 using machine learning. The generated model is stored in the model storage unit 10. This completes the model generation method 20. 【0030】 Furthermore, the method for estimating business information performed by the request response support device will be explained using Figure 4. Figure 4 is a process diagram of the business information estimation method performed by the request response support device. As shown in Figure 4, the business information estimation method 30 comprises the steps of acquiring input data for estimation in step S31, extracting characters in step S32, estimation in step S33, and display in step S34. Each step will be described below. 【0031】 Step S31, the process of acquiring input data for estimation, is a process in which the input / output unit 2 acquires newly transmitted request information from the requesting terminal 50. The acquired new request information is stored in the estimation data storage unit 11. The new request information, like the request information for training data, includes a request statement that includes the name of the submission to be submitted by the recipient, the deadline for submission of the submission, and attached documents. 【0032】 Step S32, the character extraction process, is a process in which the character extraction unit 4 refers to the estimated data storage unit 11 and extracts pre-set specific characters from the email request text and attached documents included in the new request information. The extracted character data is stored in the estimated data storage unit 11. 【0033】 Step S33, the estimation process, is a process in which the estimation unit 6 estimates business information corresponding to the character data stored in the estimation data storage unit 11 based on the model generated by the model generation unit 5. The estimated business information is associated with the extracted character data and stored in the estimation data storage unit 11. The estimated business information includes, like the business information of the training data, information on the format of the submission corresponding to the name of the submission and information on the destination of the submission corresponding to the name of the submission. 【0034】 Step S34, the display step, is the step in which the display unit 12 displays the estimated business information. This completes the business information estimation method 30. 【0035】 Furthermore, the contents of the request information will be explained using Figures 5 and 6. Figure 5 is an example of a request text included in the request information. Figure 6 is an example of an attached file included in the request information. As shown in Figure 5, the request information is the request text 40, which is the input data for estimation, and includes the names of the submissions 41a to 41c and the submission deadlines 42a to 42c that the requesting party should submit. The request text 40 also includes the contents of the attached document 43, which is titled "Attachment 1** Table". Although the names of the submissions 41c and the submission deadlines 42c are not specifically stated, they are included in the attached document 43, so the estimation unit 6 can obtain them by referring to the attached document 43. The input data for the training data also includes the names of the submissions and the submission deadlines, similar to the request text 40. 【0036】 As shown in Figure 6, the attached document 43, which is the input data for estimation, is a table containing the name of the submitted item 44 and the submission deadline 45. As mentioned above, the attached document 43 is automatically opened by the character extraction unit 4, and the name of the submitted item 44 and the submission deadline 45 are extracted as character data. 【0037】 Next, we will explain the contents of business information using Figure 7. Figure 7 is an example of business information. As shown in Figure 7, the business information 46 consists of the form information 47 of the submitted document corresponding to the name of the submitted document, and the submission destination information 48 of the submitted document corresponding to the name of the submitted document. The form information 47 is displayed as the name of the form (e.g., ○○ Specifications, ×× Survey Form) or a number that identifies the form (e.g., ◇◇ Form 1, ◇◇ Form 2). The submission destination information 48 is displayed as the name of the department where the submission destination terminal 51 is located (e.g., ○○○ Department, ××× Department). However, other display formats are also acceptable. 【0038】 As explained above, the request response support device 1 allows the estimation unit 6 to estimate the format information 47 of the submitted documents and the destination information 48, making it possible to accurately grasp these. Furthermore, since the person in charge does not need to interpret the request from scratch, the time to respond is shortened, significantly improving work efficiency. In addition, regardless of whether the person in charge has experience handling the request information, they can quickly obtain the necessary business information46, greatly reducing the burden on the person in charge. Furthermore, since the request information includes the submission deadline 45 in addition to the name of the submitted item 44, the format information 47 and the destination information 48 estimated by the estimation unit 6 will take the submission deadline into consideration. Therefore, the business information 46 will be more accurate. 【0039】 Furthermore, in each of the multiple request handling support devices 1, the estimation unit 6 estimates only the business information 46 related to the submissions that each department is responsible for, making it possible to accurately grasp the format information 47 and destination information 48 of the submissions required by each department. Furthermore, since the estimation unit 6 also uses attached document 43, which includes the name of the submitted item 44 and the submission deadline 45, as input data for estimation, it can estimate the form information 47 and the submission destination information 48 in detail, and is expected to prevent submission omissions. 【0040】 It should be noted that the request handling support device according to the present invention is not limited to those shown in the examples. For example, the input data for training data and the request information in the input data for estimation do not necessarily have to include attached documents, and the request text does not necessarily have to include a submission deadline. Furthermore, if the request text or attached documents include links, the character extraction unit 4 may automatically open these links. [Industrial applicability] 【0041】 This invention can be used as a request handling support device that extracts and outputs business information that the requesting party will handle from the request information submitted by the requesting party to the requesting party. [Explanation of Symbols] 【0042】 1…Request Response Support Device 2…Input / Output Unit 3…Control Unit 4…Character Extraction Unit 5…Model Generation Unit 6…Estimation Unit 7…Storage Unit 8…Character Data Storage Unit 9…Training Data Storage Unit 10…Model Storage Unit 11…Estimation Data Storage Unit 12…Display Unit 20…Model Generation Method 30…Business Information Estimation Method 40…Request Text 41a~41c…Submission Item Name 42a~42c…Submission Deadline 43…Attached Documents 44…Submission Item Name 45…Submission Deadline 46…Business Information 47…Form Information 48…Submission Destination Information 50…Requesting Terminal 51…Submission Destination Terminal 60…Communication Channel

Claims

[Claim 1] A request handling support device that extracts and outputs business information to be handled by the requesting party from the requesting party's request information, The aforementioned request handling support device is installed at the request destination and is capable of communicating with the requesting party's terminal via a communication channel, and comprises a model generation unit and an estimation unit. The model generation unit generates a model using machine learning from training data, which consists of the request information transmitted from the requesting terminal as input data and the business information corresponding to the request information as output data. The estimation unit uses the request information newly transmitted from the requesting terminal as estimation input data, and estimates the business information corresponding to the estimation input data based on the model. The aforementioned request information is a request document that includes at least the names of the submissions that the requesting party should submit, The request support device is characterized in that the aforementioned business information consists of format information for the submitted item corresponding to the name of the submitted item and destination information for the submitted item corresponding to the name of the submitted item. [Claim 2] Multiple request handling support devices are provided, each capable of communicating with the requesting terminal via the communication path. The request response support device according to claim 1, wherein in each of the multiple request response support devices, the model generation unit generates a different model by learning the business information, which has different content for each of the multiple request destinations, extracted from the request information transmitted by the requesting terminal. [Claim 3] The aforementioned request information includes the request text, as well as attached documents. The request response support device according to claim 1 or 2, characterized in that the request statement includes, in addition to the name of the submitted item, the deadline for submitting the submitted item.

Images