WO2020255361A1

WO2020255361A1 - Accounting processing system, accounting processing method, and accounting processing program

Info

Publication number: WO2020255361A1
Application number: PCT/JP2019/024611
Authority: WO
Inventors: 葛鴻鵬; 松原正和
Original assignee: ファーストアカウンティング株式会社
Priority date: 2019-06-20
Filing date: 2019-06-20
Publication date: 2020-12-24
Also published as: JPWO2020255361A1

Abstract

An accounting processing system 1 comprises: a sorting element extraction unit 11 which uses a sorting element extraction AI for outputting as text a sorting element from image data of a voucher to extract as text the sorting element from the image data of the voucher and outputs a degree of reliability in the extraction; a first RPA execution unit 13 which can automatically execute a process using the sorting element text in a pre-registered process sequence; and a sorting element determination unit 12 which, if a sorting element degree of reliability Re outputted by the sorting element extraction unit 11 is greater than or equal to a first threshold value T1, determines that the inputted information is valid and causes the execution of the process performed by the first RPA execution unit 13.

Description

Accounting system, accounting method, accounting program

The present invention extracts journal elements such as date, business partner, amount, and description from voucher image data such as receipts and receipts, and can automatically journal from the extracted journal elements. Accounting system, accounting method, accounting processing. Regarding the program.

Conventionally, as an accounting process for vouchers, tax accountants, accountants, and bookkeepers read the information on the vouchers one by one, enter journal elements such as date, business partner, and amount in the book, and respond to the journal elements. The journal entry was made by empirically judging the account that was created.

In this way, if a human reads the voucher one by one and makes a journal entry, the work efficiency is poor, and the accuracy of the journal entry largely depends on the experience of the person in charge, and there is a problem that the accuracy of the journal entry varies. It was.

Therefore, using an OCR (Optical Character Reader) device, simply read the contents of the voucher as electronic data and send it to the journal analysis center system via the Internet, and the result of the journal entry for the bookkeeping transaction shown in the voucher. A so-called cloud-type accounting system has been developed that allows users to refer to (see Patent Document 1).

Specifically, in the technique described in Patent Document 1, voucher data taken with a mobile terminal or the like is analyzed by a journal element extraction means, and journal element information is extracted as text format data. Specifically, it recognizes the character information group recorded in the image data of the voucher such as a receipt and converts it into text data composed of items such as user ID, date, purchase destination name, and purchased product name. .. Then, the product group corresponding to the product name included in the journal element information converted into text data is obtained from the product master, and the recommended journal is shown using the mapping table corresponding to the product group (paragraph of Patent Document 1). 0118 to paragraph 0128).

Japanese Unexamined Patent Publication No. 2014-235484

In Patent Document 1, the journal element is extracted from the voucher using an OCR device, but the display format of the voucher is various, and the information described in the voucher is also various, so that the extraction accuracy of the journal element is high. It's not easy to do. For example, if a plurality of company names are described in one voucher, it may not be possible to determine which company name is the issuer of the voucher.

Because the journal elements cannot always be extracted accurately in this way, the person in charge will eventually check all the extracted journal elements, and work efficiency cannot be improved. The same applies to journals, and recommended journals are not always correct and must be checked by the person in charge, so work efficiency cannot be improved.

Furthermore, even if the journal elements can be extracted accurately, the extracted journal elements can be entered separately in the journal entry field, or even if the correct recommended journal is output, it can be reported to the superior or the person in charge of other departments. Many simple tasks occur in accounting, such as sending emails.

The present invention has been made to solve such a problem, and the purpose of the present invention is to reduce confirmation work and simple work by a person in charge and improve the overall work efficiency of accounting processing. To provide an accounting system, accounting method, and accounting program that can be used.

In order to achieve the above object, the accounting system according to the present invention extracts the journal element as text from the image data of the voucher by the journal element extraction AI for outputting the journal element as text from the image data of the voucher. Moreover, the journal element extraction unit that outputs the reliability of the extraction, the automatic processing unit that can automatically execute the processing using the text of the journal element by the processing procedure registered in advance, and the journal element extraction unit output the data. It is provided with a determination unit that determines that the input information is appropriate and executes the processing by the automatic processing unit when the reliability is equal to or higher than a predetermined threshold value.

In order to achieve the above object, the accounting system according to the present invention extracts the journal element as text from the image data of the voucher by the journal element extraction AI for outputting the journal element as text from the image data of the voucher. The journal element extraction unit, the automatic processing unit that can automatically execute the processing using the text of the journal element by the processing procedure registered in advance, and the voucher input by a means different from the journal element extraction unit. It is stored in the journal element data storage unit that stores the text of the journal element included in the image data, the text of the journal element extracted from the journal element extraction unit, and the journal element data storage unit that corresponds to the journal element. It is provided with a determination unit that determines that the input information is appropriate and executes processing by the automatic processing unit when the texts of the journal elements are matched and matched.

In order to achieve the above object, the accounting system according to the present invention outputs the journal result to the input journal element by the journal AI for outputting the journal result according to the journal element, and the subject A journal unit that outputs the reliability of the journal result, an automatic processing unit that can automatically execute processing using the journal result by a processing procedure registered in advance, and the reliability output from the journal unit are predetermined. When the value is equal to or greater than the threshold value of, it is determined that the input information is appropriate, and a determination unit for executing the process by the automatic processing unit is provided.

Further, in the above-mentioned accounting processing device, the determination unit may generate warning display data when it determines that the input information is not valid.

Further, in the above-mentioned accounting processing apparatus, the automatic processing unit may be able to automatically execute a process of performing automatic journalizing using the journal element.

Further, in the above-mentioned accounting processing device, the automatic processing unit may be able to automatically execute a process of sending the journal entry result by e-mail to a specific destination.

Further, in order to achieve the above object, the accounting method according to the present invention uses the journal element extraction AI for outputting the journal element as text from the image data of the voucher, and uses the journal element as text from the image data of the voucher. The text of the extracted journal element is appropriate when the journal element extraction step that extracts and outputs the reliability of the extraction and the reliability output from the journal element extraction step are equal to or higher than a predetermined threshold. When it is determined that the judgment step of determining that the journal element is appropriate and the journal element extracted in the determination step is appropriate, the process using the text of the journal element is automatically executed by the pre-registered processing procedure. The processing process is performed by the computer.

Further, in order to achieve the above-mentioned object, in the accounting processing program according to the present invention, a computer is made to execute the above-mentioned accounting processing method.

According to the present invention using the above means, it is possible to reduce the confirmation work and the simple work by the person in charge and improve the overall work efficiency of the accounting process.

It is a schematic block diagram which showed the whole structure of the accounting processing system which concerns on one Embodiment of this invention. It is a flowchart which showed the flow of the accounting processing executed by the accounting processing apparatus. It is explanatory drawing which shows an example of the transition of the accounting processing executed by the accounting processing apparatus. It is explanatory drawing which shows an example of the transition of the accounting processing executed by the accounting processing apparatus.

Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a schematic configuration diagram showing an overall configuration of an accounting processing system according to an embodiment of the present invention, and the configuration of the present embodiment will be described based on the same diagram.

As shown in FIG. 1, the accounting processing system 1 according to the present embodiment accounts for each device on the user side and accounting processing on the accounting processing service provider side via a communication network 2 such as the Internet and VPN (Virtual Private Network). It is configured by being connected to the device 10. Although only one user is shown in FIG. 1 for simplification of the description, the accounting processing apparatus 10 can connect to a plurality of users via the communication network 2.

The user is, for example, an expert such as a tax accountant or an accountant, or a corporation or an individual who directly performs accounting processing, and has at least a reading device 3 and an information terminal 4.

The reading device 3 is an optical device such as a scanner or a camera, and is a device that can capture a voucher as image data. The word "voucher" in the present embodiment and claims is used for receipts, receipts, other receipts, invoices, invoices, bankbooks of financial institutions, documents certifying money transfer in accounting, and electronic money. It shall also include transaction information using IC cards such as money.

The information terminal 4 is, for example, a personal computer (hereinafter referred to as a PC), a smartphone, a tablet PC, and a mobile terminal such as a mobile phone. The information terminal 4 has an input unit 4a that accepts user operations and a display unit 4b that can display at least web information.

The user can acquire the image data of the voucher by the reading device 3 and transmit it to the accounting processing device 10 by the information terminal 4, and can also send and receive information to and from the accounting processing device 10. Although the reading device 3 and the information terminal 4 are shown as separate bodies in FIG. 1, the reading device 3 and the information terminal 4 may be integrated as in a mobile terminal with a camera. Further, the user does not have to own the reading device 3, and for example, the image data of the voucher read by an external reading device may be acquired via an e-mail or a web.

On the other hand, the accounting processing service provider (hereinafter, also simply referred to as a service provider) is a business operator that provides accounting processing services by so-called cloud computing, and is a person who manages the accounting processing device 10.

The accounting processing device 10 has one or a plurality of servers (computers) that execute journal processing based on a program, and functionally mainly, a journal element extraction unit 11 that extracts journal elements as text from the image data of a voucher. , The journal element determination unit 12 that determines the validity of the extracted journal element, and the first RPA execution unit that executes the process using the text of the journal element in the pre-registered processing procedure (first RPA workflow). 13 (Automatic processing unit), Journal unit 14 that automatically performs journals based on journal elements, Journal judgment unit 15 that determines the consistency of journal results, Predetermined processing work for journal results (second RPA workflow) It has a second RPA execution unit 16 that executes the above. In this embodiment, RPA (Robotic Process Automation) identifies an application or system screen displayed on the screen of the display unit 4b or the like of the information terminal 4 by a software robot, and performs the same processing as the user's operation. It means to perform the procedure automatically.

Further, the accounting processing device 10 generates an application database 20 (hereinafter, the database is referred to as “DB”), a journal element extraction AI, and a journal AI in which information for which the user has applied for expenses via the information terminal 4 is stored. It has a learning system 21 and an RPA setting unit 22 for setting a first RPA and a second RPA.

The journal element extraction unit 11 has a function of receiving the image data of the voucher from the user and extracting the journal element from the image data. Specifically, the journal element extraction unit 11 has a journal element extraction AI for extracting the journal element from the image data of the voucher, and the journal element can be extracted from the image data of the voucher by the plurality of journal element extraction AIs. Is.

Journal element extraction AI specifies the part (position) corresponding to the journal element from the image data of the voucher, and extracts the journal element corresponding to the content of the specified part as text. That is, the journal element extraction AI specifies an area including a portion corresponding to the journal element from the image data by machine learning in the learning system 21, and outputs the journal element corresponding to the content of the specified portion as text. It is AI who learned that. For example, the journal element extraction AI specifies a date part, an amount part, a business partner part, and a description part in the image data of a voucher, and recognizes a number in the date part and the amount part to recognize the date part and the amount amount. The text is output, and the text of the business partner and the description is output by recognizing the characters in the part corresponding to the business partner and the part corresponding to the description.

Further, the journal element extraction unit 11 not only recognizes characters and numbers by the journal element extraction AI and outputs text, but also, for example, when only a logo mark, an imprint, or a telephone number is extracted, it is not shown. It may be output as the text of the business partner name by searching from the company information DB in which the company information is stored in advance or by searching the information published on the Internet. In addition, when there is no description of the address such as a receipt, the user name that sent the image data may be output as the text of the transaction source, or the company that is the customer of the user is set in advance. The image data transmitted from the user may output the set company name as the text of the transaction source.

The journal element extraction AI is an image recognition AI that specifies a part (position) corresponding to the journal element from the image data of the voucher, and a character recognition AI that extracts the journal element corresponding to the content of the specified part as text. It may be composed of two AIs.

Journal elements include, for example, date, amount, business partner, description (including proviso and product name), business source (including address), and the corresponding numbers, letters, graphics (for example, logo mark, imprint, and other companies). There is a identifiable pattern) and the appearance of the voucher (eg, the size and color of the passbook or receipt).

For example, for a date, the journal element extraction unit 11 specifies numbers before, after, and above and below characters such as "date", "year", "month", and "day" and symbols such as "/". Amounts include symbols such as "¥", product names, "amount", "deposit", "subtotal", "total", "tax", "change", "discount", "yen", "payment", "deposit", "balance", etc. Identify the numbers before, after, and above and below the characters related to. Then, in the present embodiment, as for the amount of money, a journal element is extracted for each character or number related to the amount of money, for example, "total 1000 yen" or "change 50 yen". The journal element extraction unit 11 can also recognize the meaning of the extracted characters, and for example, the "subtotal" can specify the accounting relationship such as the total amount of the prices of the products.

For business partners, the characters before and after the characters such as "Co., Ltd.", "Co., Ltd." and "(F)", the logo mark, telephone number, and the appearance of the voucher are specified and based on this information. Identify the part that corresponds to your company or personal name. For the abstract, specify the character part following the character such as "However". For the trading source, specify the part of the character before the character such as "sama".

Note that the journal elements are not limited to this, and the numbers, letters, and figures used to extract the journal elements are not limited to this. For example, if the voucher contains the quantity of purchased items, the quantity may be included as a journal entry factor, and if information such as the names and number of attendees is stated, the attendees and the number of attendees. May be included as a journal element. In addition, a number (corporate number, business office number) set to identify each company may be extracted.

In addition, the journal element extraction unit 11 can also output the reliability of the journal element output by the journal element extraction AI (hereinafter referred to as the journal element reliability Re). The journal element reliability Re is the accuracy of the judgment of the journal element extraction AI, and can be expressed as a percentage, for example. That is, the higher the journal element reliability Re, the higher the validity of the journal element text output by the journal element extraction AI, and the lower the journal element reliability Re, the lower the validity. The journal element reliability Re is not limited to the percentage display, and is displayed in a two-step evaluation in which a predetermined threshold value or more is ○ and less than a predetermined threshold value is ×, or two threshold values are set and a three-step evaluation of ○, △, and × is performed. It may be displayed.

The journal element determination unit 12 connected to the journal element extraction unit 11 has a function of determining the validity of the text output from the journal element extraction unit 11 as a journal element. The journal element determination unit 12 determines the validity of the text of the journal element based on the journal element reliability Re output by the journal element extraction unit 11 and the matching with the information regarding the journal element input by the user.

In the determination based on the journal element reliability Re of the journal element determination unit 12, for example, the first threshold value T1 corresponding to the journal element reliability Re is set, and the journal element reliability Re is equal to or higher than the first threshold T1. If so, the extracted journal element is judged to be valid. On the other hand, if the journal element reliability Re is less than the first threshold value T1, it is determined that the extracted journal element is not valid.

Further, in the determination based on the matching with the information about the journal element input by the user of the journal element determination unit 12, the journal element extracted by the journal element extraction unit 11 and the information of the journal element stored in the application DB 20 are combined. If they match and match, it is judged to be valid, and if they are inconsistent, it is judged to be invalid. The consistency of this matching is not limited to perfect matching, and for example, the degree of matching may be calculated and determined to be appropriate when the matching is to a certain extent. This degree of agreement may be expressed as a percentage in the same manner as the journal element reliability Re, or may be evaluated on a two-point scale of ○ or × or on a three-point scale of ○, Δ, or ×.

Further, the journal element determination unit 12 generates warning display data different from the normal display for invalid journal elements, and causes the display unit 4b of the user's information terminal 4 to present the warning display data.

The application DB 20 stores the image data of the voucher read by the user by the reading device 3 and the information that the user transmits the contents of the voucher to the accounting processing device 10 via the information terminal 4. The image data of the voucher and the content information of the voucher input by the user are associated with each other, and the journal element determination unit 12 collates the image data of the associated voucher with the content information.

The first RPA execution unit 13 connected to the journal element determination unit 12 has a function of executing the first RPA workflow registered in the RPA setting unit 22. The first RPA workflow is a processing work for a journal element, and in the present embodiment, input processing for automatic journal is performed. Specifically, the first RPA execution unit 13 copies the journal elements determined to be valid by the journal element determination unit 12 displayed on the screen of the display unit 4b of the information terminal 4, and is used for automatic journal entry. Repeat the work of pasting into the journal element input field of, and after the input field is filled, select the button to execute automatic journal.

The journal unit 14 has a function of outputting an account item according to the input journal element. The output of the account item corresponding to this journal element is performed by, for example, the journal AI. The journal AI is an automatic journal AI that has learned in advance to output an account for a journal element by machine learning in the learning system 21. The account item output in the journal entry unit 14 is output to the journal entry determination unit 15 as a journal entry result together with the journal entry element.

In addition, the journal unit 14 can also output the reliability of the journal result output by the journal AI (hereinafter referred to as the journal reliability). The journal reliability Rj is the accuracy of the judgment of the journal AI, and can be expressed as a percentage, for example. That is, the higher the journal reliability Rj, the higher the validity of the journal result output by the journal AI, and the lower the journal reliability Rj, the lower the validity.

The journal determination unit 15 connected to the journal unit 14 has a function of determining the validity of the journal result output from the journal unit 14. The journal determination unit 15 determines the validity of the journal result based on the journal reliability Rj output by the journal unit 14.

In the determination based on the journal reliability Rj of the journal determination unit 15, for example, if the second threshold value T2 corresponding to the journal reliability Rj is set and the journal reliability Rj is equal to or higher than the second threshold T2, the journal result is Is judged to be valid. On the other hand, if the journal reliability Rj is less than the second threshold value T2, it is determined that the journal result is not valid.

Further, the journal determination unit 15 generates warning display data different from the normal display for an invalid journal result, and causes the display unit 4b of the user's information terminal 4 to present it.

The second RPA execution unit 16 connected to the journal determination unit 15 has a function of executing the second RPA workflow registered in the RPA setting unit 22. The second RPA workflow is a processing work for the journal entry result. For example, the journal entry result determined to be appropriate by the journal entry determination unit 15 displayed on the screen of the display unit 4b of the information terminal 4 is copied and mail software is used. After pasting it in the new e-mail text of, enter a specific e-mail address (for example, the e-mail address of the accounting staff) in the destination and send it.

The learning system 21 has a function of learning the above-mentioned journal element extraction AI, journal AI, and the like, and supplying the learned AI. Specifically, the learning system 21 generates a journal element extraction AI by performing machine learning (so-called deep learning) based on teacher data (learning data) consisting of voucher image data and journal elements included in the image data. .. Further, the learning system 21 generates a journal AI by performing machine learning based on teacher data (learning data) composed of a journal element and an account item corresponding to the journal element.

The RPA setting unit 22 has a so-called RPA tool, and has a function of storing an operation procedure performed by a worker such as a user as an RPA workflow. For example, in the present embodiment, the user executes the operations of the first RPA workflow and the second RPA workflow described above from the information terminal 4 in advance via the RPA setting unit 22, and the first RPA workflow and the first RPA workflow It remembers 2 RPA workflows. Since the RPA setting unit 22 stores the user's operation through the display unit 4b, if it can be displayed on the display unit 4b, the RPA setting unit 22 straddles a plurality of systems or a plurality of applications such as an accounting processing system and mail software. Processing is also feasible.

Here, with reference to FIG. 2, a flowchart showing the flow of accounting processing executed by the accounting processing apparatus 10 is shown, and the accounting processing method according to the present embodiment will be described in detail below with reference to the same flowchart. The accounting process starts when the image data of the voucher from the user is received.

First, as step S1, the journal element extraction unit 11 extracts the journal element from the received voucher image data by the journal element extraction AI. Then, the journal element extraction unit 11 outputs the text of the extracted journal element and the journal element reliability Re.

In step S2, the journal element determination unit 12 determines whether or not the extracted journal element is appropriate. Specifically, it is determined whether or not the journal element reliability Re is equal to or higher than the first threshold value T1 and matches with the information regarding the journal element input by the user stored in the application DB 20. The journal element determination unit 12 of the present embodiment makes both a determination based on the journal element reliability Re and a determination based on the matching, but may determine only one of them. If the determination result is false (No), that is, if the journal element reliability Re is less than the first threshold value T1, or if the matching result is inconsistent, the process proceeds to step S3.

In step S3, the journal element determination unit 12 displays a warning display on the display unit 4b of the user's information terminal 4, and prompts the user (person) to confirm and correct the extracted journal element.

On the other hand, if the determination result in step S2 is true (Yes), that is, if the journal element reliability Re is equal to or higher than the first threshold value T1 and the matching result is consistent, the process proceeds to step S4.

In step S4, the first RPA execution unit 13 performs the above-mentioned input processing for automatic journal entry by the first RPA workflow. In step S3, this input process is manually performed according to the warning display.

Then, in step S5, the journal unit 14 outputs the journal result for the input journal element and the journal reliability Rj by the journal AI.

In the next step S6, the journal determination unit 15 determines whether or not the output journal result is valid. Specifically, it is determined whether or not the journal reliability Rj is equal to or higher than the second threshold value T2. If the determination result is false (No), that is, if the journal entry reliability Rj is less than the second threshold value T2, the process proceeds to step S7.

In step S7, the journal determination unit 15 displays a warning display on the display unit 4b of the user's information terminal 4, and prompts the user (person) to confirm and correct the output journal result.

On the other hand, if the determination result in step S6 is true (Yes), that is, if the journal reliability Rj is equal to or higher than the second threshold value T2, the process proceeds to step S8.

In step S8, the second RPA execution unit 16 performs the above-mentioned journal entry completion email transmission process by the second RPA workflow, and ends the routine. In step S7, this mail transmission process is manually performed according to the warning display.

In this way, the accounting processing device 10 extracts the journal element from each image data by executing the routine every time the image data of the voucher is acquired, and automatically automatically journals if it is a valid journal element. Is executed, and if this journal result is valid, it is possible to automatically send an email.

Next, with reference to FIGS. 3 and 4, explanatory diagrams showing an example of the transition of accounting processing executed by the accounting processing apparatus according to the embodiment of the present invention are shown, and the voucher is based on these figures. The specific transition from the image data of the above to the sending of the journal entry completion email will be described. The journal element extraction field A, the journal element input field B, and the journal result field C shown in FIGS. 3 and 4 are display examples displayed on the display unit 4b of the user's information terminal 4.

The journal element extraction column A in FIG. 3 is a display example when the journal element is extracted from the image data of the receipt. The original image of the receipt scanned on the left side of the screen and the extraction result of the journal element in the center of the screen are displayed on the screen. The application details applied by the user are displayed on the right side. That is, the journal element extraction column A is a display example at the time of step S2 in FIG.

Specifically, as shown by the dotted line frame in the original image of the receipt, the journal element extraction unit 11 specifies the part 30a corresponding to the date, the part 31a corresponding to the total amount, and the part 32a corresponding to the description. However, these

enlarged views

30b, 31b, and 32b are displayed in the extraction result. It should be noted that this dotted line does not have to be actually displayed on the display unit 4b.

The extraction results include the date "October 10, 2019", the amount "1,080 yen", and the description "○○ Cafe", which are the journal elements converted into text by the journal element extraction AI on the corresponding enlarged view. Is described. Further, next to the journal element, the journal element reliability Re is also written. In FIG. 3, the date journal element is "90%", the amount journal element is "85%", and the description journal element is ". 82% "is shown. For example, in FIG. 3, the first threshold value T1 is set to 80%, and the journal element reliability Re is equal to or higher than the first threshold value T1.

Further, in FIG. 4, the date column 30c, the amount column 31c, and the description column 32c are shown as the application contents corresponding to the journal elements of the extraction result, and the date column 30c shows "October 10, 2019", the amount. "1,080 yen" is described in the column 31c, and "OO cafe" is described in the description column 32c. Next to the journal element, the degree of matching of the matching result is displayed as a percentage. In FIG. 4, the journal element of the date is "100%" and the journal element of the amount is "98%" (because only ¥ is inconsistent). ), The journal entry element of the description is shown as "85%" (because only the corporation does not match). For example, in FIG. 4, the threshold value of the degree of matching is set to 80%, and all the collation results are equal to or higher than this threshold value and are consistent to some extent.

As described above, in the display example shown in the journal element extraction column A of FIGS. 3 and 4, the determination result in the journal element determination unit 12 is true (Yes). If any of the journal elements is not valid, the characters of the journal elements that are not valid in the extraction result are displayed in red or a flag is displayed as a warning display. The following flow is the same for FIGS. 3 and 4.

Next, the first RPA execution unit 13 copies each journal element according to the first RPA workflow, pastes it into the corresponding input field of the journal element input field B, and when each input field is filled, the journal execution button. Select.

Then, based on the entered journal element, the result of journal entry by the journal AI of the journal unit 14 and the journal reliability corresponding to each account are displayed as shown in the journal result column C of FIGS. 3 and 4. Specifically, the debit includes the account "Meeting fee", amount "1,000", reliability "82%", "temporary consumption tax", amount "80", and reliability "90%". The credit is credited with the account "cash", the amount "1,080", and the reliability "95%". For example, in FIGS. 3 and 4, the second threshold value T2 is set to 80%, and the journal entry reliability Rj is equal to or higher than the second threshold value T2.

Therefore, in the display example shown in the journal entry result column C of FIGS. 3 and 4, the determination result in the journal entry determination unit 15 is true (Yes). If any of the journal entry results is not valid, for example, the characters of the inappropriate account, amount, or reliability are displayed in red or a flag is displayed as a warning display. Further, a pull-down menu is set for each account, and for example, when the triangle symbols in FIGS. 3 and 4 are selected, other account candidates may be listed.

Next, the second RPA execution unit 16 pastes it into a new mail text of the mail software according to the second RPA workflow, and then inputs a specific mail address in the destination and sends it.

In this way, the accounting processing system 1 extracts the journal element from the image data of the voucher, and if the extracted journal element is highly valid, it automatically shifts to the automatic journal according to the first RPA workflow. Therefore, if the extracted journal element is highly valid, it can be automated including the processing using the journal element, and in the present embodiment, the automatic journal is automatically performed without human check or operation. Can be migrated to.

In addition, automatic journalizing is performed according to the extracted journalizing elements, and if the journalizing result is highly valid, an email is automatically sent to the person in charge according to the second RPA workflow. As a result, when the validity of the journal entry result is high, it is possible to automate the process using the journal entry result, and in the present embodiment, an email to the person in charge is automatically sent without human check or operation. You can send.

Furthermore, if the validity of the extracted journal element is low or the validity of the journal result is low, the user is warned to that effect and urged to confirm and correct it. By issuing a warning display only when the validity of the information is low in this way, it is possible to ensure the accuracy of the information while minimizing the checks by humans.

From the above, according to the accounting processing system 1, the accounting processing method, and the accounting processing program in the present embodiment, the confirmation work and the simple work by the accounting staff of the user are reduced, and the overall work efficiency of the accounting processing is improved. be able to.

The description of the embodiment of the present invention is completed above, but the aspect of the present invention is not limited to this embodiment.

For example, in the above embodiment, the first RPA workflow performs an input process for automatic journal entry, and the second RPA workflow performs a process of sending the journal entry result by e-mail to a specific destination, but the first RPA workflow And the automatic processing set as the second RPA workflow is not limited to this, and other processing procedures may be registered.

Further, in FIG. 3, an example of this embodiment is described using a receipt as an example, but the voucher is not limited to the receipt, and can be similarly applied to other vouchers such as receipts.

Further, the accounting processing system 1 of the above embodiment is a system capable of consistently and automatically processing from the extraction of journal elements to the journals, but can be applied to, for example, a system having only extraction of journal elements and only automatic journals. is there.

1 Accounting processing system 2 Communication network 3 Reading device 4 Information terminal

4a Input unit

4b Display unit 10 Accounting processing device 11 Journal element extraction unit 12 Journal element judgment unit 13 1st RPA execution unit 14 Journal unit 15 Journal judgment unit 16 2nd RPA Execution Department 20 Application Database 21 Learning System 22 RPA Setting Department

Claims

Journal element extraction for outputting journal elements as text from voucher image data A journal element extraction unit that extracts journal elements as text from voucher image data and outputs the reliability of the extraction.
An automatic processing unit that can automatically execute processing using the text of the journal element by a processing procedure registered in advance, and
When the reliability output from the journal element extraction unit is equal to or higher than a predetermined threshold value, a determination unit that determines that the input information is valid and executes processing by the automatic processing unit, and a determination unit.
Accounting system with.
Journal element extraction unit for extracting journal elements as text from image data of vouchers, and a journal element extraction unit that extracts journal elements as text from image data of vouchers by AI.
An automatic processing unit that can automatically execute processing using the text of the journal element by a processing procedure registered in advance, and
A journal element data storage unit that stores the text of the journal element included in the image data of the voucher input by a means different from the journal element extraction unit.
When the text of the journal element extracted from the journal element extraction unit and the text of the journal element stored in the journal element data storage unit corresponding to the journal element are matched and matched, the input information is valid. A determination unit that determines that the data is true and executes processing by the automatic processing unit.
Accounting system with.
For the input journal element, the journal unit that outputs the journal result and outputs the reliability of the journal result by the journal AI for outputting the journal result according to the journal element.
An automatic processing unit that can automatically execute processing using the journal results by a processing procedure registered in advance, and
When the reliability output from the journal entry unit is equal to or higher than a predetermined threshold value, a determination unit that determines that the input information is valid and executes processing by the automatic processing unit, and a determination unit.
Accounting system with.
The accounting system according to any one of claims 1 to 3, which generates warning display data when the determination unit determines that the input information is not valid.
The accounting processing system according to claim 1 or 2, wherein the automatic processing unit can automatically execute a process of performing automatic journalizing using the journal element.
The accounting processing system according to claim 3, wherein the automatic processing unit can automatically execute a process of sending the journal entry result by e-mail to a specific destination.
Journal element extraction process for outputting journal elements as text from voucher image data A journal element extraction process that extracts journal elements as text from voucher image data and outputs the reliability of the extraction.
A determination step of determining that the text of the extracted journal element is appropriate when the reliability output from the journal element extraction step is equal to or higher than a predetermined threshold value.
An automatic processing step that automatically executes a process using the text of the journal element according to a pre-registered processing procedure when it is determined that the journal element extracted in the determination step is appropriate.
The accounting method that the computer performs.
An accounting program for causing a computer to execute the accounting method according to claim 7.