Document review method, device, equipment and medium based on RPA and AI

Through RPA and AI-based document review methods, automatic review and classification of document contents is solved, and the problem of time-consuming and laborious analysis of user review results is achieved, efficient document review and risk management is achieved, and user experience and review efficiency are improved.

CN111753090BActive Publication Date: 2024-11-22BEIJING LAIYE NETWORK TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010610561.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-06-30
Publication Date
2024-11-22
Estimated Expiration
2040-06-30

AI Technical Summary

Technical Problem

The existing document review technology can only be limited to content review. Users need to manually analyze the review results, which is time-consuming and labor-intensive, and the reading review results are inefficient.

Method used

Using RPA and AI-based document review methods, the document content is automatically reviewed and classified, and through legal risk review and grammatical order review, key information is extracted, and whether the correlation factor is a risk factor, and the risk results and levels are determined based on the risk factor. Generate a clear-level audit report.

Benefits of technology

Save time for users to read review results, improve user experience, provide clear-level focus, reduce manual review time, improve document review efficiency, and promote business processes through risk warning information.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111753090B_ABST
    Figure CN111753090B_ABST
Patent Text Reader

Abstract

The embodiment of the present invention discloses a document review method, device, equipment and medium based on RPA and AI, the method comprising: S1, reviewing the content of the document; S2, classifying the content of the document according to the review results. By adopting the above technical solution, the review of the content of the document can be automatically realized based on RPA and AI technology, and the content of the document can be classified according to the review results, providing a clear hierarchy of emphasis for users to read the review results, solving the problem of time-consuming and laborious reading of the review results by users, and improving the user experience.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a document review method, device, equipment and medium based on RPA and AI. Background Art

[0002] RPA (Robotic process automation) simulates human operations on computers through specific "robot software" and automatically executes process tasks according to rules. RPA has unique advantages: low code and non-intrusive. Low code means that RPA does not require a high level of IT to operate, and business personnel who do not understand programming can also develop processes; non-intrusive means that RPA can simulate human operations without the need for software system open interfaces. However, traditional RPA has certain limitations: it can only be based on fixed rules and has limited application scenarios. With the continuous development of AI (Artificial Intelligence) technology, the deep integration of RPA and AI has overcome the limitations of traditional RPA. RPA and AI = Hand work + Head work, which is greatly changing the value of labor.

[0003] In the process of processing tasks, RPA generally applies natural language processing (NLP) and optical character recognition (OCR) technologies. Among them, NLP is a subfield of AI that can be used to understand and classify text content. OCR refers to the process in which electronic devices, such as scanners or digital cameras, examine characters printed on paper, determine their shapes by detecting dark and light patterns, and then use character recognition methods to translate the shapes into computer text.

[0004] At present, after using OCR and NLP technology to identify the content of documents, there are products on the market that can realize the content review of paper and electronic documents. For example, it can review typos in the document content, or review its risks according to the rules and regulations required by the company. However, the document review method provided by the existing technology is limited to the review stage of the document content, and the results of the review need to be further analyzed and sorted by the user. In addition, when reading the review results, the user still needs to read the document information in full. If the document content is large, the user will spend a lot of time and energy, that is, the process of manually reading the review results is time-consuming and labor-intensive, and inefficient. Summary of the invention

[0005] The embodiments of the present invention disclose a document review method, apparatus, device and medium based on RPA and AI, which realize automatic review of document content and classification according to the review results, saving the user's time in reading the review results.

[0006] In a first aspect, an embodiment of the present invention discloses a document review method based on RPA and AI, the method comprising:

[0007] S1. Review the content of the document;

[0008] S2. Classify the document contents according to the audit results.

[0009] Optionally, the step S1 specifically includes:

[0010] S11. Conduct a legal risk review of the document content, or,

[0011] S12. Review the grammar and word order of the document content.

[0012] Optionally, the step S11 specifically includes:

[0013] S111, extracting key information from the document content;

[0014] S112. Conduct a legal risk review of the document content based on the key information.

[0015] Optionally, the step S112 specifically includes:

[0016] S1121. Searching for a correlation factor corresponding to the key information from a preset knowledge graph;

[0017] S1122, determining whether the associated factor is a risk factor;

[0018] S1123. If the associated factor is a risk factor, determine the risk result according to the weight information corresponding to the risk factor.

[0019] Optionally, the step S1122 specifically includes:

[0020] If different versions of the document exist, the contents of the correlation factors in the different versions are compared;

[0021] If the contents corresponding to the correlation factors are different, the correlation factors are used as risk factors.

[0022] Optionally, the step S1122 specifically includes:

[0023] Matching the content corresponding to the correlation factor with the risk content in the label library;

[0024] If the match is successful, the associated factor is used as a risk factor.

[0025] Optionally, the step S1123 specifically includes:

[0026] If the associated factor is a risk factor, the weight information corresponding to each risk factor is multiplied, and the product is used as the risk result.

[0027] Optionally, the step S2 specifically includes:

[0028] S21. Determine the risk level corresponding to the risk result;

[0029] S22: classify the document content according to the risk level.

[0030] Optionally, the step S22 specifically includes:

[0031] For each risk level of document content, the document content is classified according to the key information and an audit report is generated.

[0032] Optionally, the method further includes:

[0033] Generate risk warning information according to the risk level; or,

[0034] The risk content and its risk results are sent to the lawyer, and the modification method provided by the lawyer is used as a prompt information of the risk content.

[0035] Optionally, after step S1, the method further includes:

[0036] If the document contains content that matches the company's prohibition conditions, the document will be returned and a reminder email will be sent to the user.

[0037] In a second aspect, an embodiment of the present invention further provides a document review and classification device based on RPA and AI, the device comprising:

[0038] An audit module, configured to audit the document content;

[0039] The classification module is configured to classify the document content according to the audit results.

[0040] Optionally, the audit module includes:

[0041] The legal risk review submodule is configured to conduct a legal risk review on the document content, or,

[0042] The grammar and word order review submodule is configured to review the grammar and word order of the document content.

[0043] Optionally, the legal risk review submodule includes:

[0044] A key information extraction unit, configured to extract key information from the document content;

[0045] The legal risk review unit is configured to perform a legal risk review on the document content based on the key information.

[0046] Optionally, the legal risk review unit includes:

[0047] A correlation factor search subunit is configured to search for a correlation factor corresponding to the key information from a preset knowledge graph;

[0048] A risk factor determination subunit, which determines whether the associated factor is a risk factor;

[0049] The risk result determination subunit determines the risk result according to the weight information corresponding to the risk factor if it is determined that the associated factor is a risk factor.

[0050] Optionally, the risk factor determination subunit is specifically configured as follows:

[0051] If different versions of the document exist, the contents of the correlation factors in the different versions are compared;

[0052] If the contents corresponding to the correlation factors are different, the correlation factors are used as risk factors.

[0053] Optionally, the risk factor determination subunit is specifically configured as follows:

[0054] Matching the content corresponding to the correlation factor with the risk content in the label library;

[0055] If the match is successful, the associated factor is used as a risk factor.

[0056] Optionally, the risk result determination subunit is specifically configured as follows:

[0057] If the associated factor is a risk factor, the weight information corresponding to each risk factor is multiplied, and the product is used as the risk result.

[0058] Optionally, the classification module includes:

[0059] a risk level determination unit, configured to determine a risk level corresponding to the risk result;

[0060] The classification unit is configured to classify the document content according to the risk level.

[0061] Optionally, the classification unit is specifically configured as:

[0062] For each risk level of document content, the document content is classified according to the key information and an audit report is generated.

[0063] Optionally, the device further comprises:

[0064] a warning information generating module, configured to generate risk warning information according to the risk level; or,

[0065] The lawyer suggestion module is configured to send the risk content and its risk results to the lawyer, and use the modification method provided by the lawyer as prompt information of the risk content.

[0066] Optionally, the device further comprises:

[0067] The document return module is configured to return the document and send a prompt email to the user if it is identified that the document content contains content that matches the company's prohibition conditions after reviewing the document content.

[0068] In a third aspect, an embodiment of the present invention further provides a computer device, including:

[0069] A memory storing executable program code;

[0070] a processor coupled to the memory;

[0071] The processor calls the executable program code stored in the memory to execute part or all of the steps of the RPA and AI-based document review method provided in any embodiment of the present invention.

[0072] In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium storing a computer program, wherein the computer program includes instructions for executing part or all of the steps of the RPA and AI-based document review method provided in any embodiment of the present invention.

[0073] The technical solution provided by the embodiment of the present invention can automatically realize the review of document content based on RPA and AI technology, and classify the document content according to the review results, providing a clear hierarchy of focuses for users to read the review results, thereby solving the problem of users spending time and effort in reading the review results in their entirety, and improving the user experience.

[0074] Compared with the prior art, the invention includes:

[0075] 1. Based on RPA and AI technology, the document content can be automatically classified according to the review results, avoiding the problem of users having to spend time and effort reading the review results of the entire document, reducing the time users spend reviewing documents, and improving the timeliness of user feedback.

[0076] 2. By conducting a legal risk review of the document content and classifying the document content according to the risk level corresponding to the risk results, the audit report for users to read the document can be provided with a clear focus, thereby improving the user experience.

[0077] 3. Automatically generate risk warning information based on the risk level. This warning information can be used as a reference and basis for the user's next operation, thus helping to advance the business process of the document.

[0078] 4. When determining whether a correlation factor is a risk factor, the historical tracing method is adopted, that is, by comparing the contents of the correlation factors in different versions, and taking the correlation factors with different contents as risk factors. This setting takes into account the risk impact brought about by changes in key factors in the document, and provides a reliable and effective analysis basis for the legal risk review of the document.

[0079] 5. Send the risk content and its risk results to the lawyer, and use the lawyer's advice as the prompt information corresponding to the risk content. This setting can make the prompt information more targeted and professional.

[0080] 6. For risky document content, the risk result is the product of the weight values ​​corresponding to each risk factor. Since different users have different focuses on key information, the weight information of each risk factor can be set by the user, so that the risk result can meet the user's personalized needs.

[0081] 7. If the document is identified to contain content that matches the company's prohibited conditions, the document will be returned and a prompt email will be sent to the user to remind the user that the document does not meet the review conditions, saving time and resources for subsequent manual review of the document. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0083] Figure 1 is a schematic diagram of a knowledge graph provided by an embodiment of the present invention;

[0084] Figure 2It is a flowchart of a document review method based on RPA and AI provided by an embodiment of the present invention;

[0085] Figure 3 It is a schematic diagram of a process of conducting a legal risk review of document content based on key information provided by an embodiment of the present invention;

[0086] Figure 4 It is a structural diagram of a document review device based on RPA and AI provided by an embodiment of the present invention;

[0087] Figure 5 It is a structural schematic diagram of a computing device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0088] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0089] It should be noted that the terms "including" and "having" and any variations thereof in the embodiments of the present invention and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products or devices.

[0090] In the description of the present invention, "knowledge graph" is a graph-based data structure consisting of nodes (Point) and edges (Edge). In the knowledge graph, each node represents an "entity" that exists in the real world, and each edge is a "relationship" between entities. Knowledge graph is the most effective way to represent relationships. In layman's terms, knowledge graph is a large-scale semantic network obtained by connecting all different types of information together. Among them, entity is the most basic element in the knowledge graph, which refers to something that is distinguishable and exists independently, for example, Figure 1 "Wang Tao", "Guarantor" and "ID number" in the provided knowledge graph are all entities in the knowledge graph.

[0091] In the description of the present invention, “association factors” may be other entities in the knowledge graph that have a semantic relationship with a certain entity, such as Figure 1As shown, the association factors of "Wang Tao" in the knowledge graph include "mobile phone number", "ID number", "education", "taxation" and "personal credit record". In the technical solution provided by the embodiment of the present invention, if the content corresponding to the "association factor" has changed in different versions of the document, the association factor is determined as a "risk factor", that is, some key information of legal risk exists in the document.

[0092] In order to describe the content of the present invention more clearly and clearly, the technical solution provided by the embodiment of the present invention is briefly introduced below.

[0093] When reviewing a document, the semantic understanding of the document content can be performed based on NLP. Specifically, key information can be extracted from the document content, and the document content can be reviewed based on the key information. The review in the embodiment of the present invention can be a legal risk review, that is, reviewing whether the various terms of the contract meet the legal expression requirements, whether some key information in the document content has legal risks, etc. In the solution provided by the embodiment of the present invention, the legal risk review of the document content based on key information can be achieved in the following ways:

[0094] Search for the correlation factor corresponding to the key information from the preset knowledge graph; determine whether the correlation factor is a risk factor; if the correlation factor is a risk factor, determine the risk result based on the weight information corresponding to the risk factor.

[0095] Among them, judging whether the correlation factor is a risk factor can be achieved in the following ways:

[0096] As an optional implementation, if there are different versions of the document, the contents of key information in different versions are compared; if there is a difference in the content corresponding to the key information, the key information is used as a risk factor.

[0097] Specifically, taking contract A as an example, if contract A has two versions, A1 and A2, where A1 is the version before the upgrade and A2 is the upgraded version after the upgrade. For a contract, the contract name is key information, and the associated factors corresponding to the key information include but are not limited to: "legal person", "equity relationship" and "controlling relationship", etc. Compared with the A1 version of the contract, if the content of the above associated factors has changed in the A2 version, the above associated factors will be used as risk factors.

[0098] As another optional implementation method, a label library can be pre-established. The templates stored in the label library are all risk content with risks. The content corresponding to the associated factor is matched with the risk content in the label library. If the match is successful, the associated factor is used as the risk factor. The templates in the label library can be provided by the user or set according to the company's regulations.

[0099] Taking the contract as an example, if the label library contains the risk content "payment time exceeds the time threshold set by the company", for the association factor "payment time" in the contract, if the corresponding contract content is 5 days longer than the payment time threshold set by the company, it means that the content corresponding to the association factor matches the risk content in the label library. At this time, the association factor can be used as a risk factor.

[0100] The calculation scheme provided in this embodiment can multiply the weight information corresponding to each risk factor after determining the risk factor, and use the product as the risk result. Since different users have different concerns about different audit items, the weight information can be set by the user. In this way, the risk result obtained can meet the personalized needs of the user.

[0101] Furthermore, the technical solution provided by the embodiment of the present invention can classify the document content according to the risk result after the risk result is obtained. This can be achieved in the following ways:

[0102] Determine the risk level corresponding to the risk result and classify the document content according to the risk level.

[0103] Specifically, risks can be pre-classified into three risk levels: high, medium, and low, and each level corresponds to a risk range. This embodiment can determine the risk level corresponding to the risk result according to the range of the value of the risk result. After the risk level is determined, the document content can be classified according to the risk level to generate an audit report. For example, in the audit report of a contract, the risky content in the contract can be classified as high risk, medium risk, and low risk, so that customers can be more focused in the process of reading the audit report, thereby improving the user experience.

[0104] In addition, the above classification of document content according to risk level can be used as a primary category. On this basis, for the document content of each risk level, the document content can be classified according to key information. That is, on the basis of the primary category, the document content is classified into a secondary category to achieve hierarchical management of file content, so that the audit report has clear levels and improves the reading experience of users.

[0105] Furthermore, after determining the risk level of the document content, risk warning information may be generated according to the risk level, wherein the risk warning information may include the specific risk level of the document and the cause of the risk, so as to remind the user.

[0106] For example, the risk content and its risk results can also be sent to the lawyer, and the modification method provided by the lawyer can be used as the prompt information of the risk content. This arrangement can make the obtained prompt information more targeted and professional.

[0107] In addition, in the traditional document review process, the system generally conducts automatic review first and then manual review. The technical solution provided by the embodiment of the present invention automatically returns the document if the system identifies that the document content contains content that matches the company's prohibited conditions, and sends a prompt email to the user to remind the user that the contract does not meet the conditions, thereby saving time and resource consumption for subsequent manual review.

[0108] The contents provided by the embodiments of the present invention will be described in detail below in conjunction with the flowchart.

[0109] Embodiment 1

[0110] See also Figure 2 , Figure 2 This is a flowchart of a document review method based on RPA and AI provided by an embodiment of the present invention. This method is typically applied in the review process of contracts and legal documents, and can be executed by a document review device based on RPA and AI. The device can be implemented in software and / or hardware, and the embodiment of the present invention does not limit this. Figure 2 As shown, the method provided in this embodiment specifically includes:

[0111] 110. Review the document content.

[0112] The review of the document content includes but is not limited to legal risk review, grammar, word order review, and typo review, etc. The technical solution provided in this embodiment mainly introduces the legal risk review process of the document content in detail.

[0113] For example, before conducting a content review, the OCR recognition method can be used to identify the document content, and the NLP technology can be used to understand the semantics of the document content. After completing the semantic understanding of the document content, the key information in the document content can be extracted, and the legal risk review of the document content can be conducted based on the key information.

[0114] Among them, key information is the core information in the document content. Key information can be identified by keyword matching. The matching template is configured according to user needs. For example, for a contract, information such as "contract name", "contract amount" and "supplier" can be used as templates. When a keyword matching the template content is identified in the contract, the content of the matching keyword in the contract can be used as key information.

[0115] In this embodiment, Figure 3 As shown in the figure, the legal risk review of the document content based on key information may include the following steps:

[0116] 210. Find the correlation factors corresponding to the key information from the preset knowledge graph.

[0117] Among them, the association factors can be other entities in the knowledge graph that have a semantic relationship with the key information entity. One key information can correspond to multiple association factors. Taking the contract as an example, "Contract Name" is the key information, and the association factors corresponding to the key information may include: "Legal Person", "Equity Relationship", "Controlling Relationship", "Blacklist Library", "Credit Reporting Agency", "Court Ruling Network Documents" and "Aging Records of Accounts Receivable within the Company".

[0118] Since there are generally multiple correlation factors corresponding to key information, a relatively comprehensive document content will be involved. That is, when conducting a legal risk review, this embodiment can provide a comprehensive analysis basis for the review of the legal risks of the document through the correlation factors.

[0119] 220. Determine whether the correlation factor is a risk factor. If so, execute step 230; otherwise, return to execute step 210.

[0120] As a common implementation method, determining whether a correlation factor is a risk factor can be performed in the following manner:

[0121] If there is only one version of the document, it is generally possible to determine whether the specific content of the correlation factor is on the blacklist of the enterprise. If it is on the blacklist, the correlation factor is used as a risk factor. Alternatively, the credit record corresponding to the correlation factor can be queried. If the credit record is abnormal, the correlation factor is used as a risk factor.

[0122] As an optional implementation, determining whether a correlation factor is a risk factor may also be performed in the following manner:

[0123] A label library is pre-established, and the templates stored in the label library are all risky content. The templates in the label library can be provided by users or set according to the company's regulations. For example, the module can be "payment time exceeds the time threshold set by the company", "payment ratio is not coordinated and does not meet the regular ratio", etc. In addition, since different users have different concerns about different audit items, users can add weight information for different risk content.

[0124] In this embodiment, if the content corresponding to the association factor matches the risk content in the label library, the association factor is determined as a risk factor. For example, for the association factor "payment time" in the contract, if the corresponding contract content is 5 days more than the payment time threshold set by the company, it means that the content corresponding to the association factor matches the risk content in the label library. At this time, the association factor "payment time" in the contract can be used as a risk factor, and the weight information corresponding to the risk factor can be obtained through the label library.

[0125] As another optional implementation, the risk factor may also be determined by a historical retrospective method, which may be:

[0126] If there are different versions of the document, the contents of the correlation factors in different versions will be compared; if there are differences in the contents corresponding to the correlation factors, the correlation factors will be used as risk factors. This setting takes into account the risk impact brought about by changes in key factors in the document and provides a reliable and effective analysis basis for the legal risk review of the document.

[0127] Specifically, taking Contract A as an example, after comparing the contents of the pre-upgrade version A1 and the post-upgrade version A2 of Contract A, if it is identified that the contract name has changed from "Shenzhen Investment Technology Co., Ltd." in version A1 to "Shenzhen Investment Holding Technology Co., Ltd." in version A2, the key information "Contract Name" will be used as a risk factor. Alternatively, if the "Contract Name" of the contract has not changed, but the "Legal Person" associated with the key information "Contract Name" has changed, for example, from "Zhang" in the pre-upgrade version A1 to "Li" in the post-upgrade version A2, the "Legal Person" associated factor will be used as a risk factor.

[0128] 230. Determine the risk outcome based on the weight information corresponding to the risk factors.

[0129] Among them, due to different needs of users, the emphasis on risk factors is also different. Therefore, the weight information corresponding to the risk factor can be set by the user, and the weight information reflects the proportion of a certain risk factor in all risk factors. For a certain risk content, the weight values ​​of each risk factor corresponding to the risk content can be multiplied, and the product obtained is used as the risk result.

[0130] 120. Classify the document content based on the audit results.

[0131] In the process of risk review of the document in this embodiment, the review result obtained is the above risk result. After the risk result is obtained, the document content can be classified according to the risk result, which can be achieved in the following way:

[0132] Determine the risk level corresponding to the risk result and classify the document content according to the risk level.

[0133] Exemplarily, risks can be pre-classified into three risk levels: high, medium, and low, and each level corresponds to a risk range. Therefore, this embodiment can determine the risk level corresponding to the risk result according to the range in which the value of the risk result is located. After the risk level is determined, the audit results can be classified according to the risk level and an audit report can be generated. For example, in the audit report of a contract, the contract content can be classified according to high risk, medium risk, and low risk, so that customers can be more focused in the process of reading the audit report, thereby improving the user experience.

[0134] In addition, the above classification of document content according to risk level can be used as a first-level category. On this basis, for the document content of each risk level, the document content can also be classified according to key information, that is, on the basis of the first-level category, the document content can be classified at the second level. For example, for the medium-risk content in the contract, it can be further classified according to key information, such as equity relations, controlling relations, blacklist libraries, and credit reporting agencies, so as to realize hierarchical management of file content, make the audit report clearly structured, avoid the time-consuming and laborious problem of users reading the entire contract information, and improve the user's reading efficiency.

[0135] Furthermore, after determining the risk level of the document content, risk warning information may be generated according to the risk level, wherein the risk warning information may include the specific risk level of the document and the cause of the risk, so as to remind the user.

[0136] For example, the risk content and its risk results can also be sent to the lawyer, and the modification method provided by the lawyer can be used as the prompt information of the risk content. This arrangement can make the obtained prompt information more targeted and professional.

[0137] In addition, in the traditional document review process, the system generally conducts automatic review first and then manual review. The technical solution provided by the embodiment of the present invention is that when the system identifies that there is content in the document that matches the company's prohibited conditions, for example, if the payment time in the contract exceeds the time threshold set by the company, or if someone related to the contract has a bad credit record, the document will be returned and a reminder email will be sent to the user to remind the user that the contract does not meet the conditions. This setting can save time and resource consumption in subsequent manual review.

[0138] The technical solution provided by the embodiment of the present invention can automatically implement the review of document content based on RPA and AI technology, and classify the document content according to the review results. This setting provides a focus for users to read the review results of the document, solves the problem of time-consuming and laborious reading of the review results by users, and improves the user experience. The technical solution provided by this embodiment, specifically when conducting a legal risk review of the document content, classifies the document content according to the risk level corresponding to the risk results, provides users with a clear hierarchy of review points when reading the review results of the document, and improves the user experience. In addition, the technical solution provided by this embodiment can automatically generate content prompt information according to the risk level, and the prompt information can be used as a reference and basis for the user's next operation, thereby helping to advance the business process of the document.

[0139] Embodiment 2

[0140] See also Figure 4 , Figure 4 Schematic diagram of the structure of a document review device based on RPA and AI provided by an embodiment of the present invention. Figure 4 As shown, the device includes: an audit module 310 and a classification module 320; wherein,

[0141] The review module 310 is configured to review the document content;

[0142] The classification module 320 is configured to classify the document content according to the audit result.

[0143] Optionally, the review module 310 includes:

[0144] The legal risk review submodule is configured to conduct a legal risk review on the document content, or,

[0145] The grammar and word order review submodule is configured to review the grammar and word order of the document content.

[0146] Optionally, the legal risk review submodule includes:

[0147] A key information extraction unit, configured to extract key information from the document content;

[0148] The legal risk review unit is configured to perform a legal risk review on the document content based on the key information.

[0149] Optionally, the legal risk review unit includes:

[0150] A correlation factor search subunit is configured to search for a correlation factor corresponding to the key information from a preset knowledge graph;

[0151] A risk factor determination subunit, which determines whether the associated factor is a risk factor;

[0152] The risk result determination subunit determines the risk result according to the weight information corresponding to the risk factor if it is determined that the associated factor is a risk factor.

[0153] Optionally, the risk factor determination subunit is specifically configured as follows:

[0154] If different versions of the document exist, the contents of the correlation factors in the different versions are compared;

[0155] If the contents corresponding to the correlation factors are different, the correlation factors are used as risk factors.

[0156] Optionally, the risk factor determination subunit is specifically configured as follows:

[0157] Matching the content corresponding to the correlation factor with the risk content in the label library;

[0158] If the match is successful, the associated factor is used as a risk factor.

[0159] Optionally, the risk result determination subunit is specifically configured as follows:

[0160] If the associated factor is a risk factor, the weight information corresponding to each risk factor is multiplied, and the product is used as the risk result.

[0161] Optionally, the classification module includes:

[0162] a risk level determination unit, configured to determine a risk level corresponding to the risk result;

[0163] The classification unit is configured to classify the document content according to the risk level.

[0164] Optionally, the classification unit is specifically configured as:

[0165] For each risk level of document content, the document content is classified according to the key information and an audit report is generated.

[0166] Optionally, the device further comprises:

[0167] a warning information generating module, configured to generate risk warning information according to the risk level; or,

[0168] The lawyer suggestion module is configured to send the risk content and its risk results to the lawyer, and use the modification method provided by the lawyer as prompt information of the risk content.

[0169] Optionally, the device further comprises:

[0170] The document return module is configured to return the document and send a prompt email to the user if it is identified that the document content contains content that matches the company's prohibition conditions after reviewing the document content.

[0171] The RPA and AI-based document review device provided in the embodiment of the present invention can execute the RPA and AI-based document review method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method. For technical details not described in detail in the above embodiments, please refer to the RPA and AI-based document review method provided in any embodiment of the present invention.

[0172] Embodiment 3

[0173] See also Figure 5 , Figure 5 Schematic diagram of a computing device provided by an embodiment of the present invention. Figure 5 As shown, the computing device may include:

[0174] A memory 701 storing executable program codes;

[0175] a processor 702 coupled to the memory 701;

[0176] Among them, the processor 702 calls the executable program code stored in the memory 701 to execute the document review method based on RPA and AI provided by any embodiment of the present invention.

[0177] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program, wherein the computer program enables a computer to execute the RPA and AI-based document review method provided by any embodiment of the present invention.

[0178] In various embodiments of the present invention, it should be understood that the size of the serial numbers of the above-mentioned processes does not mean the necessary order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0179] In the embodiments provided by the present invention, it should be understood that "B corresponding to A" means that B is associated with A, and B can be determined according to A. However, it should also be understood that determining B according to A does not mean determining B only according to A, and B can also be determined according to A and / or other information.

[0180] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0181] If the above integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-accessible memory. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several requests for a computer device (which can be a personal computer, a server or a network device, etc., specifically a processor in a computer device) to perform some or all of the steps of the above methods of various embodiments of the present invention.

[0182] Those skilled in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0183] The above is a detailed introduction to a document review method, device, equipment and medium based on RPA and AI disclosed in an embodiment of the present invention. Specific examples are used in this article to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea; at the same time, for general technical personnel in this field, according to the idea of ​​the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.

Claims

1. A document review method based on RPA and AI, characterized in that: include: S1. Review the document content, wherein the step S1 specifically includes: S11. Perform a legal risk review on the document content, or S12. Perform a grammatical review on the document content; wherein the step S11 specifically includes: S111. Extract key information from the document content; S112. Perform a legal risk review on the document content based on the key information; wherein the step S112 specifically includes: S1121. Search for association factors corresponding to the key information from a preset knowledge graph; S1122. Determine whether the association factor is a risk factor; S1123. If the association factor is a risk factor, determine the risk result according to the weight information corresponding to the risk factor; wherein the step S1122 specifically includes: if there are different versions of the document, compare the contents of the association factors in different versions; if there are differences in the contents corresponding to the association factors, use the association factors as risk factors; the step S1123 specifically includes: if the association factor is a risk factor, multiply the weight information corresponding to each risk factor, and use the product as the risk result; After step S1, the method further comprises: If it is identified that the document contains content that matches the company's prohibition conditions, the document is returned and a reminder email is sent to the user; S2. Classifying the document contents according to the audit results; Wherein, the step S2 specifically includes: S21, determining the risk level corresponding to the risk result, wherein the risk level includes three levels: high, medium and low, and different risk levels correspond to different risk ranges; S22, classifying the document content according to the risk level to obtain a primary classification result; For each risk level of the document content, the document content is classified according to the key information to obtain a secondary classification result, and an audit report is generated according to the secondary classification result; The method further comprises: Generate risk warning information according to the risk level; or, The risk content and its risk results are sent to the lawyer, and the modification method provided by the lawyer is used as a prompt information of the risk content.

2. The method according to claim 1, characterized in that: The step S1122 specifically includes: Matching the content corresponding to the correlation factor with the risk content in the label library; If the match is successful, the associated factor is used as a risk factor.

3. A document review device based on RPA and AI, characterized in that: include: An audit module, configured to audit the document content; A classification module, configured to classify the document content according to the audit result; Wherein, the audit module includes: The legal risk review submodule is configured to perform a legal risk review on the document content, or the grammar and word order review submodule is configured to perform a grammar and word order review on the document content; The legal risk review submodule includes: A key information extraction unit, configured to extract key information from the document content; A legal risk review unit, configured to perform a legal risk review on the document content based on the key information; Wherein, the legal risk review unit includes: A correlation factor search subunit is configured to search for a correlation factor corresponding to the key information from a preset knowledge graph; A risk factor determination subunit, which determines whether the associated factor is a risk factor; A risk result determination subunit, if it is determined that the associated factor is a risk factor, determines the risk result according to the weight information corresponding to the risk factor; Wherein, the risk factor determination subunit is specifically configured as follows: If different versions of the document exist, the contents of the correlation factors in the different versions are compared; If the content corresponding to the correlation factor is different, the correlation factor is used as a risk factor; The risk result determination subunit is specifically configured as follows: If the correlation factor is a risk factor, the weight information corresponding to each risk factor is multiplied, and the product is used as the risk result; Wherein, the classification module includes: a risk level determination unit, configured to determine a risk level corresponding to the risk result, wherein the risk level includes three levels: high, medium, and low, and different risk levels correspond to different risk ranges; A classification unit is configured to classify the document content according to the risk level to obtain a primary classification result; for the document content of each risk level, classify the document content according to the key information to obtain a secondary classification result, and generate an audit report according to the secondary classification result; The device also includes: a warning information generating module, configured to generate risk warning information according to the risk level; or, The lawyer suggestion module is configured to send the risk content and its risk results to the lawyer, and use the modification method provided by the lawyer as the prompt information of the risk content; Wherein, the device further comprises: The document return module is configured to return the document and send a prompt email to the user if it is identified that the document content contains content that matches the company's prohibition conditions after reviewing the document content.

4. A computing device, characterized in that include: A memory storing executable program code; a processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the document review method based on RPA and AI as described in claim 1 or 2.

5. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the RPA and AI-based document review method as described in claim 1 or 2 is implemented.