Enterprise management industry intelligent contract review method and system based on large model

Through the intelligent contract review method based on large models, the problems of low efficiency and high cost of traditional contract review are solved, and fast and accurate risk identification and review are achieved, the cost of human resources is reduced, and the transparency of review is improved.

CN119048295APending Publication Date: 2024-11-29SHANDONG LANGCHAO YUNTOU INFORMATION TECH CO LTD
View PDF 0 Cites 9 Cited by

Patent Information

Application Number
CN202411118502.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

Traditional contract review is inefficient, subjective, high error risk and high cost. Existing artificial intelligence-based tools have limited processing capabilities, lack of in-depth understanding and high maintenance costs.

Method used

Using a smart contract review method based on big models, upload contract documents through the web interface, convert and preprocess, use big models to deeply analyze and identify risk points and extract key information, and provide standard templates and legal references in combination with the knowledge base to generate Review reports and allow users to comment and modify online.

Benefits of technology

It significantly reduces the time for contract review, improves the quality and reliability of reviews, reduces dependence on professional legal personnel, reduces human resources and operational costs, and improves user participation and review transparency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119048295A_ABST
    Figure CN119048295A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of artificial intelligence, in particular to an enterprise management industry intelligent contract review method and system based on a large model, and the method comprises the following steps: a user uploads a contract document through a Web interface; the control module receives the document and performs format conversion and processing; an algorithm module is applied to carry out document preprocessing, and prompt is generated; the large model module receives the prompt, carries out deep analysis, identifies risk points and extracts key information; the knowledge base module assists the large model and provides a standard template and legal provision reference; the enterprise management industry intelligent contract review method and system based on the large model have the beneficial effects that through automatic document uploading, format conversion, text preprocessing and risk point recognition, the time required by traditional contract review is greatly shortened, and an enterprise can quickly respond to contract review requirements.
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 specifically to a smart contract review method and system for enterprise management industry based on a large model. Background Art

[0002] In the business management industry, contract review is a vital task, which not only involves the economic interests of the enterprise, but also the legal risk management of the enterprise. Traditional contract review mainly relies on the personal experience and manual review of legal professionals. This method has the following major problems and shortcomings:

[0003] Inefficiency: Manual contract review consumes a lot of time and effort, especially for large or complex contract documents.

[0004] High subjectivity: Review methods that rely on personal experience may lead to inconsistency and subjective bias in review results.

[0005] Risk of Error: Manual review is prone to oversights and errors, especially when dealing with a large number of details and clauses.

[0006] High cost: Professional legal services are usually expensive, which is a considerable expense for companies that frequently need to review contracts.

[0007] In order to solve the above problems, some contract review tools based on artificial intelligence technology have emerged in recent years. These tools usually use natural language processing (NLP) technology to analyze contract texts and can identify key terms and potential risks in contracts. However, existing technologies still have the following limitations:

[0008] Limited processing capabilities: Most existing tools can only handle some standardized contract types and have limited processing capabilities for complex or non-standardized contract texts.

[0009] Lack of in-depth understanding: Lack of ability to understand and analyze the deep semantics and legal relationships of the contract text.

[0010] Update and maintenance costs: The algorithms and models of existing tools need to be regularly updated and maintained to adapt to changes in laws and business practices, which adds additional operating costs. Summary of the invention

[0011] The purpose of the present invention is to provide a smart contract review method and system for the enterprise management industry based on a big model to solve the problems raised in the above background technology.

[0012] To achieve the above object, the present invention provides the following technical solution: a smart contract review method for enterprise management industry based on a large model, the method comprising the following steps:

[0013] Users upload contract documents through the web interface;

[0014] The control module receives the document and performs format conversion and processing;

[0015] Apply the algorithm module to preprocess the document and generate prompt;

[0016] The large model module receives prompts and conducts in-depth analysis to identify risk points and extract key information;

[0017] The knowledge base module assists the large model and provides standard templates and legal text references;

[0018] The system generates a contract review report and displays it in the user-side module;

[0019] Users can annotate and modify review results online;

[0020] The user exports the modified contract document.

[0021] Preferably, the specific operations of the user uploading the contract document through the web interface include:

[0022] The user enters the operation page, clicks the contract upload function, and uploads the original contract that needs to be reviewed. Documents in doc, docx, and pdf formats are supported;

[0023] The specific operations of the control module receiving documents and performing format conversion and processing include:

[0024] The control module receives the document and determines whether the file belongs to doc, docx or pdf according to the file suffix. For non-text pdf files, the control module uses the integrated OCR technology to convert them into the editable docx format, and proofreads and modifies the converted documents: deletes unnecessary line breaks, and repairs the table style; then passes the file to the application algorithm module. docx and doc files can be directly passed to the application algorithm module.

[0025] Preferably, the specific operations of applying the algorithm module to preprocess the document and generate the prompt include:

[0026] Using text analysis technology, the document is divided into different sections according to the structure and content of the contract, and all the clauses in the contract are handed over to the risk clause identification model. The model detects and classifies the clauses according to the semantics, keywords and other information of each clause, including two categories: risky clauses and risk-free clauses; conducts in-depth analysis of the clauses in each section, and extracts text fragments containing key information based on the keywords in the clauses, including Party A and Party B information, amount information, payment information, and date information; guides the large model to accurately identify and extract Party A and Party B specific information from the extracted information fragments; guides the large model to describe the risk clauses according to the risk clauses and make reasonable modification suggestions based on the risk clauses.

[0027] Preferably, the large model module receives the prompt and performs in-depth analysis to identify risk points and extract key information. The specific operations include:

[0028] The prompt is passed to the big model module. Based on a full understanding of the prompt, the big model begins to accurately generate Party A and Party B information, amount information, payment information, date information element information and risk description, risk clause modification suggestion and risk identification information based on the trained knowledge; it filters out unreasonable results and analyzes them to ensure that the returned information is relevant to the query request and excludes irrelevant or misleading answers.

[0029] Preferably, the specific operations of the user online annotation and modification of the review results include:

[0030] Generate a contract review report based on the comprehensive analysis results. The report includes the verification results of contract elements, warnings of potential risks, and modification suggestions, and transmits the results to the user display module; the user-side module displays the review report, and the user-side module displays the generated review report to the user, who views the risk points identified by the system, the extracted key information, and the modification suggestions provided; the user annotates and modifies online, and the user annotates the review report online, raises questions or suggestions, and modifies the contract content directly on the Web interface, supporting real-time text editing and format adjustment; export the modified contract document, and after the user completes the online annotation and modification of the review results, the modified contract document is exported as a new file. The system supports the export of multiple file formats to meet the different needs of users.

[0031] A big model-based intelligent contract review system for the enterprise management industry, the system consists of a user module, a control module, an application algorithm module, a big model module and a knowledge base module, and realizes the automated review of contract texts in the enterprise management industry.

[0032] Preferably, the user module provides users with an interactive web interface, allowing customers to upload contract documents, including viewing contract review results after parsing is complete, including element verification, semantic risks and table errors. Users can annotate and modify the results online, and new files can be exported after the modifications are completed.

[0033] Preferably, the control module is responsible for receiving the contract document uploaded by the user. If it is a pdf file, it is converted into a docx file through the integrated OCR, and then formatted and transmitted to the control end module.

[0034] Preferably, the application algorithm module pre-processes the contract document, including text cleaning, entity recognition and key element extraction, writes prompts to provide structured input for the large model, and performs information analysis after cleaning / filtering according to the output of the large model.

[0035] Preferably, the large model module, based on training a large-scale language model, conducts in-depth analysis of the contract text, identifies potential risk points and inconsistencies, and extracts amount information, payment information, Party A and Party B information, and date;

[0036] The knowledge base module stores standard contract-related templates, provides reference information for large models, and enhances the accuracy of reviews.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] The intelligent contract review method and system for the enterprise management industry based on a big model proposed in the present invention greatly reduces the time required for traditional contract review through automated document uploading, format conversion, text preprocessing and risk point identification, enabling enterprises to quickly respond to contract review needs.

[0039] By using a large model based on the Transformer architecture to perform in-depth semantic analysis and extract key information, the present invention can more accurately identify risk points and key elements in the contract, thereby improving the quality and reliability of the review.

[0040] By reducing the reliance on professional legal personnel for review, the present invention helps to reduce the human resource costs of enterprises. At the same time, the automated review process reduces the potential costs caused by human errors.

[0041] Users can annotate and modify the review results directly on the web interface, which improves user participation and transparency of the review process. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flowchart of the smart contract review of the present invention;

[0043] Figure 2This is a system block diagram of the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose and technical solution of the present invention clearly and completely described, and the advantages more clearly understood, the embodiments of the present invention are further described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, rather than all of the embodiments, and are only used to explain the embodiments of the present invention, and are not used to limit 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.

[0045] Embodiment 1

[0046] See also Figure 1 The present invention provides a technical solution: a smart contract review method for enterprise management industry based on a large model, the method comprising the following steps:

[0047] The user enters the operation page, clicks the contract upload function, and uploads the original contract that needs to be reviewed, supporting documents of doc, docx and pdf types.

[0048] The control module receives the document and determines whether the file belongs to doc, docx or pdf according to the file suffix. For non-text pdf files, the control module uses the integrated OCR technology to convert them into editable docx format. The converted document is proofread and modified: redundant line breaks are deleted, table styles are repaired, etc. The file is then passed to the application algorithm module. docx and doc files can be directly passed to the application algorithm module.

[0049] After the application algorithm module receives the article, it performs the following operations:

[0050] 1. Section classification: Using text analysis technology, the document is divided into different sections according to the structure and content of the contract, such as: title and preface, project content, service period, contract amount, project delivery, project acceptance, payment method, Party A's responsibilities and obligations, Party B's responsibilities and obligations, force majeure, breach of contract liability, disputes and arbitration, other agreed matters, attachments and stamped signatures, etc.

[0051] 2. Risk clause identification: All clauses in the contract are submitted to the risk clause identification model, which detects and classifies the clauses according to their semantics, keywords and other information, including risky clauses and risk-free clauses.

[0052] 3. Information fragment extraction: Conduct in-depth analysis of the clauses in each section, and extract text fragments containing key information based on the keywords in the clauses, including Party A and Party B information, amount information, payment information, date information, etc. For example, based on the information such as 'total contract amount', 'amount excluding tax', 'value-added tax rate' in the clauses, it is judged that the fragment contains amount information, and based on the information such as 'service period', 'contract period', 'service time range' in the clauses, it is judged that the fragment contains date information. Based on the information such as 'account name', 'taxpayer identification number', 'name of the bank' in the clauses, it is judged that the fragment contains Party A and Party B information.

[0053] 4. Write prompts for extracting information: Guide the big model to accurately identify and extract specific information of Party A and Party B from the extracted information fragments, such as company name, bank, contact number, address, etc. Extract detailed information related to the amount, including total amount, lowercase tax-inclusive amount, uppercase tax-inclusive amount, applicable tax rate, and calculated tax amount. Let the big model identify payment-related terms, such as the installment payment arrangement, the payment ratio of each installment, etc. Extract all important dates mentioned in the contract, such as the contract signing date, effective date, payment deadline, etc.

[0054] 5. Prompt for writing risk clause modification suggestions: guide the large model to describe the risks of risk clauses according to the risk clauses and make reasonable modification suggestions based on the risk clauses.

[0055] The prompt is passed to the big model module. Based on a full understanding of the prompt, the big model begins to accurately generate factor information such as Party A and Party B information, amount information, payment information, date information, and risk identification information such as risk description and risk clause modification suggestions based on the trained knowledge.

[0056] Filter unreasonable results and analyze: Ensure that the information returned is relevant to the query request and exclude irrelevant or misleading answers. Ensure that the information returned is consistent in logic and facts, such as the consistency of dates, events and data. Check whether the amount appears in the same case format in all documents, whether uppercase or lowercase, to maintain professionalism and consistency. Verify calculations involving amounts to ensure that the results of adding or subtracting all values ​​are correct and there are no calculation errors. Ensure that the information of Party A and Party B in the contract or agreement is consistent throughout the document without confusion or incorrect mention. Identify and mark statements that may indicate risk or uncertainty, such as "maybe", "perhaps", "uncertain", etc., for further evaluation. Compare whether the data in different parts matches, such as product quantity, price, total amount, etc., to ensure that the data in the entire document is accurate. Check whether the format and grammar of the document are correct, including punctuation, spelling and grammatical structure. Ensure that the content of the document complies with relevant laws, regulations and industry standards, and there are no violations or improper expressions. Identify and delete or replace sensitive information, such as personal privacy data, business secrets, etc. Make sure the information in the document is relevant to the context and is not taken out of context or misinterpreted. Check whether the document has a coherent logical flow and whether there are logical jumps or incoherence between the parts. Make sure the correct professional terminology is used and that the terminology is used appropriately and consistently.

[0057] Generate a contract review report based on the comprehensive analysis results. The report includes the verification results of contract elements, warnings of potential risks, and modification suggestions. The results are transmitted to the user display module.

[0058] The user-side module displays the review report. The user-side module displays the generated review report to the user, and the user can view the risk points identified by the system, the key information extracted, and the modification suggestions provided.

[0059] Users can annotate and modify online. Users can annotate the review report online, raise questions or suggestions, and modify the contract content directly on the web interface. The system supports real-time text editing and format adjustment.

[0060] Export the modified contract document. After the user completes the online annotation and modification of the review results, the modified contract document can be exported as a new file. The system supports the export of multiple file formats to meet the different needs of users.

[0061] Embodiment 2

[0062] Based on Example 1, a smart contract review system for the enterprise management industry based on a big model is proposed. The system consists of a user module, a control module, an application algorithm module, a big model module and a knowledge base module to realize the automated review of contract texts in the enterprise management industry.

[0063] User Modules:

[0064] Interactive Web interface: An intuitive and easy-to-use user interface is designed to ensure that users can easily upload and manage contract documents; Multi-format support: The system supports multiple document formats such as docx, doc, pdf, etc. to meet the needs and preferences of different users; Review result display: After the document parsing is completed, the system automatically displays the review results, including but not limited to element verification, semantic risk analysis and table error detection; Online annotation and modification: Users can directly annotate the review results on the Web interface, raise questions or suggestions, and modify the text content in real time; Document export function: Allows users to export modified contract documents as new files, supporting multiple file formats for subsequent use or printing.

[0065] Control Module:

[0066] Document reception and processing: The control end module serves as the hub for interaction between the system and users and is responsible for receiving all contract documents uploaded by users; OCR technology integration: For uploaded PDF documents, OCR technology is used to convert them into editable docx format to ensure the readability and processability of the documents; Format standardization: Format standardization is performed on the converted documents, including tables, paragraphs, etc., to facilitate consistency processing in subsequent modules; Information transmission: The processed documents are passed to the application algorithm module for deeper analysis and processing.

[0067] Application algorithm module:

[0068] Text preprocessing: Perform necessary preprocessing steps on the contract document, including removing irrelevant content, text cleaning, and optimizing the document structure; Entity and element recognition: Automatically identify entities (such as names of people and organizations) and key elements (such as contract terms and conditions) in the document through natural language processing technology; Structured prompt generation: Generate structured prompts based on the extracted entities and elements to provide clear and organized input information for the large model; Information analysis and filtering: After receiving the output results of the large model, further cleaning and filtering are performed to remove invalid or redundant information, and in-depth information analysis is performed.

[0069] Large model module:

[0070] Risk identification: Train risk identification models based on customer risk data, and identify risk clauses based on contract clauses provided by customers, including scenarios such as payment collection, intellectual property rights, project changes and contract changes, legal basis, liability for breach of contract, quality and safety responsibility, etc. Risk analysis: Conduct a comprehensive risk analysis of risk clauses and provide modification suggestions based on risk clauses; Extract key information: Accurately extract key financial information (such as amount), time information (such as payment date), information about Party A and Party B, etc. in the contract to provide decision support for the review.

[0071] Knowledge Base Modules:

[0072] Standard template storage: Build and maintain a knowledge base containing standard templates of various contract types for users to refer to when writing or reviewing contracts; Legal text index: Integrate relevant legal texts and judicial interpretations to provide legal basis and support for contract review; Reference information provision: The knowledge base module provides rich reference information for the large model to help improve the accuracy and professionalism of contract review.

[0073] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A smart contract review method for enterprise management industry based on big model, Features: The method comprises the following steps: Users upload contract documents through the web interface; The control module receives the document and performs format conversion and processing; Apply the algorithm module to preprocess the document and generate prompt; The large model module receives prompts and conducts in-depth analysis to identify risk points and extract key information; The knowledge base module assists the large model and provides standard templates and legal text references; The system generates a contract review report and displays it in the user-side module; Users can annotate and modify review results online; The user exports the modified contract document.

2. According to claim 1, a method for reviewing smart contracts in the enterprise management industry based on a large model, Features: The specific operations for users to upload contract documents through the web interface include: The user enters the operation page, clicks the contract upload function, and uploads the original contract that needs to be reviewed. Documents in doc, docx, and pdf formats are supported; The specific operations of the control module receiving documents and performing format conversion and processing include: The control module receives the document and determines whether the file belongs to doc, docx or pdf according to the file suffix. For non-text pdf files, the control module uses the integrated OCR technology to convert them into the editable docx format, and proofreads and modifies the converted documents: deletes unnecessary line breaks, and repairs the table style; then passes the file to the application algorithm module. docx and doc files can be directly passed to the application algorithm module.

3. According to claim 2, a method for reviewing smart contracts in the enterprise management industry based on a large model, Features: The specific operations of applying the algorithm module to preprocess the document and generate prompts include: Using text analysis technology, the document is divided into different sections according to the structure and content of the contract, and all the clauses in the contract are handed over to the risk clause identification model. The model detects and classifies the clauses according to the semantics, keywords and other information of each clause, including two categories: risky clauses and risk-free clauses; conducts in-depth analysis of the clauses in each section, and extracts text fragments containing key information based on the keywords in the clauses, including Party A and Party B information, amount information, payment information, and date information; guides the large model to accurately identify and extract Party A and Party B specific information from the extracted information fragments; guides the large model to describe the risk clauses according to the risk clauses and make reasonable modification suggestions based on the risk clauses.

4. According to claim 1, a method for reviewing smart contracts in the enterprise management industry based on a large model, Features: The large model module receives prompts and performs in-depth analysis to identify risk points and extract key information. The specific operations include: The prompt is passed to the big model module. Based on a full understanding of the prompt, the big model begins to accurately generate Party A and Party B information, amount information, payment information, date information element information and risk description, risk clause modification suggestion and risk identification information based on the trained knowledge; it filters out unreasonable results and analyzes them to ensure that the returned information is relevant to the query request and excludes irrelevant or misleading answers.

5. According to claim 1, a method for reviewing smart contracts in the enterprise management industry based on a large model, Features: The specific operations for users to annotate and modify review results online include: Generate a contract review report based on the comprehensive analysis results. The report includes the verification results of contract elements, warnings of potential risks, and modification suggestions, and transmits the results to the user display module; the user-side module displays the review report, and the user-side module displays the generated review report to the user, who views the risk points identified by the system, the extracted key information, and the modification suggestions provided; the user annotates and modifies online, and the user annotates the review report online, raises questions or suggestions, and modifies the contract content directly on the Web interface, supporting real-time text editing and format adjustment; export the modified contract document, and after the user completes the online annotation and modification of the review results, the modified contract document is exported as a new file. The system supports the export of multiple file formats to meet the different needs of users.

6. A big model-based enterprise management industry smart contract review system according to any one of claims 1 to 5, Features: The system consists of a user module, a control module, an application algorithm module, a large model module and a knowledge base module, and realizes the automated review of contract texts in the enterprise management industry.

7. According to claim 6, a large model-based enterprise management industry intelligent contract review system, Features: The user module provides users with an interactive web interface, allowing customers to upload contract documents, including viewing contract review results after parsing, including element verification, semantic risks and table errors. Users can annotate and modify the results online, and export new files after the modifications are completed.

8. According to claim 6, a large model-based enterprise management industry intelligent contract review system, Features: The control module is responsible for receiving the contract document uploaded by the user. If it is a pdf file, it is converted into a docx file through the integrated OCR, and then formatted and transmitted to the control end module.

9. According to claim 6, a large model-based enterprise management industry intelligent contract review system, Features: The application algorithm module pre-processes the contract document, including text cleaning, entity recognition and key element extraction, writes prompts to provide structured input for the big model, and performs information analysis after cleaning / filtering according to the output of the big model.

10. According to claim 6, a large model-based enterprise management industry intelligent contract review system, Features: The large model module, based on training a large-scale language model, conducts in-depth analysis of the contract text, identifies potential risk points and inconsistencies, and extracts amount information, payment information, Party A and Party B information, and date; The knowledge base module stores standard contract-related templates, provides reference information for large models, and enhances the accuracy of reviews.

Citation Information

Cited By

  • Multi-format purchase order analysis and input method based on artificial intelligence

    CN119830894A

  • Contract risk assessment and compliance check method, system and equipment based on large language model, and medium

    CN119863119A

  • Method and system for automatically classifying construction site labor contracts

    CN120198061A

  • Order remark information processing method, medium, equipment and program product

    CN120374237A

  • Automatic contract auditing system and method based on large language model

    CN120952708A