Legal affair consultation method and system based on AI interaction

Abstract

The invention relates to the technical field of legal consultation, in particular to a legal affair consultation method and system based on AI interaction. The method comprises the following steps: acquiring consultation information; analyzing the consultation information, and determining a consultant type and case description; analyzing the case description, and determining a legal subject and a dispute focus; and determining a case handling lawyer according to the legal subject, the dispute focus and the consultant type. The consultation starting efficiency is improved, the initial interaction time is shortened, meanwhile, the data availability is guaranteed, and flow interruption caused by information omission is prevented. The interference of redundant data is eliminated, the accuracy of user identification and the clarity of case content are improved, and consultation deviation caused by identity misjudgment or content mixing is reduced. Error judgment of entities is avoided, analysis errors are reduced, the case processing period is shortened, and the correlation of legal suggestions is enhanced. The lawyer matching accuracy and efficiency are improved, resource waste is reduced, and the case processing quality is enhanced.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of legal consultation, in particular to a legal affair consultation method and system based on AI interaction. BACKGROUND

[0002] In the field of modern legal affair consultation, AI interaction legal affair consultation systems have been widely used to improve consultation efficiency and accessibility. AI interaction legal affair consultation systems usually analyze user input consultation information based on natural language processing technology and automatically provide legal advice or match professional lawyers.

[0003] However, AI interaction legal affair consultation systems only focus on the analysis of case content, and for consultation seekers with clear purposes, dynamic adjustment cannot be achieved, which may result in a large waste of time for consultation seekers and is not conducive to making up for the non-professional nature of consultation seekers and legal consultation business conversion. SUMMARY

[0004] The application provides a legal affair consultation method and system based on AI interaction to solve the above problems.

[0005] In a first aspect, the application provides a legal affair consultation method based on AI interaction, which comprises: obtaining consultation information; analyzing the consultation information to determine the type of the consultation seeker and the description of the case; analyzing the case description to determine the legal subject and the dispute focus; determining the case handling lawyer according to the legal subject, the dispute focus and the type of the consultation seeker.

[0006] According to the scheme, the consultation information is obtained to improve the consultation starting efficiency, shorten the initial interaction time, ensure the data availability and prevent process interruption caused by information omission. The consultation information is analyzed to determine the type of the consultation seeker and the description of the case, which eliminates redundant data interference, improves the accuracy of user identification and the clarity of case content, and reduces consultation deviation caused by identity misjudgment or mixed content. The case description is analyzed to determine the legal subject and the dispute focus, which avoids entity misjudgment, reduces analysis errors, shortens the case processing period and enhances the relevance of legal advice. The case handling lawyer is determined according to the legal subject, the dispute focus and the type of the consultation seeker, which improves the accuracy and efficiency of lawyer matching, reduces resource waste and enhances the case processing quality.

[0007] Optionally, the analysis of the consultation information to determine the type of the consultation seeker comprises: analyzing the consultation information to determine the user registration source and the consultation history; determining the type of the consultation seeker according to the user registration source and the consultation history.

[0008] The application analyzes the consultation information, determines the user registration source and consultation history, realizes complete data acquisition of the user background, and provides structured input basis for the classification of the consultant type. According to the user registration source and the consultation history, the consultant type is determined, the processing logic is dynamically adjusted according to the type, and the overall consultation efficiency is improved.

[0009] Optionally, the determining of the case handling lawyer according to the legal subject, the dispute focus and the consultant type comprises: According to the legal subject, the AI question and answer database is called to determine the AI reply graph; According to the dispute focus, the AI reply graph is analyzed to determine the AI reply level; According to the AI reply level, it is determined whether to allocate a lawyer; If it is determined to allocate a lawyer, the case handling lawyer is determined according to the consultant type.

[0010] According to the legal subject, the AI question and answer database is called to determine the AI reply graph, the historical data based on similar legal subjects is analyzed to improve the pertinence and efficiency of data retrieval, reduce irrelevant information interference, and provide accurate data basis for dispute focus analysis. According to the dispute focus, the AI reply graph is analyzed to determine the AI reply level, the processing capacity of AI is accurately evaluated, the reliability of AI in the dispute focus scene is ensured, the objective basis for the decision of whether to allocate a lawyer is provided, and the consultation error or resource misallocation caused by insufficient AI capacity is avoided. According to the AI reply level, it is determined whether to allocate a lawyer, the dynamic resource optimization is realized, the case handling accuracy and user satisfaction are improved, the lawyer resources are saved, and the overall efficiency is improved. If it is determined to allocate a lawyer, the case handling lawyer is determined according to the consultant type, the personalized lawyer matching is realized, the user trust and the consultation conversion rate are enhanced, and the lawyer allocation process is efficient and targeted.

[0011] Optionally, the consultant type comprises a referral type, and the determining of the case handling lawyer according to the consultant type comprises: If the consultant type is the referral type, the referral information is obtained; the lawyer information is determined according to the referral information; The lawyer information is analyzed to determine the field of expertise; The lawyer matching degree is determined according to the field of expertise and the legal subject; The case handling lawyer is determined according to the lawyer matching degree.

[0012] According to the scheme, if the consultant type is a referral type, the referral information is obtained, the recommendation relationship information is used in the matching process, and the personalized processing is enhanced. According to the referral information, the lawyer information is determined, the lawyer detailed data related to the recommendation is ensured to be obtained, and the basis for professional field analysis is provided. The lawyer information is analyzed, the field of expertise is determined, and it is ensured that the matching is based on the actual expertise of the lawyer. According to the field of expertise and the legal subject, the lawyer matching degree is determined, the matching degree of the lawyer and the case is quantified, the matching decision is objectivized, and subjective deviation is avoided. According to the lawyer matching degree, the case handling lawyer is determined, the efficient and accurate distribution of the lawyer is ensured, and the case handling efficiency is improved.

[0013] Optionally, the analyzing the case description, determining the legal subject and the dispute focus, comprises: performing natural language processing on the case description, and extracting entity information, claim expression and legal relationship; determining the legal subject corresponding to the entity information according to a preset legal element classification model; analyzing the claim expression and the legal relationship, and determining the dispute focus.

[0014] According to the scheme, the case description is subjected to natural language processing, the entity information, the claim expression and the legal relationship are extracted, and it is ensured that the case elements are fully captured. According to the preset legal element classification model, the legal subject corresponding to the entity information is determined, the entity misjudgment is avoided, the correct identification of the legal subject is ensured, and reliable basis is provided for case analysis and resource matching. The claim expression and the legal relationship are analyzed, the dispute focus is determined, the case analysis is accurate, and information omission or incompleteness is reduced.

[0015] Optionally, the determining the consultant type according to the user registration source and the consultation history, comprises: analyzing the consultation history, and determining consultation frequency and consultation field; analyzing the consultation history based on the consultation field, and determining consultation depth; predicting the user's legal knowledge mastery degree according to the consultation depth; determining the consultant type according to the user registration source, the legal knowledge mastery degree and the consultation frequency.

[0016] By the scheme, the consultation history is analyzed, the consultation frequency and the consultation field are determined, the consultation activity and the historical participation degree of the user are reflected, and the interest focus and the common legal problem direction of the user are revealed, avoiding generalization processing. Based on the consultation field, the consultation history is analyzed, the consultation depth is determined, irrelevant data interference is avoided, the pertinence and accuracy of the analysis are improved, and the misjudgment risk is reduced. According to the consultation depth, the legal knowledge level of the user is predicted, the legal literacy level of the user is indicated, subjective judgment is avoided, and the integrity and operability of the user portrait are enhanced. According to the user registration source, the legal knowledge level, and the consultation frequency, the consultant type is determined, the fine identification of the user type is realized, the randomness of type division is avoided, and personalized consultation processing is supported.

[0017] Optionally, the AI reply level is determined according to the dispute focus, the AI reply graph is analyzed, and the AI reply level is determined. Based on the legal subject, the dispute focus is analyzed, and the legal relationship complexity is determined; The AI question and answer database and the AI reply graph are analyzed, and the consultation refutation situation is determined; According to the consultation refutation situation, the reply accuracy rate is determined; According to the legal relationship complexity and the reply accuracy rate, the AI reply level is determined.

[0018] Through the scheme, based on the legal subject, the dispute focus is analyzed, the legal relationship complexity is determined, the overall difficulty of the case is reflected, and the evaluation weight of the AI reply level is affected. The AI question and answer database and the AI reply graph are analyzed, the consultation refutation situation is determined, the real-time interaction evidence of the user to the AI reply is captured, a basic data set is provided for the reply accuracy rate calculation, and it is ensured that the refutation situation reflects the real user behavior. According to the consultation refutation situation, the reply accuracy rate is determined, an objective index is provided for measuring the performance quality of the AI, and is used as an input parameter for the AI reply level calculation. According to the legal relationship complexity and the reply accuracy rate, the AI reply level is determined, it is ensured that the resource allocation is based on the case complexity and the AI reliability, and the consultation efficiency is optimized.

[0019] Optionally, the lawyer matching degree is determined according to the lawyer matching degree, and the case lawyer is determined. When the lawyer matching degree is lower than a preset matching threshold, the cooperation network of the introduced lawyer is obtained based on the lawyer information; The cooperation network is analyzed, and the practice field intersection degree of the cooperation lawyer is determined; According to the practice field intersection degree, the alternative lawyer is determined; The lawyer information of the alternative lawyer is obtained and analyzed, and the case victory rate and the current case load of the alternative lawyer are determined; According to the case victory rate and the current case load, a comprehensive matching coefficient is calculated; determine the case handling lawyer according to the comprehensive matching coefficient.

[0020] According to the scheme, when the lawyer matching degree is lower than the preset matching threshold, the collaboration network of the referral lawyer is obtained based on the lawyer information, avoiding resource waste caused by initial matching failure, and ensuring that the matching process does not interrupt. Analyzing the collaboration network, determining the cross degree of the practice field of the collaborative lawyer, reducing the interference of irrelevant lawyers, providing objective basis for screening candidate lawyers, and improving the professional relevance of matching. According to the cross degree of the practice field, determine the candidate lawyer, ensure that the professional background of the candidate lawyer is highly consistent with the demand of the case, avoid inefficient allocation caused by field mismatch, and optimize the quality of the candidate pool. Obtain and analyze the lawyer information of the candidate lawyer, determine the case success rate and current case load of the candidate lawyer, ensure that the matching process considers actual constraints, and avoid overloading or inefficient lawyers. According to the case success rate and the current case load, calculate the comprehensive matching coefficient, simplify the decision logic, and facilitate the evaluation and sorting of the applicability of the candidate lawyer. According to the comprehensive matching coefficient, determine the case handling lawyer, improve the case handling efficiency and success rate, avoid the need for manual intervention, and realize closed-loop resource optimization.

[0021] Optionally, the determining the lawyer matching degree according to the field of expertise and the legal subject includes: Obtain the historical cross-field cases of the referral lawyer; the historical cross-field cases include case handling time and case handling result; Based on the field of expertise, obtain the historical handling cases; the historical handling cases include the average handling time; According to the average handling time and the case handling time, determine the handling difference of the cross-field case; Based on the dispute focus, analyze the historical cross-field cases to determine the handling similarity; Statistically analyze the case handling result to determine the success rate of the cross-field case; According to the handling similarity and the success rate, calculate the field adaptation coefficient; According to the field adaptation coefficient, the field of expertise and the legal subject, determine the lawyer matching degree.

[0022] By the scheme, the historical cross-field cases of the introduced lawyer are acquired, the actual performance data of the introduced lawyer in the cross-field cases is ensured to be accessed, and a data foundation is laid for calculating and processing the difference and the winning rate and the like indexes. Based on the proficient field, the historical processing cases are acquired, a work efficiency reference point of the introduced lawyer in the familiar field is provided, and the processing efficiency deviation of the cross-field case is used for comparison. According to the average processing time length and the case processing time length, the processing difference of the cross-field case is determined, the processing efficiency difference of the cross-field case compared with the case in the proficient field is revealed, and an efficiency dimension basis is provided for evaluating the cross-field adaptability of the lawyer. Based on the dispute focus, the historical cross-field cases are analyzed, the processing similarity is determined, the similarity degree of the historical cases and the current case on the dispute focus is quantified, and a similarity index is provided for calculating the field adaptation coefficient. The case processing result is counted, the winning rate of the cross-field case is determined, and a performance basis is provided for evaluating the cross-field reliability. According to the processing similarity and the winning rate, the field adaptation coefficient is calculated, and the overall adaptability of the introduced lawyer to process the similar cross-field case as the current case is quantified. According to the field adaptation coefficient, the proficient field and the legal subject, the lawyer matching degree is determined, and it is ensured that the matching result is based on the comprehensive evaluation of the professional field coverage and the cross-field adaptability.

[0023] In a second aspect, the application provides an AI interaction-based legal matter consultation system, which comprises: An information analysis module, configured to acquire consultation information, analyze the consultation information, and determine a consultant type and a case description; A case analysis module, configured to analyze the case description and determine a legal subject and a dispute focus; A case handling determination module, configured to determine a case handling lawyer according to the legal subject, the dispute focus and the consultant type.

[0024] Optionally, when the information analysis module analyzes the consultation information and determines the consultant type, it is configured to: analyze the consultation information to determine a user registration source and a consultation history; determine the consultant type according to the user registration source and the consultation history.

[0025] Optionally, when the case handling determination module determines the case handling lawyer according to the legal subject, the dispute focus and the consultant type, it is configured to: based on the legal subject, call an AI question and answer database to determine an AI reply graph; analyze the AI reply graph according to the dispute focus to determine an AI reply level; determine whether to assign a lawyer according to the AI reply level; if it is determined to assign a lawyer, determine a case handling lawyer according to the consultant type.

[0026] Optionally, the consultant type includes a referral type, and the case handling determination module determines the case handling lawyer based on the consultant type, for: If the consultant type is the referral type, obtaining referral information; and determining lawyer information based on the referral information; Analyze the lawyer's information and determine their areas of expertise; Determine the lawyer's matching degree based on the areas of expertise and the legal subjects; The lawyer handling the case is determined based on the lawyer matching degree.

[0027] Optionally, when the case analysis module analyzes the case description and determines the legal subject and the focus of the dispute, it is used to: Perform natural language processing on the case description to extract entity information, claim statements, and legal relationships; Determine the legal subject corresponding to the entity information based on a preset legal element classification model; Analyze the statement of the claim and the legal relationship to determine the focus of the dispute.

[0028] Optionally, when the information parsing module determines the consultant type based on the user registration source and the consultation history, it is configured to: Analyze the consultation history to determine the frequency and areas of consultation; Analyze the consultation history based on the consultation area to determine the depth of consultation; Based on the depth of the consultation, predict the user's level of legal knowledge; The consultant type is determined based on the user's registration source, the level of legal knowledge mastered, and the consultation frequency.

[0029] Optionally, the case handling determination module analyzes the AI ​​response map based on the dispute focus to determine the AI ​​response level, and is used to: Based on the legal entities, analyze the focus of the dispute and determine the complexity of the legal relationship; Analyze the AI ​​question-answer database and the AI ​​response graph to determine the consultation and rebuttal situation; Determine the accuracy of the response based on the consultation and rebuttal; The AI ​​response level is determined based on the complexity of the legal relationship and the accuracy of the response.

[0030] Optionally, when the case handling determination module determines the lawyer to handle the case based on the lawyer matching degree, it is used to: When the lawyer matching degree is lower than a preset matching threshold, obtaining a collaborative network of referring lawyers based on the lawyer information; Analyze the collaborative network to determine the degree of overlap in the collaborative lawyers' practice areas; determine the candidate lawyers according to the cross degree of the practice field; obtain and analyze the lawyer information of the candidate lawyers, determine the case success rate and current case load of the candidate lawyers; calculate a comprehensive matching coefficient according to the case success rate and the current case load; determine the case handling lawyer according to the comprehensive matching coefficient.

[0031] Optionally, when the case handling determination module determines the lawyer matching degree according to the proficient field and the legal subject, it is used for: obtain historical cross-field cases of the referral lawyer; the historical cross-field cases include case processing time and case processing result; obtain historical processing cases based on the proficient field; the historical processing cases include average processing time; determine the processing difference of the cross-field cases according to the average processing time and the case processing time; analyze the historical cross-field cases based on the dispute focus, and determine the processing similarity; statistically determine the success rate of the cross-field cases according to the case processing result; calculate a field adaptation coefficient according to the processing similarity and the success rate; determine the lawyer matching degree according to the field adaptation coefficient, the proficient field and the legal subject. BRIEF DESCRIPTION OF DRAWINGS

[0032] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0033] Figure 1 An application scenario schematic diagram is provided for an embodiment of the present application; Figure 2 A flowchart of a legal affairs consulting method based on AI interaction is provided for an embodiment of the present application; Figure 3 A structural schematic diagram of a legal affairs consulting system based on AI interaction is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0034] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative work fall within the protection scope of the present application.

[0035] In addition, the term "and / or" herein is only to describe the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can represent the three cases of A alone, A and B together, and B alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects unless otherwise specified.

[0036] The embodiments of the present application will be described in further detail below with reference to the drawings of the specification.

[0037] The AI interaction legal affairs consulting system only focuses on the analysis of case content, and for the purpose of clear consultants, it is impossible to realize dynamic adjustment, which may lead to a large amount of waste of the time of the consultants, and is not conducive to making up for the non-professional nature of the consultants and the conversion of legal consulting business.

[0038] Based on this, the present application provides an AI interaction-based legal affairs consulting method and system, which acquires consulting information, improves the consulting starting efficiency, shortens the initial interaction time, guarantees data availability, and prevents process interruption caused by information omission. Analyzing the consulting information, determining the consultant type and case description, eliminating redundant data interference, improving the accuracy of user identification and the clarity of case content, reducing the consulting deviation caused by identity misjudgment or content mixing. Analyzing the case description, determining the legal subject and the dispute focus, avoiding entity misjudgment, reducing analysis errors, shortening the case processing period, and enhancing the relevance of legal advice. According to the legal subject, the dispute focus and the consultant type, determine the case handling lawyer, improve the accuracy and efficiency of lawyer matching, reduce resource waste, and enhance the case processing quality.

[0039] Figure 1 An application scenario schematic diagram provided by the present application is provided when legal affairs consulting is performed.

[0040] Specifically, the method provided in the application is applied to any server, the server interacts with a user equipment, and consultation information is obtained through the user equipment. The consultation information is analyzed to determine the type of the consultant and the case description. The case description is analyzed to determine the legal subject and the dispute focus, to avoid entity misjudgment, reduce analysis errors, shorten the case processing period, and enhance the relevance of legal advice. According to the legal subject, the dispute focus, and the type of the consultant, a case handling lawyer is determined, the accuracy and efficiency of lawyer matching are improved, resource waste is reduced, and the case processing quality is enhanced. The specific implementation mode can refer to the following embodiments.

[0041] Figure 2 A flowchart of a legal transaction consultation method based on AI interaction provided by an embodiment of the application. The method of the embodiment can be applied to the server in the above scenarios. As shown in the figure, the method comprises the following steps. Figure 2 S201, obtaining consultation information; analyzing the consultation information to determine the type of the consultant and the case description; The consultation information can be legal consultation related data.

[0042] The type of the consultant can be a user classification.

[0043] The case description can be the case related content extracted from the consultation information.

[0044] Specifically, the consultation information is obtained through the user equipment. Then, a natural language processing model established based on natural language processing theory is used for text analysis and entity extraction of the consultation information. Then, a user database established based on relational database management theory is accessed to retrieve the user's consultation history. Then, based on the user registration input theory, the user registration source collected through the form is classified using a rule engine to determine the type of the consultant. At the same time, the natural language processing model extracts the part of the consultation information related to the case as the case description.

[0045] S202, analyzing the case description to determine the legal subject and the dispute focus; The legal subject can be an entity object involved in the case.

[0046] The dispute focus can be the core conflict point of the case.

[0047] Specifically, the case description is subjected to deep semantic analysis. Then, an entity recognition model established based on named entity recognition theory is used to scan the case description text to extract the legal subject. At the same time, a legal element classification model established based on supervised text classification theory is used to analyze the case description to identify the dispute focus.

[0048] S203, determining a case handling lawyer according to the legal subject, the dispute focus, and the type of the consultant.

[0049] ​The case handling lawyer can be a lawyer who handles the case.

[0050] Specifically, SQL queries are used to filter the lawyer's expertise that aligns with the legal subject and the dispute focus; then, the priority is adjusted according to the consultant type to determine the case handling lawyer.

[0051] Through this solution, the consultation information is obtained, the consultation starting efficiency is improved, the initial interaction time is shortened, the data availability is ensured, and the process interruption caused by information omission is prevented. The consultation information is analyzed, the consultant type and the case description are determined, the redundant data interference is eliminated, the accuracy of user identification and the clarity of case content are improved, and the consultation deviation caused by identity misjudgment or content confusion is reduced. The case description is analyzed, the legal subject and the dispute focus are determined, the entity misjudgment is avoided, the analysis error is reduced, the case processing period is shortened, and the relevance of legal advice is enhanced. According to the legal subject, the dispute focus and the consultant type, the case handling lawyer is determined, the accuracy and efficiency of lawyer matching are improved, the resource waste is reduced, and the case processing quality is enhanced.

[0052] In some embodiments, the consultation information is analyzed, and the user registration source and the consultation history are determined; according to the user registration source and the consultation history, the consultant type is determined.

[0053] The user registration source can be the source information input by the user through the form when registering.

[0054] The consultation history can be the consultation record of the user.

[0055] Specifically, a string matching algorithm is used to scan the user ID in the consultation information; then, a database query statement is constructed according to the user ID, and a request is sent to the pre-defined user database; then, in the user database, the user registration source corresponding to the user ID is queried; then, the consultation history associated with the user ID in the user database is queried. Then, the user registration source and the consultation history are input into the rule engine, and finally, the input is processed using conditional statements to determine the consultant type.

[0056] Through this solution, the consultation information is analyzed, the user registration source and the consultation history are determined, the complete data acquisition of the user background is realized, and the structured input basis for the consultant type classification is provided. According to the user registration source and the consultation history, the consultant type is determined, the processing logic is dynamically adjusted according to the type, and the overall consultation efficiency is improved.

[0057] In some embodiments, based on the legal subject, an AI question and answer database is called to determine an AI reply graph; according to the dispute focus, the AI reply graph is analyzed to determine the AI reply level; according to the AI reply level, it is determined whether to assign a lawyer; if it is determined to assign a lawyer, the case handling lawyer is determined according to the consultant type.

[0058] The AI question and answer database can be a database storing historical AI consultation session data.

[0059] The AI reply graph can be structured data representing an association network between questions and answers.

[0060] The AI reply level can be a value quantitatively representing the reliability of the AI in handling different controversy focuses.

[0061] Specifically, based on the legal subject, a number of historical question and answer records related to the legal subject in the AI question and answer database are retrieved; then, the historical question and answer records are organized as structured data to form an AI reply graph. Subsequently, a text analysis algorithm is applied to scan the content in the AI reply graph, focusing on the question and answer part related to the controversy focus; further, a matching degree index is calculated, and based on a preset threshold established by the simple question and answer matching accuracy, the AI reply level is determined. Subsequently, the AI reply level is compared with the preset threshold to determine whether to allocate a lawyer; if the AI reply level is lower than the preset threshold, it is determined that the AI reply level is insufficient, and a lawyer is allocated; further, when a lawyer is determined to be allocated, a lawyer information database is accessed; finally, based on the consultant type, a classification rule is applied to screen the qualified case handling lawyer from the lawyer pool.

[0062] Through the scheme, based on the legal subject, the AI question and answer database is retrieved, the AI reply graph is determined, the analysis based on the historical data of similar legal subjects is ensured, the pertinence and efficiency of data retrieval are improved, and irrelevant information interference is reduced, thereby providing an accurate data basis for controversy focus analysis. According to the controversy focus, the AI reply graph is analyzed, the AI reply level is determined, the processing capacity of the AI is accurately evaluated, the reliability of the AI in the controversy focus scenario is ensured, an objective basis is provided for the decision of whether to allocate a lawyer, and consultation errors or resource misallocation caused by insufficient AI capacity are avoided. According to the AI reply level, it is determined whether to allocate a lawyer, dynamic resource optimization is realized, case handling accuracy and user satisfaction are improved, lawyer resources are saved, and overall efficiency is improved. If it is determined to allocate a lawyer, the case handling lawyer is determined according to the consultant type, personalized lawyer matching is realized, user trust and consultation conversion rate are enhanced, and the lawyer allocation process is efficient and highly targeted.

[0063] In some embodiments, if the consultant type is a referral type, referral information is obtained; based on the referral information, lawyer information is determined; the lawyer information is analyzed to determine the field of expertise; based on the field of expertise and the legal subject, a lawyer matching degree is determined; and based on the lawyer matching degree, a case handling lawyer is determined.

[0064] The referral type can be a type of consultant who initiates consultation through recommendation.

[0065] The referral information can be data such as a referrer identifier related to the source of recommendation and a source channel.

[0066] The lawyer information can be professional data records about the lawyer.

[0067] The field of expertise can be the professional direction of the lawyer.

[0068] The lawyer matching degree can be a quantitative score of the matching degree between the lawyer and the legal subject of the case.

[0069] Specifically, when the consultant type is the referral type, the referral information is retrieved from the user input; then, based on the referral information, the lawyer information database is accessed; subsequently, a retrieval operation is performed in the database to match the lawyer information associated with the referral information. Furthermore, the classification code in the lawyer information is parsed to determine the field of expertise. Then, using a preset matching rule, it is determined whether the field of expertise covers the typical case type of the legal subject; subsequently, according to the matching result, the lawyer matching degree is determined. Finally, a preset threshold value is established according to the historical average matching degree; furthermore, the lawyer matching degree is compared with the preset threshold value, and if the lawyer matching degree exceeds the preset threshold value, the case handling lawyer is determined.

[0070] Through the scheme, if the consultant type is the referral type, the referral information is obtained, so that the matching process utilizes the recommendation relationship information and enhances personalized processing. According to the referral information, the lawyer information is determined to ensure that detailed data of the lawyer related to the recommendation is obtained, providing a basis for professional field analysis. The lawyer information is parsed to determine the field of expertise, ensuring that the matching is based on the actual expertise of the lawyer. According to the field of expertise and the legal subject, the lawyer matching degree is determined to quantify the matching degree between the lawyer and the case, making the matching decision objective and avoiding subjective bias. According to the lawyer matching degree, the case handling lawyer is determined to ensure efficient and accurate allocation of lawyers and improve the efficiency of case handling.

[0071] In some embodiments, the case description is subjected to natural language processing to extract entity information, expression of appeal and legal relationship; according to a preset legal element classification model, the legal subject corresponding to the entity information is determined; the expression of appeal and the legal relationship are analyzed to determine the focus of dispute.

[0072] The entity information can be a specific entity object.

[0073] The expression of appeal can be the core request of the user.

[0074] The legal relationship can be a legal contract relationship, a tort relationship and an ownership relationship.

[0075] The legal element classification model can be a preset classification model for classifying entity information.

[0076] Specifically, based on the case description, natural language processing such as word segmentation, part-of-speech tagging, and syntax analysis is performed; then, entity information is extracted through named entity recognition technology; at the same time, the expression of appeal is extracted through intent recognition technology; and the legal relationship is extracted through relationship extraction technology. Subsequently, the entity information is input into the pre-set legal element classification model; then, each entity information is classified to determine the corresponding legal subject. Then, the keywords and semantics in the expression of appeal are analyzed; then, the legal relationship is associated and cross-verified with the expression of appeal; finally, based on the pre-set rules established based on historical legal cases and general legal text patterns, the dispute focus is determined.

[0077] Through the scheme, the case description is subjected to natural language processing to extract entity information, expression of appeal, and legal relationship, ensuring that the case elements are fully captured. According to the pre-set legal element classification model, the legal subject corresponding to the entity information is determined to avoid entity misjudgment and ensure the correct identification of the legal subject, providing a reliable basis for case analysis and resource matching. By analyzing the expression of appeal and the legal relationship, the dispute focus is determined to ensure accurate case analysis and reduce information omission or incompleteness.

[0078] In some embodiments, the consultation history is analyzed to determine the consultation frequency and the consultation field; based on the consultation field, the consultation history is analyzed to determine the consultation depth; according to the consultation depth, the user's legal knowledge level is predicted; and according to the user's registration source, the legal knowledge level, and the consultation frequency, the type of the consultant is determined.

[0079] The consultation frequency can be the number of consultations of the user per unit time.

[0080] The consultation field can be a legal field tag that the user mainly consults.

[0081] The consultation depth can be a numerical value representing the depth of consultation.

[0082] The legal knowledge level can be a classification tag of the user's legal knowledge level.

[0083] Specifically, the consultation history is analyzed, and a plurality of consultation records of the user are read; then, the consultation records are sorted based on timestamps, and a consultation frequency in a unit time is calculated; meanwhile, a preset legal field classifier is applied to analyze a consultation field of the consultation records. Subsequently, based on the consultation field, a plurality of consultation records belonging to the consultation field are filtered out from the consultation history; further, a length and a detail degree of the consultation records are extracted; then, in combination with the consultation interaction data extracted from the historical log, a depth score of each consultation record is calculated; subsequently, the depth scores of the plurality of consultation records are averaged, so as to determine a consultation depth. Further, a preset threshold is established based on the average value of the statistical historical consultation data and an expert experience rule; then, the consultation depth is compared with the preset threshold, and a legal knowledge mastery degree of the user is predicted. Finally, based on a user registration source, in combination with the legal knowledge mastery degree and the consultation frequency; further, rule matching is performed through conditional judgment logic, so as to determine a consultant type.

[0084] Through the scheme, the consultation history is analyzed, the consultation frequency and the consultation field are determined, the consultation activity and the historical participation degree of the user are reflected, and the interest focus and the common legal problem direction of the user are revealed, so as to avoid generalization processing. Based on the consultation field, the consultation history is analyzed, the consultation depth is determined, irrelevant data interference is avoided, the pertinence and the accuracy of the analysis are improved, and the misjudgment risk is reduced. According to the consultation depth, the legal knowledge mastery degree of the user is predicted, the legal literacy level of the user is indicated, subjective judgment is avoided, and the completeness and the operability of the user portrait are enhanced. According to the user registration source, the legal knowledge mastery degree and the consultation frequency, the consultant type is determined, the fine identification of the user type is realized, the randomness of the type division is avoided, and personalized consultation processing is supported.

[0085] In some embodiments, based on a legal subject, a dispute focus is analyzed, a legal relationship complexity is determined; an AI question and answer database and an AI reply graph are analyzed, and a consultation refutation situation is determined; according to the consultation refutation situation, a reply accuracy rate is determined; according to the legal relationship complexity and the reply accuracy rate, an AI reply level is determined.

[0086] The legal relationship complexity can be a complexity level for quantifying a case structure complexity.

[0087] The consultation refutation situation can be user feedback data.

[0088] The reply accuracy rate can be a percentage numerical value quantifying AI reply reliability.

[0089] Specifically, according to the type of legal subject and the content of the dispute focus, classification processing is performed; then, according to the classification processing result, the number of legal subjects is counted and the relationship type in the dispute focus is analyzed to determine the complexity of the legal relationship. Subsequently, the AI question and answer database is accessed to retrieve historical question and answer records related to the current consultation; further, the AI reply graph is loaded, and the graph is traversed to extract the user reply part; then, the user reply content is analyzed, the refutation mark is identified, and the consultation refutation situation is determined. Subsequently, the consultation refutation situation is quantified, and the refutation situation is converted into a reply accuracy rate by applying a preset mapping rule. Finally, the complexity of the legal relationship and the reply accuracy rate are combined, and the AI reply level is calculated through conditional rules.

[0090] Through the scheme, based on the legal subject, the dispute focus is analyzed, the complexity of the legal relationship is determined, the overall difficulty level of the case is reflected, and the evaluation weight of the AI reply level is affected. The AI question and answer database and the AI reply graph are analyzed, the consultation refutation situation is determined, the real-time interaction evidence of the user to the AI reply is captured, and a basic data set is provided for the reply accuracy rate calculation, ensuring that the refutation situation reflects the real user behavior. According to the consultation refutation situation, the reply accuracy rate is determined, an objective index is provided for measuring the performance quality of the AI, and is used as an input parameter for the AI reply level calculation. According to the complexity of the legal relationship and the reply accuracy rate, the AI reply level is determined, ensuring that the resource allocation is based on the complexity of the case and the reliability of the AI, and the consultation efficiency is optimized.

[0091] In some embodiments, when the lawyer matching degree is lower than the preset matching threshold, based on the lawyer information, the collaboration network of the referral lawyer is obtained; the collaboration network is analyzed to determine the practice field intersection degree of the collaboration lawyer; according to the practice field intersection degree, the candidate lawyer is determined; the lawyer information of the candidate lawyer is obtained and analyzed to determine the case victory rate and the current case load of the candidate lawyer; according to the case victory rate and the current case load, the comprehensive matching coefficient is calculated; and according to the comprehensive matching coefficient, the case handling lawyer is determined.

[0092] The preset matching threshold can be a pre-set matching degree critical value. It is pre-stored in the server and called when used.

[0093] The referral lawyer can be a lawyer associated through a referral relationship.

[0094] The collaboration network can be a relationship network constituted by the practice field intersection degree between lawyers.

[0095] The collaboration lawyer can be another lawyer who has a cooperation relationship with the referral lawyer.

[0096] The practice field intersection degree can be the overlap ratio of the fields of expertise of two lawyers.

[0097] The candidate lawyer can be a candidate lawyer.

[0098] The case success rate can be a success ratio of historical cases of the lawyer.

[0099] The current case load can be a workload value of the current case handled by the lawyer.

[0100] The comprehensive matching coefficient can be a matching coefficient value index calculated based on the case success rate and the current case load.

[0101] Specifically, a preset matching threshold is set according to the historical matching success rate and industry experience value; then, the lawyer matching degree is compared with the preset matching threshold, if the lawyer matching degree is lower than the preset matching threshold, the collaboration network of the referral lawyer is retrieved from the stored lawyer information. Then, the collaboration network is analyzed to identify the collaboration lawyers; then, for each collaboration lawyer, the field of expertise is extracted from the lawyer information and compared with the field of expertise of the referral lawyer, so as to determine the practice field intersection degree of the collaboration lawyer. Subsequently, a preset intersection degree threshold is set based on the historical statistical analysis of the practice field intersection degree and expert knowledge; then, a number of collaboration lawyers are screened, only the lawyers with the practice field intersection degree score higher than the preset intersection degree threshold are retained, and are marked as candidate lawyers. Then, based on the lawyer information, for each candidate lawyer, the historical case data is extracted to determine the case success rate; at the same time, the current case record of the candidate lawyer is queried to count the number of cases and determine the current case load. Then, based on the case success rate and the current case load, a preset calculation rule is applied for comprehensive processing to calculate a comprehensive matching coefficient. Finally, the comprehensive matching coefficients of the candidate lawyers are compared, and the candidate lawyers are sorted; then, the candidate lawyer with the highest comprehensive matching coefficient is selected as the case handling lawyer.

[0102] Through the scheme, when the lawyer matching degree is lower than the preset matching threshold, the collaboration network of the referral lawyer is obtained based on the lawyer information, avoiding resource waste caused by initial matching failure, and ensuring that the matching process does not interrupt. The collaboration network is analyzed to determine the practice field intersection degree of the collaboration lawyer, reducing the interference of irrelevant lawyers, providing an objective basis for screening candidate lawyers, and improving the professional relevance of matching. According to the practice field intersection degree, the candidate lawyers are determined to ensure that the professional background of the candidate lawyers is highly consistent with the case demand, avoiding inefficient allocation caused by field mismatch, and optimizing the quality of the candidate pool. The lawyer information of the candidate lawyers is obtained and analyzed to determine the case success rate and the current case load of the candidate lawyers, ensuring that the matching process considers actual constraints and avoids overloading or inefficient lawyers. According to the case success rate and the current case load, the comprehensive matching coefficient is calculated to simplify the decision logic, facilitating the evaluation and sorting of the applicability of the candidate lawyers. According to the comprehensive matching coefficient, the case handling lawyer is determined to improve the case handling efficiency and success rate, avoiding the need for manual intervention, and realizing closed-loop resource optimization.

[0103] In some embodiments, the history cross-field case of the introduced lawyer is obtained; the history handled case is obtained based on the proficient field; the history handled case includes the average handling time; the handling difference of the cross-field case is determined according to the average handling time and the case handling time; the handling similarity is determined by analyzing the history cross-field case based on the dispute focus; the winning rate of the cross-field case is determined by counting the case handling result; the field adaptation coefficient is calculated according to the handling similarity and the winning rate; and the lawyer matching degree is determined according to the field adaptation coefficient, the proficient field and the legal subject.

[0104] The history cross-field case can be the past case record of the case handling time and the case handling result of the case not belonging to the proficient field handled by the introduced lawyer.

[0105] The case handling time can be the time length from acceptance to conclusion of the case.

[0106] The case handling result can be the final state of the case.

[0107] The history handled case can be the past case record of the average handling time of the introduced lawyer in the proficient field.

[0108] The average handling time can be the time length determined by calculating the average value of the handling time of the history handled case.

[0109] The handling difference can be the difference value between the case handling time and the average handling time.

[0110] The handling similarity can be the similarity score of the history cross-field case and the current case in the dispute focus.

[0111] The winning rate can be the proportion of the number of winning cases in the cross-field case to the total number of cases.

[0112] The field adaptation coefficient can be an index of the adaptation ability of the lawyer in similar cross-field cases.

[0113] Specifically, the history cross-field cases of the referral lawyer are retrieved from the lawyer information. Then, the history cross-field cases are parsed, and the case processing time length of each case is determined according to the time length from acceptance to conclusion of the case. At the same time, the case processing result is determined according to the final state of the case, whether the case is won or lost. Subsequently, the history processing cases in the specialized field of the referral lawyer are retrieved from the same database according to the specialized field of the referral lawyer. Further, the history processing cases are parsed, and the average processing time length of a plurality of related cases is calculated to determine the average processing time length. Then, the case processing time length of each history cross-field case is compared with the average processing time length, and the processing difference is determined by calculating the difference. Subsequently, based on the dispute focus, the processing similarity is determined by comparing the dispute focus and the case description of the history cross-field case through a text analysis technique. Then, the case processing results of the history cross-field cases are summarized, and the winning rate of the cross-field cases is determined by calculating the proportion of the number of winning cases to the total number of cases. Further, based on the processing similarity and the winning rate, the field adaptation coefficient is determined by comprehensive processing through a weighted average formula. Finally, whether the specialized field covers the field related to the legal subject is evaluated, and the field relevance of the referral lawyer is established according to the referral relationship. Subsequently, the lawyer matching degree is determined by combining the field adaptation coefficient and the field relevance through a preset matching rule.

[0114] By the scheme, the history cross-field cases of the referral lawyer are obtained, the actual performance data of the referral lawyer in the cross-field cases is ensured, and a data foundation is laid for calculating the processing difference and the winning rate and the like. Based on the specialized field, the history processing cases are obtained, a reference point of the work efficiency of the referral lawyer in the familiar field is provided, and the processing efficiency deviation of the cross-field cases is used for comparison. According to the average processing time length and the case processing time length, the processing difference of the cross-field cases is determined, the processing efficiency difference of the cross-field cases compared with the specialized field cases is revealed, and an efficiency dimension basis is provided for evaluating the cross-field adaptation ability of the lawyer. Based on the dispute focus, the history cross-field cases are analyzed, the processing similarity is determined, and the similarity of the history cases in the dispute focus with the current case is quantified, a similarity index is provided for calculating the field adaptation coefficient. The case processing results are counted, the winning rate of the cross-field cases is determined, and a performance basis is provided for evaluating the cross-field reliability. According to the processing similarity and the winning rate, the field adaptation coefficient is calculated, and the overall adaptation ability of the referral lawyer in processing the similar cross-field cases to the current case is quantified. According to the field adaptation coefficient, the specialized field and the legal subject, the lawyer matching degree is determined, and it is ensured that the matching result is based on the comprehensive judgment of the professional field coverage and the cross-field adaptation ability.

[0115] Figure 3 A structural schematic diagram of a legal affair consulting system based on AI interaction provided by an embodiment of the present application is shown in Figure 3The AI interaction-based legal matter consultation system 300 of the embodiment includes an information analysis module 301, a case analysis module 302, and a case handling determination module 303.

[0116] The information analysis module 301 is configured to obtain consultation information, analyze the consultation information, and determine a consultant type and a case description. The case analysis module 302 is configured to analyze the case description, and determine a legal subject and a dispute focus. The case handling determination module 303 is configured to determine a case handling lawyer according to the legal subject, the dispute focus, and the consultant type.

[0117] Optionally, when the information analysis module 301 analyzes the consultation information and determines the consultant type, the information analysis module 301 is configured to: analyze the consultation information, and determine a user registration source and a consultation history; determine the consultant type according to the user registration source and the consultation history.

[0118] Optionally, when the case handling determination module 303 determines the case handling lawyer according to the legal subject, the dispute focus, and the consultant type, the case handling determination module 303 is configured to: based on the legal subject, call an AI question and answer database, and determine an AI reply graph; analyze the AI reply graph according to the dispute focus, and determine an AI reply level; determine whether to assign a lawyer according to the AI reply level; if it is determined to assign a lawyer, determine a case handling lawyer according to the consultant type.

[0119] Optionally, the consultant type includes a referral type, and when the case handling determination module 303 determines the case handling lawyer according to the consultant type, the case handling determination module 303 is configured to: if the consultant type is the referral type, obtain referral information, and determine lawyer information according to the referral information; analyze the lawyer information, and determine a field of expertise; determine a lawyer matching degree according to the field of expertise and the legal subject; determine the case handling lawyer according to the lawyer matching degree.

[0120] Optionally, when the case analysis module 302 analyzes the case description and determines the legal subject and the dispute focus, the case analysis module 302 is configured to: perform natural language processing on the case description, and extract entity information, a statement of claim, and a legal relationship; determine a legal subject corresponding to the entity information according to a preset legal element classification model; Analyze the appeal expression and the legal relationship, and determine the dispute focus.

[0121] Optionally, when the information analysis module 301 determines the consultant type according to the user registration source and the consultation history, is used for: Analyzing the consultation history to determine the consultation frequency and the consultation field; Based on the consultation field, analyze the consultation history to determine the consultation depth; According to the consultation depth, predict the user's legal knowledge level; According to the user registration source, the legal knowledge level, the consultation frequency, determine the consultant type.

[0122] Optionally, when the case determination module 303 determines the AI reply level according to the dispute focus, analyzes the AI reply graph, is used for: Based on the legal subject, analyze the dispute focus to determine the legal relationship complexity; Analyzing the AI question and answer database and the AI reply graph to determine the consultation refutation; According to the consultation refutation, determine the reply accuracy; According to the legal relationship complexity and the reply accuracy, determine the AI reply level.

[0123] Optionally, when the case determination module 303 determines the case lawyer according to the lawyer matching degree, is used for: When the lawyer matching degree is lower than the preset matching threshold, based on the lawyer information, the cooperation network of the introduced lawyer is obtained; Analyze the cooperation network to determine the cross degree of the practice field of the cooperation lawyer; According to the cross degree of the practice field, determine the alternative lawyer; Get and analyze the lawyer information of the alternative lawyer, determine the case success rate of the alternative lawyer and the current case load; According to the case success rate and the current case load, calculate the comprehensive matching coefficient; According to the comprehensive matching coefficient, determine the case lawyer.

[0124] Optionally, when the case determination module 303 determines the lawyer matching degree according to the expertise field and the legal subject, is used for: Get the historical cross-field case of the introduced lawyer; the historical cross-field case contains the case processing time and the case processing result; Based on the expertise field, obtain the historical processing case; the historical processing case includes the average processing time; According to the average processing time and the case processing time, determine the processing difference of cross-domain cases; Based on the dispute focus, analyze the historical cross-domain cases to determine the processing similarity; Statistical analysis of the case processing results, determine the success rate of the cross-domain cases; According to the processing similarity and the success rate, calculate the field adaptation coefficient; According to the field adaptation coefficient, the field of expertise and the legal subject, determine the lawyer matching degree.

[0125] The system of the embodiment can be used to execute the method of any of the above embodiments, and the implementation principles and technical effects are similar, which will not be described here.

Claims

1. A legal affairs consulting method based on AI interaction, characterized in that: include: Obtain consultation information; analyze the consultation information to determine the consultant type and case description; Analyze the case description to determine the legal subject and the focus of the dispute; The lawyer to handle the case will be determined based on the legal entity, the focus of the dispute and the type of consultant.

2. The method according to claim 1, characterized in that The step of analyzing the consultation information and determining the type of the consultant includes: Analyze the consultation information to determine the user's registration source and consultation history; The consultant type is determined based on the user registration source and the consultation history.

3. The method according to claim 1, characterized in that Determining the lawyer to handle the case based on the legal entity, the focus of the dispute, and the type of consultant includes: Based on the legal entity, retrieve the AI ​​question and answer database to determine the AI ​​answer graph; Analyze the AI ​​response map based on the dispute focus to determine the AI ​​response level; Determining whether to assign a lawyer based on the AI ​​response level; If a lawyer is assigned, the lawyer who will handle the case will be determined based on the type of consultant.

4. The method according to claim 3, characterized in that The consultant type includes a referral type, and the case lawyer is determined based on the consultant type, including: If the consultant type is the referral type, obtaining referral information; and determining lawyer information based on the referral information; Analyze the lawyer's information and determine their areas of expertise; Determine the lawyer's matching degree based on the areas of expertise and the legal subjects; The lawyer handling the case is determined based on the lawyer matching degree.

5. The method according to claim 1, wherein The analysis of the case description and determination of the legal subject and the focus of the dispute include: Perform natural language processing on the case description to extract entity information, claim statements, and legal relationships; Determine the legal subject corresponding to the entity information based on a preset legal element classification model; Analyze the statement of the claim and the legal relationship to determine the focus of the dispute.

6. The method according to claim 2, characterized in that Determining the consultant type based on the user registration source and the consultation history includes: Analyze the consultation history to determine the frequency and areas of consultation; Analyze the consultation history based on the consultation area to determine the depth of consultation; Based on the depth of the consultation, predict the user's level of legal knowledge; The consultant type is determined based on the user's registration source, the level of legal knowledge mastered, and the consultation frequency.

7. The method according to claim 3, characterized in that Analyzing the AI ​​response map based on the dispute focus to determine the AI ​​response level includes: Based on the legal entities, analyze the focus of the dispute and determine the complexity of the legal relationship; Analyze the AI ​​question-answer database and the AI ​​response graph to determine the consultation and rebuttal situation; Determine the accuracy of the response based on the consultation and rebuttal; The AI ​​response level is determined based on the complexity of the legal relationship and the accuracy of the response.

8. The method according to claim 4, characterized in that Determining the lawyer to handle the case based on the lawyer matching degree includes: When the lawyer matching degree is lower than a preset matching threshold, obtaining a collaborative network of referring lawyers based on the lawyer information; Analyze the collaborative network to determine the degree of overlap in the collaborative lawyers' practice areas; Identify potential lawyers based on the overlap between the practice areas; Obtaining and analyzing attorney information of the candidate lawyers to determine the case winning rate and current case load of the candidate lawyers; Calculating a comprehensive matching coefficient based on the case win rate and the current case load; The lawyer to handle the case is determined based on the comprehensive matching coefficient.

9. The method according to claim 4, characterized in that The determination of lawyer matching based on the areas of expertise and the legal subject includes: Obtain the referral lawyer's historical cross-disciplinary cases; the historical cross-disciplinary cases include case handling time and case handling results; Based on the areas of expertise, obtain historical cases; the historical cases include average processing time; Determine the differences in handling cross-disciplinary cases based on the average handling time and the case handling time; Based on the focus of the dispute, analyze the historical cross-disciplinary cases and determine the similarity of their treatment; Collect statistics on the case handling results and determine the winning rate of the cross-field cases; Calculating a domain adaptation coefficient based on the processing similarity and the winning rate; The lawyer matching degree is determined based on the field adaptability coefficient, the field of expertise and the legal subject.

10. A legal affairs consulting system based on AI interaction, characterized in that: The method as claimed in any one of claims 1 to 9 comprises: An information analysis module is used to obtain consultation information; analyze the consultation information to determine the consultant type and case description; The case analysis module is used to analyze the case description and determine the legal subject and the focus of the dispute; The case handling determination module is used to determine the lawyer handling the case based on the legal entity, the focus of the dispute and the type of consultant.

Images