Insurance policy auditing method and device, equipment and medium

Abstract

The invention relates to a data processing technology, can be applied to business system platforms of medical health, financial science and technology and the like, and discloses an insurance policy auditing method, device and equipment and a medium, and the method comprises the steps: obtaining an insurance policy to be audited, carrying out the data extraction of the insurance policy to be audited, and generating structured insurance policy data; matching the structured insurance policy data with an auditing strategy in an auditing model library to obtain a matched auditing strategy; based on the auditing strategy, auditing the structured insurance policy data to generate an auditing result; and generating a corresponding risk level report based on a preset risk level rule and the audit result, and displaying the risk level report. The method can automatically extract the structured data of the insurance policy, thoroughly gets rid of the dependence of artificial experience, is different from single-dimensional forward verification, performs deep logic deduction from multiple angles, combines sequential logic verification, can accurately recognize hidden risks which are difficult to cover by a traditional system, and improves the auditing efficiency.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an insurance policy review method, device, equipment and medium. Background Art

[0002] Insurance is a financial instrument that provides financial protection to individuals, families or businesses through a risk transfer mechanism. Its core value lies in coping with uncertain risks and reducing potential economic losses.

[0003] In the healthcare sector, an insurance policy is a legally binding contract between an insurer and the policyholder, providing coverage for medical expenses. Medical insurance policy types include hospitalization insurance, outpatient insurance, critical illness insurance, and accidental insurance. Medical insurance policies include information such as the policy details, policy duration, insured amount, and premium. The coverage details which medical expenses are eligible for reimbursement, such as hospitalization, surgery, and medication costs.

[0004] In the financial sector, insurance policies vary depending on the insured object and risk type, including life insurance, health insurance, property insurance, liability insurance, investment insurance, and group insurance. Life insurance policies insure the life of the insured. Upon the insured's death during the policy period or upon the fulfillment of contractually agreed conditions, the insurance company pays the beneficiary a benefit, typically including death benefits, survival benefits, and disability benefits. Health insurance policies focus on protecting the insured's medical expenses in the event of illness or accidental injury, covering hospitalization, surgery, and outpatient expenses. They typically include provisions such as medical expense reimbursement and critical illness coverage. Property insurance policies primarily cover personal or business property, such as homes, vehicles, and equipment, providing compensation in the event of loss. They typically include provisions such as property damage compensation and theft compensation. Liability insurance policies address the insured's legal liability, including bodily injury or property damage to third parties caused by fault or negligence, and typically include compensation limits and deductibles. Investment-type insurance policies not only provide basic insurance coverage but also include investment features designed to achieve capital appreciation. These policies typically include participating, universal, and unit-linked policies. Group insurance policies are typically purchased by employers for their employees, covering multiple insureds. They are intended for specific groups or organizations and typically include standardized benefits and premium rates.

[0005] During the insurance application process, the policyholder must sign an insurance policy with the insurance company. The policy serves as proof of the insurance company's commitment to the policyholder and a contract between the two parties, specifically outlining the rights and obligations of both parties. Policy review is a crucial step in the insurance business, ensuring the authenticity, completeness, compliance, and accuracy of the policy to safeguard the smooth operation of the insurance business and the legitimate rights and interests of both clients and the company.

[0006] Existing insurance policy reviews mainly rely on manual experience, which has problems such as low efficiency, high cost, and high risk of subjective errors. Especially when regulatory rules are frequently updated, it is difficult to ensure the consistency of review standards. Summary of the Invention

[0007] In view of the above-mentioned deficiencies in the prior art, the present invention provides a policy review method, device, equipment and medium, which aims to solve the problems in the prior art that the review of insurance documents mainly relies on manual experience, has low efficiency, high cost, and high risk of subjective errors, especially when regulatory rules are frequently updated, it is difficult to ensure the consistency of review standards.

[0008] The technical solutions of the present invention are as follows:

[0009] A first embodiment of the present invention provides a policy review method, the method comprising:

[0010] Obtaining insurance policies to be reviewed, extracting data from the insurance policies to be reviewed, and generating structured insurance policy data;

[0011] Matching the structured policy data with the audit policies in the audit model library to obtain matching audit policies;

[0012] Auditing the structured policy data based on the audit strategy to generate an audit result;

[0013] Based on the preset risk level rules and the audit results, a corresponding risk level report is generated and displayed.

[0014] Another embodiment of the present invention provides an insurance policy review device, comprising:

[0015] A data extraction module is used to obtain the insurance policies to be reviewed, extract data from the insurance policies to be reviewed, and generate structured insurance policy data;

[0016] A data matching module, configured to match the structured policy data with the audit policies in the audit model library to obtain matching audit policies;

[0017] A data review module, configured to review the structured policy data based on the review strategy and generate a review result;

[0018] The risk report generation module is used to generate a corresponding risk level report based on the preset risk level rules and the audit results, and display the risk level report.

[0019] Another embodiment of the present invention provides a computer device, the computer device comprising at least one processor; and

[0020] a memory communicatively connected to the at least one processor; wherein,

[0021] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the above-mentioned insurance policy review method.

[0022] Another embodiment of the present invention further provides a computer-readable storage medium, which stores computer-executable instructions. When the computer-executable instructions are executed by one or more processors, the one or more processors can execute the steps of the above-mentioned insurance policy review method.

[0023] Beneficial effects: The insurance policy review method, device, equipment and medium of the embodiment of the present invention can be used through the insurance policy review method, device, equipment and medium, including: obtaining the insurance policy to be reviewed, extracting data from the insurance policy to be reviewed, and generating structured insurance policy data; matching the structured insurance policy data with the audit strategy in the audit model library to obtain the matching audit strategy; auditing the structured insurance policy data based on the audit strategy to generate an audit result; generating a corresponding risk level report based on the preset risk level rules and the audit result, and displaying the risk level report. The embodiment of the present invention completely gets rid of the reliance on manual experience by automatically extracting the structured data of the insurance policy, and is different from the single-dimensional forward verification. It performs in-depth logical deduction from multiple angles and combines with temporal logic verification to accurately identify hidden risks that are difficult for traditional systems to cover, thereby improving the audit efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments of the present invention. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0025] Figure 1 A schematic diagram of an application environment of an embodiment of a policy review method of the present invention;

[0026] Figure 2 This is a flow chart of a preferred embodiment of a policy review method of the present invention;

[0027] Figure 3 This is a schematic diagram of the functional modules of a preferred embodiment of an insurance policy review device of the present invention;

[0028] Figure 4A schematic structural diagram of a preferred embodiment of a computer device of the present invention;

[0029] Figure 5 This is another structural diagram of a preferred embodiment of a computer device of the present invention. DETAILED DESCRIPTION

[0030] To make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail below. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0031] The following describes the embodiments of the present invention in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0032] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, and this is merely a way of distinguishing the objects of the same attributes when describing them in the embodiments of the present application. Here, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0033] The insurance policy review method provided by the embodiment of the present invention can be applied in the following situations: Figure 1In an application environment, the client communicates with the server through a network. The client accesses the network or business platform of the server, and the server can obtain the insurance policy to be reviewed, extract data from the insurance policy to be reviewed, and generate structured insurance policy data; match the structured insurance policy data with the audit strategy in the audit model library to obtain a matching audit strategy, wherein the audit strategy includes reverse verification parameters, sequential logic verification parameters, compliance verification parameters, and insurance data verification parameters; based on the audit strategy, the structured insurance policy data is audited to generate an audit result; based on the preset risk level rules and the audit result, a corresponding risk level report is generated, and the risk level report is displayed. In the present invention, by automatically extracting the structured data of the insurance policy, it completely gets rid of the dependence on manual experience, and is different from the single-dimensional forward verification. In-depth logical deduction is performed from multiple angles, combined with sequential logic verification, it can accurately identify hidden risks that are difficult for traditional systems to cover, thereby improving the audit efficiency. Among them, the client can be, but is not limited to, various personal computers, laptops, smart phones, tablet computers, and portable wearable devices. The server can be implemented with an independent server or a server cluster consisting of multiple servers. The present invention is described in detail below through specific examples.

[0034] In response to the above problems, the present invention provides a policy review method, please refer to Figure 2 , Figure 2 This is a flow chart of a preferred embodiment of the insurance policy review method of the present invention. Figure 2 As shown, it includes:

[0035] Step S100: Obtain the insurance policy to be reviewed, extract data from the insurance policy to be reviewed, and generate structured insurance policy data.

[0036] The embodiments of the present invention are mainly aimed at the inability to flexibly adapt to the characteristics of different insurance products (such as differentiated rules for life insurance and health insurance) during policy review in the existing technology, and the inability to synchronize the latest regulatory rules in real time. It is used to review insurance policies submitted by users. It mainly automatically extracts structured data from insurance application documents. Structured policy data includes but is not limited to identity card number, effective date, insurance type information, insurance plan, insurance amount, premium, etc.

[0037] In the medical field, medical insurance policies include insurance coverage: clearly listing which medical expenses can be compensated by insurance, such as hospitalization expenses, surgical expenses, drug expenses, etc. Insurance period: stipulates the validity period of the insurance contract, that is, the time period during which the insurer assumes insurance liability. Insurance amount: stipulates the maximum compensation limit that the insurer will bear for the insured due to medical matters during the insurance period. Insurance premium: the fee that the insured needs to pay to the insurer in exchange for insurance coverage. Other agreements: including other terms agreed upon by both parties, such as deductibles, waiting periods, and claims procedures. For example, the commonly used terms of medical insurance mainly include deductible clauses, proportional payment clauses, and payment limit clauses.

[0038] In the financial sector, insurance policy content includes basic information, insurance liability clauses, cost-related clauses, analysis of special clauses, and key points in the claims process. Basic information includes the contract code, policyholder information, insured information, beneficiary information, and the policy period. Insurance liability clauses include core coverage and liability exemptions. Cost-related clauses include data such as premiums, grace periods, cash values, and policy loans. Analysis of special clauses includes, but is not limited to, waiting periods, cooling-off periods, and reinstatement clauses. Key points in the claims process include timelines, document lists, and review deadlines.

[0039] Step S100, i.e. obtaining the insurance policy to be reviewed, extracting data from the insurance policy to be reviewed, and generating structured insurance policy data, includes:

[0040] Step S101: When a policy review instruction is detected, obtain the policy to be reviewed;

[0041] Step S102: Determine whether the type of the policy to be reviewed is text or picture. If the type of the policy to be reviewed is text, execute step S103; if the type of the policy to be reviewed is picture, execute step S104;

[0042] Step S103: extract keywords from the insurance policy to be reviewed to generate structured insurance policy data;

[0043] Step S104: perform image recognition on the insurance policy to be reviewed to obtain text content in the image, perform keyword extraction on the text content, and generate structured insurance policy data.

[0044] Users can submit policies for review through the insurance submission system. The policy review instruction is triggered by, for example, triggering a review button. After detecting the policy review instruction, the policy for review is retrieved. The policy may be text data submitted online or a paper policy signed offline. The system determines whether the policy for review is text or image. If the policy is text, keyword extraction technology is used to generate structured policy data.

[0045] If the type of insurance policy to be reviewed is a picture, an OCR recognition algorithm or an image recognition model is used to perform image recognition, and then keywords are extracted from the text content to generate structured insurance policy data. Among them, OCR (Optical Character Recognition) is a core technology that converts text information in an image into editable and searchable machine text. The OCR algorithm is specifically as follows: Image preprocessing: After acquiring the image through a scanner, digital camera or other device, the image needs to be grayscaled, binarized, denoised and other processes to eliminate the impact of factors such as uneven lighting and background interference on recognition accuracy. For example, converting a color image into a black and white binary image can highlight the contrast between the text area and the background. Text detection and segmentation: Use deep learning models (such as CRNN, EAST) to detect text areas in the image and segment continuous text into single characters or words. This stage needs to solve problems such as text direction correction and tilt adjustment to ensure clear character boundaries. Feature extraction and recognition: Traditional OCR relies on manually designed features (such as stroke structure and edge density), while modern methods automatically learn character features through convolutional neural networks (CNN). The recognition process combines pattern matching and semantic analysis, for example, using lexical constraints to correct easily confused characters like "0" and "O." Post-processing optimization: Recognition results are grammatically checked and contextually corrected, and structured data (such as tables and paragraphs) are output. Some systems incorporate natural language processing (NLP) technology to further enhance semantic coherence.

[0046] Among them, step S104, that is, if the type of the insurance policy to be reviewed is an image, then performing image recognition on the insurance policy to be reviewed to obtain text content in the image, performing keyword extraction on the text content, and generating structured insurance policy data, includes:

[0047] Step S1041: constructing an image recognition model based on a deep learning network;

[0048] Step S1042: If the type of the policy to be reviewed is an image, obtain a policy image corresponding to the policy to be reviewed, input the policy image into the image recognition model, and obtain the text content in the image based on the recognition result of the image recognition model;

[0049] Step S1043: extract keywords from the text content to generate structured insurance policy data.

[0050] Build a deep learning model. Classic model architectures include the CNN series, such as LeNet-5 (handwritten digit recognition), and other models. Access public datasets (such as CIFAR-10 and ImageNet), crawlers, and proprietary business data. Obtain annotation tools such as LabelImg (object detection), Labelme (semantic segmentation), and CVAT (an enterprise-level annotation platform). Generate data samples using data augmentation from public datasets. Data augmentation includes basic operations such as rotation, flipping, scaling, and cropping, as well as advanced strategies such as MixUp (sample mixing), CutMix (region replacement), and AutoAugment (automatic search augmentation strategy).

[0051] Set hyperparameter tuning: learning rate: use a learning rate scheduler (such as CosineAnneal ingLR); batch size: adjust according to video memory (usually 32 / 64 / 128); regularization: L2 regularization (weight decay), Dropout (random inactivation).

[0052] A deep learning model is trained based on dataset samples and hyperparameters to generate an image recognition model. The policy image corresponding to the policy to be reviewed is obtained and fed into the image recognition model. Based on the model's recognition results, the text content within the image is obtained. Keyword extraction is then performed on the text content to generate structured policy data.

[0053] Step S200: Match the structured policy data with the audit strategy in the audit model library to obtain the matching audit strategy.

[0054] The audit model library contains pre-configured audit strategies. Multiple audit strategies can be configured within the audit strategy. After obtaining structured policy data, the audit model library can be queried to determine if there are corresponding rules within the audit model library. If a matching audit strategy is found, it is used as the policy audit strategy. Audit strategies include, but are not limited to, reverse verification parameters, temporal logic verification parameters, compliance verification parameters, and policy data verification parameters. For example, reverse verification parameters include regulatory compliance (such as the legitimacy of the policyholder's qualifications), product compatibility (such as whether the insured product covers designated illnesses), and reasonableness of the risk coverage (such as whether the cumulative coverage exceeds the limit). Temporal logic verification parameters include parameters such as detecting underwriting dates prior to the effective date and detecting unusual behavior such as multiple short-term policies by the same policyholder. Compliance verification parameters verify policy data against regulatory red lines. Policy data verification parameters primarily verify the accuracy of policy data, including recalculation of insured sums and premiums, post-claim settlement policy type status, and compliance with the policy application and underwriting time sequence.

[0055] Step S200, that is, matching the structured policy data with the audit policy in the audit model library to obtain the matching audit policy, includes:

[0056] Step S201: pre-build an audit model library, wherein the audit strategy of the audit model library includes product insurance rules, regulatory rules, product terms and risk insurance amount rules;

[0057] Step S202: Obtaining policy parameters from the structured policy data, and assembling the policy parameters to generate assembled parameters;

[0058] Step S203: matching the assembly parameters with the audit policies in the audit model library, and obtaining corresponding audit policies according to the matching results.

[0059] In the embodiment of the present invention, the pre-built audit model library is a dynamic rule library that supports real-time updating and loading of product insurance rules, regulatory rules, regulatory red lines, product terms (such as deductible limits), and risk coverage rules (such as the cumulative coverage limit). The logical relationship between association rules (such as "When the insured is over 60 years old, the risk coverage threshold is automatically reduced by 30%"), underage risk coverage, personal insurance risk coverage limit, etc. The audit model library supports common queries and makes it easy to see which products cover which rules. It is displayed clearly according to the knowledge graph.

[0060] The insurance knowledge graph is a structured semantic knowledge base that aggregates a large amount of insurance domain knowledge and describes concepts, entities and their relationships in the form of a graph structure, providing knowledge management and decision support for the insurance industry.

[0061] The data sources for constructing the insurance knowledge graph include multi-source heterogeneous data, including insurance contract texts, business process records, customer information, industry reports, etc. This data undergoes rigorous cleaning and pre-processing to ensure accuracy and consistency, providing a reliable data foundation for the knowledge graph.

[0062] Construction method: Knowledge extraction: extract key information from the contract text, such as terms, responsibilities, fees, etc.; extract operating specifications and restrictions from the business process. This is a key step in converting scattered business data into structured knowledge. Semantic understanding: through natural language processing technology, understand the deep meaning of the text and provide rich semantic support for the construction of the knowledge graph. For example, accurately parse the professional terms in insurance terms. Knowledge fusion: Integrate data from different sources, handle inconsistencies, duplications and missing issues in the data, and ensure the overall quality of the knowledge graph. Knowledge storage: Use graph databases (such as Neo4j, JanusGraph, etc.) or other technologies suitable for storing and querying large-scale graph data to build a knowledge graph to ensure efficient query performance and scalability.

[0063] Step S300: Audit the structured policy data based on the audit strategy and generate an audit result.

[0064] After matching the audit strategy, the structured data is checked against the audit strategy. If all structured policy data meets the audit strategy, the audit is completed and a positive audit result is generated, indicating that there are no exceptions in the policy audit. If the structured policy data does not meet the audit strategy, the policy audit is considered to be abnormal.

[0065] In step S300, the structured policy data is audited based on the audit strategy to generate an audit result, including:

[0066] Step S301: Obtain reverse check parameters, sequential logic check parameters, compliance check parameters, and insurance data check parameters in the audit policy;

[0067] Step S302: Verify the structured policy data based on the reverse verification parameter to generate a first verification result;

[0068] Step S303: Verify the structured policy data based on the temporal logic verification parameter to generate a second verification result;

[0069] Step S304: Verify the structured policy data based on the compliance verification parameters to generate a third verification result;

[0070] Step S305: Verify the structured policy data based on the insurance data verification parameter to generate a fourth verification result;

[0071] Step S306: Combine the first verification result, the second verification result, the third verification result, and the fourth verification result to generate an audit result.

[0072] Obtain the reverse verification parameters, timing logic verification parameters, compliance verification parameters, and insurance data verification parameters in the audit policy.

[0073] A multi-dimensional reverse check generates the first verification result. Specifically, reverse logic deduction is conducted from dimensions such as regulatory compliance (such as the legality of the insured's qualifications), product compatibility (such as whether the insured product covers designated diseases), and the rationality of the risk coverage (such as whether the cumulative coverage exceeds the limit).

[0074] The second verification result is generated by sequential logic verification. Specifically, it analyzes the insurance behavior based on the time axis (such as the underwriting time being earlier than the effective time, and the abnormal behavior detection of the same insured person taking out multiple insurance policies in a short period of time).

[0075] Compliance checks generate third-party verification results. This checks whether regulatory red lines have been exceeded. If a customer exceeds these limits by a significant amount, it identifies the customer as a potential insurance risk and flags such customers.

[0076] The insurance data is checked and the fourth verification result is generated. The fourth verification result includes the verification of the correctness of the insurance data, including the recalculation of the insurance premium, the status of the insurance type after the claim is settled, and whether the insurance application time and underwriting time sequence are consistent.

[0077] The first verification result, the second verification result, the third verification result and the fourth verification result are combined to generate a final audit result.

[0078] Step S400: Generate a corresponding risk level report based on the preset risk level rules and the audit results, and display the risk level report.

[0079] A risk level report is generated in advance based on the audit results and corresponding conditions. For example, the risk level can be set as high, medium, or low. The risk type corresponding to each level of the risk level report can be configured as needed. The risk level report is displayed on the visual terminal. High, medium, and low risk labels can also be generated in the risk level report. For example, high risk is marked in red, medium risk is marked in yellow, and low risk is marked in green.

[0080] Step S400, i.e., generating a corresponding risk level report based on the preset risk level rules and the audit results, and displaying the risk level report, includes:

[0081] Step S401: Pre-set the risk level and corresponding audit parameters to generate risk level rules;

[0082] Step S402: matching the audit result with the preset risk level rules;

[0083] Step S403: Generate a corresponding risk level report based on the matching result, and display the risk level report on a visualization panel.

[0084] Risk levels and corresponding audit parameters are set in advance to generate risk level rules. Insurance product risk levels are usually expressed as high, medium, or low. As the risk level increases, the risk the user needs to bear also increases.

[0085] A low risk rating indicates that the overall risk level of the product is low, or that although there are some factors that may adversely affect the safety of the product's premiums paid and returns, the possibility of the product incurring losses is low.

[0086] Medium-risk products have a moderate overall risk level, with certain fluctuations in profits and a greater possibility of losses.

[0087] High-risk products have a high overall risk level and a high possibility of product losses.

[0088] Step S403, i.e. generating a corresponding risk level report based on the matching result and displaying the risk level report on a visualization panel, includes:

[0089] Step S4031: Generate a corresponding risk level report based on the matching result, and determine whether the risk level in the risk level report is greater than or equal to a preset risk level threshold. If the risk level in the risk level report is greater than or equal to the preset risk level threshold, execute step S4032; if the risk level in the risk level report is less than the preset risk level threshold, execute step S4033;

[0090] Step S4032: triggering an early warning operation in real time, sending the early warning information to the management terminal, and displaying the risk level report and policy modification suggestions on a visual panel;

[0091] Step S4033: Display the risk level report on a visualization panel.

[0092] When high-risk orders are confirmed, early warnings are triggered and the verification results (such as rule conflict details) are displayed through a visual panel.

[0093] Common types of policy anomalies include: Abnormalities during the insurance application process; False information: such as forged IDs, income certificates, or health status certificates; Duplicate insurance: The same customer purchases high-value life insurance policies from multiple companies (possibly for suicide fraud); and Correlated risks: There is a transfer of benefits between the policyholder and the beneficiary (e.g., undisclosed family relationships).

[0094] Abnormal policy duration. Abnormal premium payment: Long-term non-payment without triggering surrender, or frequent changes in payment methods. Abnormal policy changes: For example, multiple changes to beneficiaries or reductions in coverage within a short period of time. Abnormal customer behavior: For example, a high-risk customer suddenly requests to cancel the policy (possibly transferring risk).

[0095] Abnormal claims processing. False claims: Forged accident certificates, exaggerated losses (e.g., falsely reporting repair costs for auto insurance). High-frequency claims: Frequent small claims from the same client or related parties (possibly insurance fraud). Temporal anomalies: If an insurance claim occurs shortly after purchase (possibly due to a pre-existing medical condition).

[0096] Compliance and operational risks. Policy violations: Policy terms and conditions that do not comply with regulatory requirements (e.g., premium rates exceeding the limit). System operation anomalies: Policy information illegally tampered with. Corrective action suggestions are provided (e.g., supplemental documentation or adjustments to the insured amount).

[0097] The alert trigger and response method is as follows: Through graded alerts: different response mechanisms are set according to the risk level (high / medium / low). If the risk is low: automatic reminders (such as premium overdue notices) are sent. High risk: the policy is frozen, manual review is initiated, or an alarm is issued.

[0098] In some other embodiments, individual or batch verification is also supported, and relevant data can be retrieved according to information such as insurance time, insurance type, age, etc.

[0099] Compared with the existing technology, the embodiments of the present invention have the following technical advantages:

[0100] Intelligence and automation significantly improve efficiency. The system uses a rule engine module to achieve real-time updates and logical associations in the dynamic rule base. Combined with a data assembly module, it automatically extracts structured data from insurance documents, completely eliminating reliance on manual experience.

[0101] Multi-dimensional reverse verification accurately identifies complex risks. Unlike single-dimensional forward verification, the reverse verification module conducts in-depth logical deduction from perspectives such as regulatory compliance, product compatibility, and the rationality of risk coverage. Combined with temporal logic verification (such as abnormal underwriting time sequence and multiple insurance policies within a short period of time), it can accurately identify hidden risks that are difficult for traditional systems to cover.

[0102] Dynamic rule adaptation and knowledge graph visualization ensure compliance consistency. The rule engine supports flexible configuration of differentiated rules for different insurance products (such as regulatory requirements for life and health insurance) and real-time synchronization of the latest policies (such as the dynamic compliance requirements of the China Banking and Insurance Regulatory Commission).

[0103] Full-process risk warning and decision support. High-risk documents are marked in real time through a visual dashboard, providing conflict details and correction suggestions (such as supplementary documents or adjustment of insurance coverage), helping business personnel respond quickly to issues.

[0104] Scalability and cost optimization. Through modular design, it can adapt to the needs of insurance institutions of different sizes. It can handle both individual document verification and batch processing of millions of data points, reducing the cost of software and hardware iteration for enterprises.

[0105] It should be noted that there is not necessarily a certain order between the above steps. A person skilled in the art can understand, based on the description of the embodiments of the present invention, that in different embodiments, the above steps may have different execution orders, that is, they may be executed in parallel, or may be executed interchangeably, etc.

[0106] Another embodiment of the present invention provides a policy review device, which corresponds to the policy review method of the above embodiment. Figure 3 As shown, the device 1 includes:

[0107] The data extraction module 100 is used to obtain the insurance policies to be reviewed, extract data from the insurance policies to be reviewed, and generate structured insurance policy data;

[0108] The data matching module 200 is used to match the structured policy data with the audit policies in the audit model library to obtain matching audit policies;

[0109] A data review module 300 is configured to review the structured policy data based on the review strategy and generate a review result;

[0110] The risk report generation module 400 is used to generate a corresponding risk level report based on preset risk level rules and the audit results, and display the risk level report.

[0111] The specific implementation method is shown in the method embodiment, which will not be repeated here.

[0112] In one embodiment, the data extraction module 100 is specifically configured to:

[0113] When a policy review instruction is detected, obtain the policy to be reviewed;

[0114] Determine whether the type of the insurance policy to be reviewed is text or image;

[0115] If the type of the policy to be reviewed is text, perform keyword extraction on the policy to be reviewed to generate structured policy data;

[0116] If the type of the insurance policy to be reviewed is a picture, image recognition is performed on the insurance policy to be reviewed to obtain text content in the image, and keywords are extracted from the text content to generate structured insurance policy data.

[0117] The specific implementation method is shown in the method embodiment, which will not be repeated here.

[0118] In one embodiment, the data matching module 200 is specifically configured to:

[0119] Pre-build an audit model library, the audit strategy of which includes product insurance rules, regulatory rules, product terms and risk insurance amount rules;

[0120] Obtaining policy parameters from the structured policy data, and assembling the policy parameters to generate assembled parameters;

[0121] Matching is performed based on the assembly parameters with the audit policies in the audit model library, and obtaining corresponding audit policies according to the matching results.

[0122] The specific implementation method is shown in the method embodiment, which will not be repeated here.

[0123] In one embodiment, the data review module 300 is specifically configured to:

[0124] Obtaining reverse verification parameters, timing logic verification parameters, compliance verification parameters, and insurance data verification parameters in the audit policy;

[0125] Verifying the structured policy data based on the reverse verification parameter to generate a first verification result;

[0126] Verifying the structured policy data based on the temporal logic verification parameter to generate a second verification result;

[0127] Verifying the structured policy data based on the compliance verification parameters to generate a third verification result;

[0128] Verifying the structured policy data based on the insurance data verification parameter to generate a fourth verification result;

[0129] The first verification result, the second verification result, the third verification result and the fourth verification result are combined to generate an audit result.

[0130] The specific implementation method is shown in the method embodiment, which will not be repeated here.

[0131] In one embodiment, the risk report generation module 400 is specifically configured to:

[0132] Set the risk level and corresponding audit parameters in advance to generate risk level rules;

[0133] Matching the audit results with the preset risk level rules;

[0134] A corresponding risk level report is generated according to the matching result, and the risk level report is displayed on a visualization panel.

[0135] The specific implementation method is shown in the method embodiment, which will not be repeated here.

[0136] In one embodiment, the risk report generation module 400 is further configured to:

[0137] Generate a corresponding risk level report based on the matching result, and determine whether the risk level in the risk level report is greater than or equal to a preset risk level threshold;

[0138] If the risk level in the risk level report is greater than or equal to the preset risk level threshold, an early warning operation is triggered in real time, the early warning information is sent to the manager terminal, and the risk level report and policy modification suggestions are displayed on the visual panel;

[0139] If the risk level in the risk level report is less than the preset risk level threshold, the risk level report is displayed on a visualization panel.

[0140] The specific implementation method is shown in the method embodiment, which will not be repeated here.

[0141] In one embodiment, the data extraction module 100 is further configured to:

[0142] Build an image recognition model based on deep learning network;

[0143] If the type of the insurance policy to be reviewed is an image, obtaining an insurance policy image corresponding to the insurance policy to be reviewed, inputting the insurance policy image into the image recognition model, and obtaining the text content in the image based on the recognition result of the image recognition model;

[0144] Keywords are extracted from the text content to generate structured insurance policy data.

[0145] The specific implementation method is shown in the method embodiment, which will not be repeated here.

[0146] The present invention provides an insurance policy review device that completely gets rid of reliance on manual experience by automatically extracting the structured data of insurance policies. Different from single-dimensional positive verification, it conducts in-depth logical deduction from multiple angles and combines temporal logic verification. It can accurately identify hidden risks that are difficult to cover in traditional systems, thereby improving review efficiency.

[0147] Another embodiment of the present invention provides a computer device, which may be a server, and its internal structure diagram may be as shown in FIG. Figure 4 As shown. The computer device includes a processor, memory, network interface and database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external client via a network connection. When the computer program is executed by the processor, it implements the functions or steps on the service side of a policy review method.

[0148] In one embodiment, a computer device is provided. The computer device may be a client, and its internal structure diagram may be as follows: Figure 5As shown. The computer device includes a processor, memory, a network interface, a display screen, and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external server via a network connection. When executed by the processor, the computer program implements the functions or steps on the client side of a policy review method.

[0149] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed:

[0150] Obtaining insurance policies to be reviewed, extracting data from the insurance policies to be reviewed, and generating structured insurance policy data;

[0151] Matching the structured policy data with the audit policies in the audit model library to obtain matching audit policies;

[0152] Auditing the structured policy data based on the audit strategy to generate an audit result;

[0153] Based on the preset risk level rules and the audit results, a corresponding risk level report is generated and displayed.

[0154] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0155] Obtaining insurance policies to be reviewed, extracting data from the insurance policies to be reviewed, and generating structured insurance policy data;

[0156] Matching the structured policy data with the audit policies in the audit model library to obtain matching audit policies;

[0157] Auditing the structured policy data based on the audit strategy to generate an audit result;

[0158] Based on the preset risk level rules and the audit results, a corresponding risk level report is generated and displayed.

[0159] It should be noted that the above functions or steps that can be implemented by the computer-readable storage medium or computer device can be found in the relevant descriptions of the server side and the client side in the aforementioned method embodiment. To avoid repetition, they will not be described one by one here.

[0160] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchl ink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0161] The embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the objectives of the present embodiments as needed.

[0162] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, or of course, can also be implemented by hardware. Based on this understanding, the essence of the above technical solution or the part that contributes to the relevant technology can be embodied in the form of a software product. This computer software product can be present in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or certain parts of the embodiment.

[0163] It should be noted that if software tools or components other than those of our company appear in the embodiments of this application, they are only used for illustration and do not represent actual use.

[0164] Conditional language such as "can," "may," or "might," among others, unless specifically stated otherwise or otherwise understood within the context as used, is generally intended to convey that particular embodiments can include, while other embodiments do not, particular features, elements, and / or operations. Thus, such conditional language is also generally intended to imply that features, elements, and / or operations are anyway required for one or more embodiments or that one or more embodiments must include logic for determining, with or without input or prompting, whether such features, elements, and / or operations are included or to be performed in any particular embodiment.

[0165] What has been described herein in this specification and the accompanying drawings includes examples that can provide policy audit methods and apparatus. Of course, it is not possible to describe every conceivable combination of elements and / or methods for the purpose of describing the various features of the present disclosure, but it will be appreciated that many additional combinations and permutations of the disclosed features are possible. It will therefore be apparent that various modifications can be made to the present disclosure without departing from the scope or spirit of the present disclosure. In addition, or in the alternative, other embodiments of the present disclosure may be apparent from consideration of the present specification and the accompanying drawings and from practice of the disclosure as presented herein. It is intended that the examples set forth in this specification and the accompanying drawings be considered in all respects to be illustrative and not restrictive. Although specific terms are employed herein, they are used in a general and descriptive sense and not for purposes of limitation.

Claims

1. A policy review method, characterized by , the method comprises: Obtaining insurance policies to be reviewed, extracting data from the insurance policies to be reviewed, and generating structured insurance policy data; Matching the structured policy data with the audit policies in the audit model library to obtain matching audit policies; Auditing the structured policy data based on the audit strategy to generate an audit result; Based on the preset risk level rules and the audit results, a corresponding risk level report is generated and displayed.

2. The insurance policy review method according to claim 1, characterized in that: The obtaining of the insurance policy to be reviewed, extracting data from the insurance policy to be reviewed, and generating structured insurance policy data includes: When a policy review instruction is detected, obtain the policy to be reviewed; Determine whether the type of the insurance policy to be reviewed is text or image; If the type of the policy to be reviewed is text, perform keyword extraction on the policy to be reviewed to generate structured policy data; If the type of the insurance policy to be reviewed is a picture, image recognition is performed on the insurance policy to be reviewed to obtain text content in the image, and keywords are extracted from the text content to generate structured insurance policy data.

3. The insurance policy review method according to claim 1, characterized in that: The step of matching the structured policy data with the audit policies in the audit model library to obtain the matching audit policies includes: Pre-build an audit model library, the audit strategy of which includes product insurance rules, regulatory rules, product terms and risk insurance amount rules; Obtaining policy parameters from the structured policy data, and assembling the policy parameters to generate assembled parameters; Matching is performed based on the assembly parameters with the audit policies in the audit model library, and obtaining corresponding audit policies according to the matching results.

4. The insurance policy review method according to claim 1, characterized in that: The step of auditing the structured policy data based on the audit strategy and generating an audit result includes: Obtaining reverse verification parameters, timing logic verification parameters, compliance verification parameters, and insurance data verification parameters in the audit policy; Verifying the structured policy data based on the reverse verification parameter to generate a first verification result; Verifying the structured policy data based on the temporal logic verification parameter to generate a second verification result; Verifying the structured policy data based on the compliance verification parameters to generate a third verification result; Verifying the structured policy data based on the insurance data verification parameter to generate a fourth verification result; The first verification result, the second verification result, the third verification result and the fourth verification result are combined to generate an audit result.

5. The insurance policy review method according to claim 1, characterized in that: The risk level report is generated based on the preset risk level rules and the audit results, and the risk level report is displayed, including: Set the risk level and corresponding audit parameters in advance to generate risk level rules; Matching the audit results with the risk level rules; A corresponding risk level report is generated according to the matching result, and the risk level report is displayed on a visualization panel.

6. The insurance policy review method according to claim 5, characterized in that: Generating a corresponding risk level report according to the matching result and displaying the risk level report on a visualization panel includes: Generate a corresponding risk level report based on the matching result, and determine whether the risk level in the risk level report is greater than or equal to a preset risk level threshold; If the risk level in the risk level report is greater than or equal to the preset risk level threshold, an early warning operation is triggered in real time, the early warning information is sent to the manager terminal, and the risk level report and policy modification suggestions are displayed on the visual panel; If the risk level in the risk level report is less than the preset risk level threshold, the risk level report is displayed on a visualization panel.

7. The insurance policy review method according to claim 2, characterized in that: If the type of the insurance policy to be reviewed is an image, image recognition is performed on the insurance policy to obtain text content in the image, and keywords are extracted from the text content to generate structured insurance policy data, including: Build an image recognition model based on deep learning network; If the type of the insurance policy to be reviewed is an image, obtaining an insurance policy image corresponding to the insurance policy to be reviewed, inputting the insurance policy image into the image recognition model, and obtaining the text content in the image based on the recognition result of the image recognition model; Keywords are extracted from the text content to generate structured insurance policy data.

8. An insurance policy review device, characterized in that: The device comprises: A data extraction module is used to obtain the insurance policies to be reviewed, extract data from the insurance policies to be reviewed, and generate structured insurance policy data; A data matching module, configured to match the structured policy data with the audit policies in the audit model library to obtain matching audit policies; A data review module, configured to review the structured policy data based on the review strategy and generate a review result; The risk report generation module is used to generate a corresponding risk level report based on the preset risk level rules and the audit results, and display the risk level report.

9. A computer device, characterized in that: The computer device includes at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the insurance policy review method described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which, when executed by one or more processors, enable the one or more processors to execute the steps of the insurance policy review method described in any one of claims 1-7.

Images