Insurance data intelligent auditing method, device and equipment and storage medium

Abstract

The invention relates to an intelligent auditing method for insurance data, and the method comprises the steps: obtaining client verification data which comprises an identity card picture and a target insurance voucher picture, and is used for auditing insurance handling data; and identifying the customer verification data by using a pre-trained identification model, and generating a verification data identification result containing the target insurance voucher information. And according to a predefined auditing standard, extracting auditing point information from the verification data identification result. And performing comparative analysis on the auditing point information and corresponding information in the insurance handling data to judge whether the information in the insurance handling data is consistent with the information in the customer verification data. And if the comparison analysis result shows that the information is consistent, confirming that the insurance handling data passes the auditing. Through automatic processing and intelligent identification, the auditing efficiency is improved, manual intervention is reduced, and the error rate is reduced, so that the processing speed of the insurance service sheet is accelerated, and the customer experience and the service processing accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence technology and financial technology, and in particular to an insurance information intelligent review method, device, equipment and storage medium. Background Art

[0002] In the traditional auto insurance application process, customers need to provide a series of relevant information to complete the insurance application. This information includes but is not limited to vehicle information, driver's license information and personal identification. Customers usually need to hand over this information to the insurance company's salesperson, who will enter the data provided by the customer into the billing system to generate an insurance business form.

[0003] The salesperson manually fills out the insurance business form based on the information provided by the customer. This process includes entering various customer information and insurance details into the business form. Errors may occur during the manual filling process, such as missing information, incorrect data entry, or inconsistent formatting.

[0004] Insurance companies usually have underwriters review completed insurance business forms. Underwriters check whether the information in the business form is complete and accurate. Manual review usually includes checking customer information and insurance terms item by item to ensure the accuracy and compliance of the business form.

[0005] The manual review process takes a long time because underwriters need to check each data point of the insurance business form item by item. This item-by-item review method makes the entire review process slow. Due to the long manual review time, the generation and review of insurance business forms are also affected, which prolongs the waiting time for customers.

[0006] A slow review process may cause customers to wait too long after applying for insurance, affecting the customer experience. When customers apply for car insurance, long waiting times and uncertain review processes may lead to unsatisfactory user experience.

[0007] In the modern fast-paced society, customers expect to complete the insurance application and review process quickly, but the existing manual review method fails to meet this demand. Summary of the invention

[0008] The main purpose of the present invention is to provide an intelligent insurance data review method, device, equipment and storage medium, aiming to solve the technical problems in existing insurance business that the manual input and review process is time-consuming and easily affected by human errors and negligence, resulting in slow review speed of insurance business orders and insufficient data accuracy.

[0009] To achieve the above object, the present invention provides an insurance information intelligent review method, comprising:

[0010] Obtaining customer verification information, including a picture of the customer's ID card and a picture of the target insurance certificate, for reviewing insurance application information;

[0011] Identify the customer verification information based on the pre-trained recognition model to generate a verification information recognition result, wherein the verification information recognition result includes target insurance certificate information;

[0012] Extracting audit point information from the verification data identification result according to predefined audit standards;

[0013] Compare and analyze the audit point information with the corresponding information in the insurance processing materials to determine whether the information in the insurance processing materials is consistent with the information in the customer verification materials;

[0014] If the comparison and analysis result shows that the information is consistent, it is confirmed that the insurance processing information has passed the review.

[0015] In one embodiment, after determining whether the information in the insurance application materials is consistent with the information in the customer verification materials, the method further includes:

[0016] If the comparison and analysis result shows that the information is inconsistent, it is confirmed that the insurance application materials have not passed the review;

[0017] Identify and determine the fields and corresponding positions in the insurance processing materials that are inconsistent with the audit point information;

[0018] Generating audit prompt information based on the inconsistent fields and corresponding positions, the audit prompt information including data items that need to be resubmitted or supplemented;

[0019] The audit reminder information is sent to the salesperson, prompting him to resubmit or supplement the audit point information that has not passed the review in the insurance processing materials.

[0020] In one embodiment, obtaining customer verification information includes:

[0021] Generate an H5 document upload short link, which includes the URL of the upload page and parameters for identifying the customer's identity information and document type;

[0022] Send the upload short link to the customer's mobile terminal via SMS;

[0023] When the upload short link is triggered, an upload page is provided, and corresponding data upload options are displayed according to the data type;

[0024] Receiving the ID card image and the target insurance certificate image uploaded by the data upload option as the customer verification data;

[0025] Check whether the customer verification information complies with the format requirements;

[0026] The customer verification data that meets the format requirements is stored in the database, and the customer verification data is associated with the customer identity information.

[0027] In one embodiment, before identifying the customer verification information based on the pre-trained recognition model, the method further includes:

[0028] Obtaining sample data, and dividing the sample data into a training set and a test set;

[0029] Inputting the training set into a deep learning model for training, adjusting the recognition parameters of the deep learning model root by learning the data features in the training set, and generating a preliminary recognition model;

[0030] Inputting the test set into the preliminary recognition model to obtain a test result;

[0031] Analyzing the recognition accuracy of the preliminary recognition model based on the test results, and adjusting the recognition parameters of the preliminary recognition model according to the recognition accuracy;

[0032] The preliminary recognition model is trained multiple times using the test set until the recognition accuracy reaches an accuracy threshold, the training is stopped, and a final recognition model is generated.

[0033] In one embodiment, according to a predefined audit standard, extracting audit point information from the verification document recognition result includes:

[0034] Based on the target insurance business needs, define the review standards for the insurance application materials;

[0035] Perform a correspondence analysis based on the fields in the audit standard and the fields in the verification document identification result to determine the audit point information that needs to be extracted;

[0036] The audit point information is extracted and formatted based on the format standard of the insurance processing materials.

[0037] In one embodiment, comparing and analyzing the audit point information with the corresponding information in the insurance application materials to determine whether the information in the insurance application materials is consistent with the information in the customer verification materials includes:

[0038] Extracting data fields matching the type of the audit point information from the insurance handling documents as information to be audited;

[0039] Compare each data field of the information to be audited with the corresponding field in the audit point information one by one;

[0040] If each field in the information to be reviewed is completely consistent with the corresponding field in the audit point information, the comparison and analysis result is confirmed to be information consistency; if there is a data field in the information to be reviewed that is inconsistent with the audit point information, the comparison and analysis result is confirmed to be information inconsistency;

[0041] When the comparison analysis is completed, a comparison analysis result report is generated to record the comparison result of each field.

[0042] In one embodiment, after confirming that the insurance application materials have passed the review, the method further includes:

[0043] Determine the business type of the insurance handling materials, and determine a business form template based on the business type;

[0044] Fill the information of the insurance handling materials into the business form template to generate an insurance business form;

[0045] Verify whether the information in the insurance business form is consistent with the corresponding audit point information;

[0046] The verified and consistent insurance business documents shall be stored in the insurance archive.

[0047] Furthermore, to achieve the above-mentioned purpose, the present invention also provides an insurance document intelligent review device, which includes a memory, a processor, and an insurance document intelligent review program stored in the memory and executable on the processor, and the insurance document intelligent review program, when executed by the processor, implements the steps of the insurance document intelligent review method as described above.

[0048] Furthermore, to achieve the above-mentioned purpose, the present invention also provides a computer storage medium, on which an insurance information intelligent review program is stored, and when the insurance information intelligent review program is executed by a processor, the steps of the insurance information intelligent review method described above are implemented.

[0049] Beneficial effects: The present invention relates to an intelligent review method for insurance documents, which obtains customer verification information, including an ID card image and a target insurance certificate image, for reviewing insurance processing documents. The customer verification information is identified using a pre-trained recognition model to generate a verification document recognition result containing the target insurance certificate information. According to predefined review standards, the review point information is extracted from the verification document recognition result. The review point information is compared and analyzed with the corresponding information in the insurance processing documents to determine whether the information in the insurance processing documents is consistent with the information in the customer verification documents. If the comparison and analysis results show that the information is consistent, it is confirmed that the insurance processing documents have passed the review. Through automated processing and intelligent recognition, the present invention improves review efficiency, reduces manual intervention, and reduces error rates, thereby speeding up the processing of insurance business orders and improving customer experience and the accuracy of business processing. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] The present invention will be further described below with reference to the accompanying drawings and embodiments, in which:

[0051] Figure 1 A schematic diagram of a flow chart of an embodiment of an intelligent insurance data review method of the present invention;

[0052] Figure 2 A schematic diagram of functional modules of a preferred embodiment of the insurance information intelligent review device of the present invention;

[0053] Figure 3 It is a structural diagram of the equipment hardware operating environment involved in the embodiment of the insurance information intelligent review equipment of the present invention. DETAILED DESCRIPTION

[0054] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0055] In the traditional auto insurance application process, customers need to provide a series of relevant information to complete the insurance application. This information includes but is not limited to vehicle information, driver's license information and personal identification. Customers usually need to hand over this information to the insurance company's salesperson, who will enter the data provided by the customer into the billing system to generate an insurance business form.

[0056] The salesperson manually fills out the insurance business form based on the information provided by the customer. This process includes entering various customer information and insurance details into the business form. Errors may occur during the manual filling process, such as missing information, incorrect data entry, or inconsistent formatting.

[0057] Insurance companies usually have underwriters review completed insurance business forms. Underwriters check whether the information in the business form is complete and accurate. Manual review usually includes checking customer information and insurance terms item by item to ensure the accuracy and compliance of the business form.

[0058] The manual review process takes a long time because underwriters need to check each data point of the insurance business form item by item. This item-by-item review method makes the entire review process slow. Due to the long manual review time, the generation and review of insurance business forms are also affected, which prolongs the waiting time for customers.

[0059] A slow review process may cause customers to wait too long after applying for insurance, affecting the customer experience. When customers apply for car insurance, long waiting times and uncertain review processes may lead to unsatisfactory user experience.

[0060] In the modern fast-paced society, customers expect to complete the insurance application and review process quickly, but the existing manual review method fails to meet this demand.

[0061] See also Figure 1 , Figure 1 This is a flow chart of an embodiment of the intelligent insurance data audit method provided by the present invention. It should be noted that although the logical order is shown in the flow chart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0062] like Figure 1 As shown, the insurance information intelligent review method proposed by the present invention includes the following steps:

[0063] S10, obtaining customer verification information, the customer verification information including the customer's ID card image and the target insurance certificate image, which is used to review the insurance application information;

[0064] In this embodiment, verification information for review is obtained from the customer, wherein the verification information includes a picture of the customer's ID card and a picture of the target insurance certificate. These materials are used to confirm the authenticity and completeness of the information provided by the customer.

[0065] The target insurance certificate is a verification document directly related to the subject of insurance, which is used to confirm the legality and detailed information of the subject of insurance. Specifically, it includes formal certification documents used in insurance business, which record the specific information of the insurance contract and serve as important credentials in the application, claim settlement or management process. The main function of the target insurance certificate is to provide a legal and valid proof of the subject of insurance and ensure the authenticity and integrity of the insurance business. For example, the target insurance certificate can be a driver's license, an official document used to prove an individual's driving qualifications. It records the basic information of the driver and his driving authority, including the driver's license number, the name of the holder, date of birth, address, issuing agency, driving category, etc. It can also be a vehicle owner registration certificate, a document used to confirm vehicle ownership. It is usually issued by the vehicle registration management department, and the content includes the owner's name, vehicle registration number, vehicle brand and model, vehicle identification number (VIN), registration date, etc. It can also be a health insurance card, a document or card used to prove the coverage of personal health insurance, including the policy number, cardholder name, insurance company name, insurance coverage, validity period, etc.

[0066] When applying for insurance business, customers upload verification information through the insurance company's online platform or mobile application. These platforms should support multiple image formats, such as JPEG and PNG. The system provides a user-friendly upload interface, including clear instructions and upload buttons, so that customers can upload pictures of their ID cards and insurance certificates. The system specifies the format of uploaded pictures (such as JPEG or PNG) to ensure that the pictures can be processed and recognized correctly. The system sets a file size limit (for example, no more than 5MB) to ensure the quality of uploaded pictures and the processing efficiency of the system. After receiving the pictures uploaded by the customer, the system conducts a preliminary check of the pictures, including file format, resolution, size, etc., to confirm that the pictures meet the system requirements. Pictures that meet the requirements are stored in the system's database and associated with other information of the customer (such as customer ID) to facilitate subsequent review and processing. The uploaded verification information will be encrypted and stored to ensure that customer privacy is protected and prevent unauthorized access. The system complies with the requirements of relevant laws and regulations on data protection and ensures the legality of processing and storing customer data.

[0067] In one specific implementation, on an online insurance application platform, the customer fills out an application form and uploads photos of his ID card and insurance certificate. The system checks the file format and size, then stores the qualified files in a cloud storage service and associates them with the customer's account.

[0068] By optimizing the process of obtaining customer verification information, the efficiency and accuracy of insurance information review has been improved. The system automatically processes and verifies uploaded ID cards and insurance certificate images, reducing the need for manual operations and the risk of human error.

[0069] S20, identifying the customer verification information based on the pre-trained recognition model, and generating a verification information recognition result, wherein the verification information recognition result includes target insurance certificate information;

[0070] In this embodiment, the trained recognition model is used to recognize the verification information uploaded by the customer. The verification information usually includes an ID card image and a target insurance certificate image. The recognition model extracts key information from the image by learning a large amount of data.

[0071] Use deep learning algorithms (such as convolutional neural networks (CNN), Transformer, etc.) for training to accurately identify and extract information from various verification materials. Use a training set containing labeled data (including ID card and insurance certificate images under different conditions) to train the model. The training data includes various samples to enhance the model's adaptability to different image quality and formats. After training, verify the performance of the model through a test set to ensure its recognition accuracy in different situations.

[0072] The ID card and target insurance certificate images uploaded by the customer through the system will be fed into the recognition model for processing. Before recognition, the system pre-processes the uploaded images, such as adjusting the image resolution, removing noise, and correcting tilt, to ensure that the image quality is suitable for model recognition.

[0073] After the model recognizes the uploaded verification data, the generated results include the target insurance certificate information extracted from it. These recognition results are the basis for subsequent data processing and review. Extract personal information from the ID card image, such as name, ID card number, date of birth, etc. For example, extract key information from the insurance certificate image, such as insurance policy number, insurance company name, insurance amount, insurance validity period, etc. Organize the raw data output by the recognition model to ensure that the information is formatted in a standardized data format to facilitate subsequent data processing and comparison. Store the extracted verification data information in the database to ensure data integrity and accessibility to support the subsequent review process.

[0074] Intelligent recognition of customer verification information based on pre-trained recognition models has greatly improved the accuracy of information extraction and processing efficiency. The automated recognition process reduces reliance on manual operations, reduces error rates, and speeds up information processing. It not only improves the overall efficiency of insurance information review, but also enhances the accuracy of data processing, optimizes the processing flow of insurance business, and improves customer service experience.

[0075] S30, extracting audit point information from the verification document recognition result according to a predefined audit standard;

[0076] In this embodiment, according to the pre-set audit criteria, key audit point information is extracted from the model recognition results. Audit point information is the data point used to verify and confirm the accuracy of the insurance business form, including key information that needs to be checked in the ID card and the target insurance certificate.

[0077] Set the audit criteria in the system and specify the data fields that need to be checked. For example, the audit criteria for an ID card may include name, ID number, and date of birth, while the audit criteria for an insurance certificate may include insurance number, insurance company name, and insurance amount, etc. Define the mapping relationship between the fields in the audit criteria and the fields in the recognition results to ensure that each field can be accurately extracted from the recognition results. Parse and organize the raw data obtained from the recognition model (such as ID card and insurance certificate information). The recognition results may include multiple data fields, and the system needs to filter and extract them according to the audit criteria. Extract key field information that meets the audit criteria from the recognition results. For example, ID card: extract name, ID number, date of birth, etc. Insurance certificate: extract insurance number, insurance company name, insurance amount, etc.

[0078] Format the extracted audit point information and organize it into a standardized data format for subsequent comparison and verification. Store the formatted audit point information in the database to ensure the integrity and consistency of the data for subsequent audit steps.

[0079] By extracting audit point information from the recognition results according to predefined audit criteria, the accuracy and consistency of information extraction are improved. Clear audit criteria help the system accurately filter out key data from the recognition results, ensuring that the data in the insurance business form meets the audit requirements. It reduces manual intervention and errors, optimizes the data processing process, speeds up the audit of insurance business forms, and thus improves overall business processing efficiency and customer experience.

[0080] S40, comparing and analyzing the audit point information with the corresponding information in the insurance processing materials to determine whether the information in the insurance processing materials is consistent with the information in the customer verification materials;

[0081] In this embodiment, the audit point information extracted from the customer verification information is compared with the relevant data in the insurance processing information to confirm whether the information between the two is consistent. The comparison and analysis process is a key step to ensure the accuracy of the insurance business form. It can verify whether the information provided by the customer meets the requirements of the insurance processing information. Through systematic comparison and analysis, information inconsistencies can be identified and corrected, thereby ensuring the integrity and reliability of the insurance processing process.

[0082] Extract key information from the customer's verification documents (such as ID cards and insurance certificate images). This information usually includes ID card number, policyholder name, insurance policy number, etc. Use image recognition models to convert text data extracted from images uploaded by customers into a structured data format. For example, convert the name and ID card number extracted from the ID card into text fields.

[0083] Relevant data extracted from insurance application materials (such as insurance application forms, electronic insurance policies, etc.). These data include policy number, policyholder name, insurance amount, etc. Automatically extract field data from the insurance processing system and perform structured processing. For example, extract the policy number and insurance amount from the electronic insurance policy.

[0084] Compare the extracted audit point information with the corresponding fields in the insurance application materials one by one. Ensure that each field (such as insurance policy number, policyholder name) is completely consistent between the two. Apply character matching technology, regular expression analysis or data cleaning algorithms to ensure accurate comparison of information. Deal with possible format differences, inconsistent character case, extra spaces and other issues.

[0085] Verify that the format of the data field is consistent. For example, the insurance policy number should conform to a specific format, and the ID number should be the correct character length and format. Check whether the extracted data is logically consistent. For example, whether the insured's name matches in the ID card and insurance policy, and whether the insurance amount complies with the policy terms.

[0086] The system records the comparison results of each field, including detailed information on consistency and inconsistency. For example, the content, location, expected value and actual value of the inconsistent field are recorded. A detailed report on each inconsistent field is generated, explaining the cause and location of the inconsistency, providing a basis for subsequent processing.

[0087] If the comparison results show inconsistent information, the system will mark the review status as "failed". The system will generate a notification or prompt message to inform the salesperson or auditor of the information items that need to be re-reviewed or corrected.

[0088] Record the comparison results of each field in detail, indicating consistent and inconsistent data. List the abnormalities found during the comparison process and provide solutions or next steps. Provide it to auditors to help them further analyze and deal with data inconsistencies. As part of the system record, it is convenient for future audits and data backtracking.

[0089] By comparing and analyzing the audit point information with the corresponding information in the insurance processing materials, the accuracy and efficiency of data auditing are improved. The system's automated comparison process reduces manual intervention, reduces the risk of operational errors, and speeds up information processing. It not only improves the accuracy of insurance processing materials, but also ensures the consistency of customer verification materials and insurance processing materials, thereby optimizing the audit process of insurance business, improving customer experience, and improving the efficiency of overall business processing.

[0090] S50: If the comparison and analysis result shows that the information is consistent, it is confirmed that the insurance processing information has passed the review.

[0091] In this embodiment, the final result is obtained from the comparison and analysis step to determine whether the insurance processing information has passed the review. The comparison and analysis result is the output of the consistency check between the customer verification information and the insurance processing information. In the comparison and analysis step, the system stores the comparison results, including the comparison status of each field (consistent or inconsistent). The overall consistency information is extracted from the stored comparison results to prepare for subsequent audit status confirmation. The comparison and analysis results are summarized into a processable format, integrating the comparison status of all fields to generate the final audit decision basis.

[0092] Compare and analyze the results for consistency and confirm whether all extracted information is completely consistent. This step determines whether the insurance processing documents are marked as approved. The system checks the results of each comparison field to confirm whether all fields are completely consistent in the insurance processing documents and customer verification documents. If the comparison results of all fields are consistent, the information is judged to be consistent as a whole; otherwise, the inconsistent information is recorded and marked as failed to pass the review. Apply predefined rules and algorithms, such as character matching algorithms or data verification technology, to ensure that the information is completely consistent. Based on the comparison results of all fields, an overall consistency judgment is made to decide whether to confirm that the information has passed the review

[0093] Based on the consistency of the information, the review status of the insurance application materials is updated. If the information is consistent, the information is confirmed to have passed the review and the system status is updated. Based on the consistency judgment result, the system updates the review status of the insurance application materials to "passed". The updated status is recorded in the system database to ensure subsequent tracking and processing. The system generates confirmation information of the review passing, marks the insurance business form as reviewed and passed, and updates the processing progress of the insurance application. The status update information and related data are archived for subsequent audit and query.

[0094] Generate and store audit pass records. Records contain detailed information about the audit decision for subsequent use and audit. Generate detailed audit pass record reports that record audit status updates, a summary of the comparison and analysis results, and any relevant comments. Organize the generated records into a standardized format for easy storage and subsequent access. Store audit pass records in the insurance company's database to ensure data integrity and traceability. Back up and archive audit records to ensure long-term storage and future reference.

[0095] When the comparison and analysis results show that the information is consistent, the insurance application materials are confirmed to have passed the review, which improves the automation level and processing efficiency of information review. The automated comparison process reduces manual intervention, reduces the risk of operational errors, and speeds up the review process. Accurate information consistency judgment and status updates ensure the integrity and accuracy of the insurance application process, thereby optimizing the processing efficiency of insurance business, enhancing customer experience, and improving the reliability and efficiency of the overall business process.

[0096] The present invention relates to an intelligent review method for insurance documents, which obtains customer verification documents, including an ID card image and a target insurance certificate image, for reviewing insurance processing documents. The customer verification documents are recognized using a pre-trained recognition model to generate a verification document recognition result containing target insurance certificate information. According to predefined review standards, review point information is extracted from the verification document recognition result. The review point information is compared and analyzed with the corresponding information in the insurance processing documents to determine whether the information in the insurance processing documents is consistent with the information in the customer verification documents. If the comparison and analysis results show that the information is consistent, it is confirmed that the insurance processing documents have passed the review. Through automated processing and intelligent recognition, the present invention improves review efficiency, reduces manual intervention, and reduces error rates, thereby speeding up the processing speed of insurance business orders and improving customer experience and the accuracy of business processing.

[0097] In one embodiment, in the above S40, after determining whether the information in the insurance application materials is consistent with the information in the customer verification materials, the following further comprises:

[0098] S401, if the comparison and analysis result shows that the information is inconsistent, then confirming that the insurance application materials have not passed the review;

[0099] S402, identifying and determining the fields and corresponding positions in the insurance handling materials that are inconsistent with the audit point information;

[0100] S403, generating review prompt information based on the inconsistent fields and corresponding positions, wherein the review prompt information includes data items that need to be resubmitted or supplemented;

[0101] S404, sending the audit prompt information to the salesperson, prompting the salesperson to resubmit or supplement the audit point information in the insurance processing materials that have not passed the audit.

[0102] In this embodiment, during the comparison and analysis process, if the information in the insurance application materials is found to be inconsistent with the information in the customer verification materials, the system will confirm that the insurance application materials have not passed the review. This process ensures that inconsistent information is properly handled and measures are taken so that the customer can correct or supplement the information.

[0103] The system checks the results of the comparison and analysis. If any field is found to be inconsistent, the insurance application document is marked as "review failed". The system updates the review status to "failed" and records the specific information of the inconsistency. The system identifies which fields are inconsistent from the comparison and analysis results, such as insurance policy number, policyholder name, etc. The specific location of the inconsistent fields in the insurance application document is marked to help accurately locate the problem. The content of the inconsistent fields, the difference between the expected value and the actual value, and the location of the field in the document are recorded in detail.

[0104] Integrate all inconsistent fields and locations into a detailed record for generating audit prompts. Generate detailed audit prompts based on the identified inconsistent fields and locations. This information includes the data items that need to be resubmitted or supplemented. Organize the prompts into a standard format to ensure that they are clear and easy to understand. Integrate the inconsistent fields and locations into the prompts to provide clear instructions. The generated prompts will be used to notify the salesperson or customer so that they can take appropriate actions. Send the audit prompts to the salesperson through appropriate channels (such as email, SMS or system notification). The notification includes detailed audit prompts, prompting the salesperson to resubmit or supplement the audit point information that failed the audit. The system tracks the salesperson's feedback on the prompts, provides necessary support and guidance, and ensures that the problem is resolved.

[0105] This embodiment compares and analyzes the audit point information with the corresponding information in the insurance application materials, so that the system can automatically detect and handle information inconsistencies, reduce manual intervention, and improve data accuracy and processing efficiency. It optimizes the insurance business review process, speeds up information processing, ensures the integrity and reliability of the data, and ultimately improves customer experience and business efficiency.

[0106] In one embodiment, in the above S10, obtaining customer verification information includes:

[0107] S101, generating an H5 document upload short link, wherein the upload short link includes a URL of an upload page and parameters for identifying customer identity information and document type;

[0108] S102, sending the upload short link to the customer's mobile terminal via SMS;

[0109] S103, when the upload short link is triggered, providing an upload page and displaying corresponding data upload options according to the data type;

[0110] S104, receiving the ID card image and the target insurance certificate image uploaded by the data upload option as the customer verification data;

[0111] S105, checking whether the customer verification information meets the format requirements;

[0112] S106, storing the customer verification information that meets the format requirements in a database, and associating the customer verification information with the customer identity information.

[0113] In this embodiment, an H5 short link is created that includes the upload page URL and parameters that identify the customer's identity information and the type of information. A URL for an upload page is generated, which is used to receive the verification information uploaded by the customer. Parameters used to identify the customer's identity and the type of information required to be uploaded are embedded in the short link to ensure that the link can uniquely identify the customer and his or her upload requirements. A short link generation tool is used to shorten the complete URL and parameters to a short link that is easy to send.

[0114] Send the generated H5 short link to the customer's mobile terminal via SMS service. Send the generated short link to the customer's mobile phone number via the SMS sending service platform. Make sure the SMS contains clear instructions to inform the customer of the type of information required to click the link to upload.

[0115] When the upload short link is triggered, for example, after the customer clicks the short link, the system provides an upload page and displays the corresponding upload options based on the document type. The short link guides the customer to the upload page, which displays the corresponding upload options based on the type of document the customer needs to upload (such as ID card, insurance certificate). The page displays relevant upload fields and instructions to facilitate customers to upload accurate information.

[0116] Receive the ID card image and the target insurance certificate image submitted by the customer through the upload page. The customer submits the image files of the ID card and the target insurance certificate through the upload page. The system receives the uploaded images and performs preliminary verification, such as file format check and size limit.

[0117] Verify whether the information uploaded by the customer meets the predetermined format requirements. Check whether the uploaded image file meets the system's required format (such as JPEG, PNG), and whether it meets the size and resolution requirements. For information that does not meet the requirements, the system will provide an error message and require the customer to re-upload the file in the required format.

[0118] Store customer verification information that meets format requirements in the database and associate it with customer identity information. Store format-verified ID card and insurance certificate image files in the system's database. Associate the stored information with customer identity information (such as customer ID, mobile phone number) to ensure data integrity and traceability.

[0119] This embodiment optimizes the submission process of customer verification materials by automatically generating and managing H5 material upload short links. The automated upload page provides clear upload instructions, reducing errors in customer operations. Systematic file format checks ensure data quality, and automatic storage and information association improve the efficiency and accuracy of data management. It improves the convenience of the customer verification process, reduces the need for manual operations, and improves the efficiency of overall business processing and customer experience.

[0120] In one embodiment, in the above S20, before identifying the customer verification information based on the pre-trained recognition model, the process further includes:

[0121] S201, obtaining sample data, and dividing the sample data into a training set and a test set;

[0122] S202, inputting the training set into a deep learning model for training, adjusting the recognition parameters of the deep learning model root by learning the data features in the training set, and generating a preliminary recognition model;

[0123] S203, inputting the test set into the preliminary recognition model to obtain a test result;

[0124] S204, analyzing the recognition accuracy of the preliminary recognition model based on the test result, and adjusting the recognition parameters of the preliminary recognition model according to the recognition accuracy;

[0125] S205, using the test set to train the preliminary recognition model multiple times until the recognition accuracy reaches an accuracy threshold, stopping the training and generating a final recognition model.

[0126] In this embodiment, sample data for training and testing, including image data of customer verification materials (such as ID card pictures, insurance certificate pictures, etc.), are collected and divided into training sets and test sets. Representative sample data, including image data under different conditions, are obtained from various sources (such as historical data, public data sets, or manual collection). The collected sample data is divided into training sets and test sets according to a certain ratio (such as 70% for training and 30% for testing). Ensure that the training set and test set contain different samples to improve the generalization ability of the model.

[0127] The divided training set data is input into the deep learning model for training to learn the features in the image data and adjust the recognition parameters of the model. The training set data is input into the model using a deep learning framework (such as TensorFlow, PyTorch), and the model parameters (such as weights and biases) are adjusted through multiple iterations (epochs) to optimize the recognition ability of the model. The deep learning model extracts key features (such as text areas, characters, etc.) from the image through a convolutional neural network (CNN) or other deep learning architecture, and gradually learns how to recognize and classify data.

[0128] Input the test set data into the trained preliminary recognition model to obtain the recognition results of the model for evaluating the performance of the model. Provide the test set data to the trained preliminary recognition model, and the model will recognize the test data and generate prediction results (such as recognized text content, information fields, etc.). Perform a preliminary analysis of the test results to evaluate the performance of the model on unseen samples, including recognition accuracy, recall rate and other performance indicators.

[0129] By analyzing the accuracy of the test results, the performance of the preliminary recognition model is evaluated, and the model parameters are adjusted according to the recognition accuracy to optimize the recognition effect. The recognition accuracy of the model on the test set is calculated, including the correct recognition ratio, the wrong recognition ratio, etc. According to the analysis results, the hyperparameters of the model (such as learning rate, regularization parameters, etc.) are adjusted to optimize the model to improve the accuracy.

[0130] The preliminary recognition model is trained and adjusted multiple times until the recognition accuracy of the model reaches the preset accuracy threshold, and finally an optimized recognition model is generated. The training process is repeated, and the improvement of the model is verified using the test set. The parameters of the model are updated after each training until the accuracy of the model on the test set reaches the preset threshold. When the recognition accuracy of the model meets the requirements, the training is stopped and the final recognition model is generated for practical application.

[0131] This embodiment can ensure the high accuracy and reliability of the recognition model when processing actual data by recognizing customer verification information based on a pre-trained recognition model. Through multiple training and adjustments, the recognition performance of the model is optimized, the misrecognition rate is reduced, and the accuracy of information extraction is improved. The recognition model finally generated can effectively process the verification information uploaded by the customer, ensure the efficiency and accuracy of the insurance processing process, and significantly improve the speed and quality of business processing.

[0132] In one embodiment, the above S30 includes:

[0133] S301, based on the target insurance business requirements, defining the review standards for the insurance handling materials;

[0134] S302, performing a correspondence analysis based on the fields in the audit standard and the fields in the verification document recognition result to determine the audit point information to be extracted;

[0135] S303, extracting the audit point information, and formatting the audit point information based on the format standard of the insurance processing documents.

[0136] In this embodiment, in order to ensure the accuracy and completeness of the insurance application materials, it is first necessary to define the audit standards according to the needs of the target insurance business. These standards will be used to extract key information from the verification data identification results. Establish clear audit standards that cover all types of information that need to be verified in the insurance application materials. For example, for car insurance, the audit standards may include the owner's name, insurance policy number, insurance amount, etc. Define the corresponding fields and data types for each insurance type or business process, such as the name and ID number on the ID card, the insurance policy number and insurance amount on the insurance certificate, etc.

[0137] Analyze the fields in the audit standards and the fields identified from the verification materials to determine which information needs to be extracted. Establish a field mapping relationship based on the field definition in the audit standards and the field information in the verification material identification results. For example, map the "PolicyNumber" field in the insurance certificate identification results to the "Insurance Policy Number" field in the audit standards. Through comparison and analysis, the system identifies the fields that meet the audit standards and determines the information that needs to be extracted.

[0138] Extract key information fields that meet the audit criteria from the verification document recognition results. This step organizes the relevant information in the original recognition results into audit point information. The system extracts field data that matches the audit criteria from the recognition results. For example, extract the name and ID card number from the ID card recognition results, and extract the insurance policy number and insurance amount from the insurance certificate recognition results. Organize the extracted fields into a standard format for subsequent use and analysis.

[0139] Format the extracted audit point information to make it conform to the format standards of insurance processing materials. Develop formatting standards based on the requirements of insurance processing materials. For example, the ID card number should be 15 or 18 characters, and the insurance policy number should conform to a specific character length and format. Adjust the extracted audit point information to a standardized format, perform necessary conversions and corrections, and ensure that the information is consistent with the format of insurance processing materials. Verify whether the formatted data meets the standards and make necessary corrections.

[0140] This embodiment ensures the accuracy and consistency of information extraction by extracting audit point information from the verification document recognition results according to predefined audit standards. Clear audit standards and automated extraction processes reduce the need for manual intervention, reduce error rates, and speed up data processing. Formatting ensures the consistency of information with insurance application materials, optimizes the insurance audit process, and improves overall business efficiency and customer experience.

[0141] In one embodiment, the above S40 includes:

[0142] S405, extracting data fields matching the type of the audit point information from the insurance handling documents as information to be audited;

[0143] S406, comparing each data field of the information to be reviewed with the corresponding field in the review point information one by one;

[0144] S407, if each field in the information to be reviewed is completely consistent with the corresponding field in the audit point information, confirm that the comparison and analysis result is that the information is consistent; if there is a data field in the information to be reviewed that is inconsistent with the audit point information, confirm that the comparison and analysis result is that the information is inconsistent;

[0145] S408, when the comparison analysis is completed, a comparison analysis result report is generated to record the comparison result of each field.

[0146] In this embodiment, data fields that match the information type of the audit point are extracted from the insurance processing documents. This step ensures that the data obtained from the insurance processing documents are the parts that need to be compared. Identify and extract data fields in the insurance processing documents that match the information type of the audit point. For example, extract the policy number and the name of the insured from the insurance application form and match them with the corresponding fields extracted from the customer verification data. Arrange the extracted fields into the format of the information to be audited, ready for further comparison.

[0147] Compare each field of the information to be reviewed extracted from the insurance application materials with the corresponding field in the review point information to confirm the consistency of the information. Compare each field in the information to be reviewed with the corresponding field in the review point information one by one. For example, compare the insurance policy number with the policy number in the review point information one by one. Use character matching, data verification, or regular expressions and other technologies to compare each field in detail to handle data format differences and possible errors.

[0148] Based on the comparison results, determine whether the information in the insurance application materials is consistent with the information in the customer verification materials. If each field in the information to be reviewed is completely consistent with the corresponding field in the audit point information, confirm that the information is consistent. The system updates the comparison result to "Information is consistent" and the insurance application materials pass the review. If any field to be reviewed is inconsistent with the corresponding field in the audit point information, record the data inconsistency. The system updates the comparison result to "Information is inconsistent" and the insurance application materials fail the review.

[0149] After the comparison analysis is completed, a report containing the comparison results is generated, recording the comparison of each field. The comparison analysis results are recorded in detail, including the comparison results of each field, the consistency status, and any inconsistent fields found. The generated report is stored in the system database for subsequent review and processing.

[0150] This embodiment improves the automation level and accuracy of data review by comparing and analyzing the audit point information with the corresponding information in the insurance processing materials. The systematic comparison process reduces manual operations, reduces the error rate, and speeds up processing. Accurate consistency judgment and detailed comparison result reports ensure the integrity and reliability of information, optimize the insurance review process, and improve overall business efficiency and customer experience.

[0151] In one embodiment, in the above S50, after confirming that the insurance application materials have passed the review, the following step is further included:

[0152] S501, determining the business type of the insurance handling document, and determining a business form template based on the business type;

[0153] S502, filling the information of the insurance handling materials into the business form template to generate an insurance business form;

[0154] S503, verifying whether the information in the insurance business form is consistent with the corresponding audit point information;

[0155] S504, storing the verified insurance business documents into the insurance archive.

[0156] In this embodiment, after confirming that the insurance application materials have passed the review, it is first necessary to determine the type of business application (such as automobile insurance, health insurance, etc.) in order to select a suitable business form template. The system automatically identifies the business type based on key information in the insurance application materials (such as insurance item name, product number, etc.). According to the predefined classification criteria, the insurance application materials are classified into corresponding business type categories.

[0157] According to the identified business type, select the appropriate template from the preset business form template library to generate the insurance business form. The system selects the corresponding business form template according to the business type (such as automobile insurance template, health insurance template). Apply the selected template to provide the format and layout basis for generating the insurance business form.

[0158] Fill the information in the insurance application materials into the selected business form template to generate the final insurance business form. Automatically fill the data in the insurance application materials (such as the name of the insured, the insurance amount, the policy number, etc.) into the corresponding fields of the business form template. Generate a complete insurance business form, ensuring that all relevant information is accurately displayed in the template.

[0159] Verify the generated insurance business form to check whether the information is consistent with the previous audit point information. The system compares the information in the insurance business form with the audit point information to ensure that all information is consistent. Use the comparison tool to check each field to ensure that the data in the insurance business form is consistent with the audit point information.

[0160] Store verified and consistent insurance business documents in the insurance archive for subsequent query and management. Store verified and consistent insurance business documents in the insurance archive to ensure data security and long-term preservation. Create archive records for each insurance business document to facilitate subsequent access, auditing and data backtracking.

[0161] This embodiment can effectively improve the accuracy and processing efficiency of insurance business form generation by automatically selecting a suitable business form template and filling in information to generate an insurance business form after the insurance processing materials are reviewed and approved. The automated template selection and information filling process reduces manual intervention, reduces the error rate, and speeds up the generation of business forms. The verification process ensures that the generated business form information is consistent with the audit point information, further improving the accuracy of the data. Finally, the verified and consistent insurance business forms are stored in the insurance archive, ensuring the integrity and traceability of the data, thereby optimizing the entire insurance processing process and improving the efficiency of business processing and customer satisfaction.

[0162] The present invention also provides an insurance information intelligent review device, referring to Figure 2 , Figure 2 This is a functional module diagram of a preferred embodiment of the insurance information intelligent review device of the present invention. The insurance information intelligent review device comprises:

[0163] A data acquisition module is used to acquire customer verification information, including a picture of the customer's ID card and a picture of the target insurance certificate, which is used to review insurance application information;

[0164] A data identification module, used to identify the customer verification information based on a pre-trained recognition model and generate a verification information identification result, wherein the verification information identification result includes target insurance certificate information;

[0165] An audit point extraction module, used to extract audit point information from the verification data identification result according to a predefined audit standard;

[0166] A consistency review module, used to compare and analyze the review point information with the corresponding information in the insurance processing materials, and determine whether the information in the insurance processing materials is consistent with the information in the customer verification materials;

[0167] The audit result confirmation module is used to confirm that the insurance processing materials have passed the audit if the comparison and analysis results show that the information is consistent.

[0168] The specific implementation of the insurance document intelligent review device of the present invention is basically the same as the embodiments of the above-mentioned insurance document intelligent review method, and will not be described in detail here.

[0169] The present invention also provides an insurance information intelligent review device, such as Figure 3 As shown, the insurance information intelligent review device may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0170] Those skilled in the art will understand that Figure 3 The hardware structure of the insurance document intelligent review device shown in the figure does not constitute a limitation on the insurance document intelligent review device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0171] like Figure 3As shown, the memory 1005 as a storage medium may include an operating system, a network communication module, a user interface module, and an insurance data intelligent review program. The operating system is a program for managing and controlling the insurance data intelligent review device and software resources, and supports the operation of the network communication module, the user interface module, the insurance data intelligent review program, and other programs or software; the network communication module is used to manage and control the network interface 1004; and the user interface module is used to manage and control the user interface 1003.

[0172] exist Figure 3 In the hardware structure of the insurance information intelligent review device shown, the network interface 1004 is mainly used to connect to the background server and communicate data with the background server; the user interface 1003 is mainly used to connect to the client and communicate data with the client; the processor 1001 can call the insurance information intelligent review program stored in the memory 1005 and perform the same operation as the insurance information intelligent review method.

[0173] The specific implementation of the insurance document intelligent review device of the present invention is basically the same as the embodiments of the above-mentioned insurance document intelligent review method, and will not be repeated here.

[0174] In addition, an embodiment of the present invention further provides a computer storage medium on which an insurance information intelligent review program is stored. When the insurance information intelligent review program is executed by a processor, the steps of the insurance information intelligent review method described above are implemented.

[0175] The specific implementation of the computer storage medium of the present invention is basically the same as the above-mentioned embodiments of the intelligent review method for insurance information, and will not be described in detail here.

[0176] It should be noted that if software tools or components other than those of our company appear in the embodiments of the present application, they are only used for illustrative purposes and do not represent actual use.

Claims

1. An intelligent insurance data review method, characterized in that: The following steps are involved: Obtaining customer verification information, including a picture of the customer's ID card and a picture of the target insurance certificate, for reviewing insurance application information; Identify the customer verification information based on the pre-trained recognition model to generate a verification information recognition result, wherein the verification information recognition result includes target insurance certificate information; Extracting audit point information from the verification data identification result according to predefined audit standards; Compare and analyze the audit point information with the corresponding information in the insurance processing materials to determine whether the information in the insurance processing materials is consistent with the information in the customer verification materials; If the comparison and analysis result shows that the information is consistent, it is confirmed that the insurance processing information has passed the review.

2. The intelligent insurance data review method according to claim 1, characterized in that: After determining whether the information in the insurance application materials is consistent with the information in the customer verification materials, it also includes: If the comparison and analysis result shows that the information is inconsistent, it is confirmed that the insurance application materials have not passed the review; Identify and determine the fields and corresponding positions in the insurance processing materials that are inconsistent with the audit point information; Generating audit prompt information based on the inconsistent fields and corresponding positions, the audit prompt information including data items that need to be resubmitted or supplemented; The audit reminder information is sent to the salesperson, prompting him to resubmit or supplement the audit point information that has not passed the review in the insurance processing materials.

3. The intelligent insurance data review method according to claim 1, characterized in that: Obtain customer verification information, including: Generate an H5 document upload short link, which includes the URL of the upload page and parameters for identifying the customer's identity information and document type; Send the upload short link to the customer's mobile terminal via SMS; When the upload short link is triggered, an upload page is provided, and corresponding data upload options are displayed according to the data type; Receiving the ID card image and the target insurance certificate image uploaded by the data upload option as the customer verification information; Check whether the customer verification information complies with the format requirements; The customer verification data that meets the format requirements is stored in the database, and the customer verification data is associated with the customer identity information.

4. The intelligent insurance data review method according to claim 1, characterized in that: Before identifying the customer verification information based on the pre-trained recognition model, the method further includes: Obtaining sample data, and dividing the sample data into a training set and a test set; Inputting the training set into a deep learning model for training, adjusting the recognition parameters of the deep learning model root by learning the data features in the training set, and generating a preliminary recognition model; Inputting the test set into the preliminary recognition model to obtain a test result; Analyzing the recognition accuracy of the preliminary recognition model based on the test results, and adjusting the recognition parameters of the preliminary recognition model according to the recognition accuracy; The preliminary recognition model is trained multiple times using the test set until the recognition accuracy reaches an accuracy threshold, the training is stopped, and a final recognition model is generated.

5. The intelligent insurance data review method as claimed in claim 1, characterized in that: According to the predefined audit criteria, audit point information is extracted from the verification data identification result, including: Based on the target insurance business needs, define the review standards for the insurance application materials; Perform a correspondence analysis based on the fields in the audit standard and the fields in the verification document identification result to determine the audit point information that needs to be extracted; The audit point information is extracted, and the audit point information is formatted based on the format standard of the insurance processing materials.

6. The intelligent insurance data review method according to claim 1, characterized in that: Comparing and analyzing the audit point information with the corresponding information in the insurance application materials to determine whether the information in the insurance application materials is consistent with the information in the customer verification materials, including: Extracting data fields matching the type of the audit point information from the insurance handling documents as information to be audited; Compare each data field of the information to be audited with the corresponding field in the audit point information one by one; If each field in the information to be reviewed is completely consistent with the corresponding field in the audit point information, the comparison and analysis result is confirmed to be information consistency; if there is a data field in the information to be reviewed that is inconsistent with the audit point information, the comparison and analysis result is confirmed to be information inconsistency; When the comparison analysis is completed, a comparison analysis result report is generated to record the comparison result of each field.

7. The intelligent insurance data audit method according to claim 1, characterized in that: After confirming that the insurance application materials have passed the review, it also includes: Determine the business type of the insurance handling materials, and determine a business form template based on the business type; Fill the information of the insurance handling materials into the business form template to generate an insurance business form; Verify whether the information in the insurance business form is consistent with the corresponding audit point information; The verified and consistent insurance business documents shall be stored in the insurance archive.

8. An intelligent insurance information review device, characterized in that: The insurance information intelligent review device comprises: A data acquisition module is used to acquire customer verification information, including a picture of the customer's ID card and a picture of the target insurance certificate, which is used to review insurance application information; A data identification module, used to identify the customer verification information based on a pre-trained recognition model and generate a verification information identification result, wherein the verification information identification result includes target insurance certificate information; An audit point extraction module, used to extract audit point information from the verification data identification result according to a predefined audit standard; A consistency review module, used to compare and analyze the review point information with the corresponding information in the insurance processing materials, and determine whether the information in the insurance processing materials is consistent with the information in the customer verification materials; The audit result confirmation module is used to confirm that the insurance processing materials have passed the audit if the comparison and analysis results show that the information is consistent.

9. An intelligent insurance data review device, characterized in that: The insurance data intelligent review device includes a memory, a processor, and an insurance data intelligent review program stored in the memory and executable on the processor. When the insurance data intelligent review program is executed by the processor, the steps of the insurance data intelligent review method as described in any one of claims 1 to 7 are implemented.

10. A computer storage medium, characterized in that: The storage medium stores an insurance document intelligent review program, which, when executed by the processor, implements the steps of the insurance document intelligent review method according to any one of claims 1 to 7.

Images