An insurance quotation information processing method, system, electronic device and storage medium

Abstract

The application relates to the field of artificial intelligence and data processing technology, and particularly discloses an insurance quotation information processing method and system, an electronic device and a storage medium. The application analyzes an attachment file through a large language model, generates an initial data object in combination with supplementary text, and performs semantic matching in a professional knowledge base and a clause knowledge base after integrity verification to obtain standard professional information and clause information, finally fuses to form a standardized quotation data object and converts and outputs, solves the technical problems of low efficiency of manual processing of unstructured quotation files, easy errors in data extraction, inaccurate professional risk matching and complex clause responsibility correspondence, and improves the quotation processing efficiency and accuracy.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and data processing technology, and in particular to a method, system, electronic device, and storage medium for processing insurance quote information. Background Technology

[0002] Currently, in the insurance industry, especially in non-motor insurance businesses such as employer's liability insurance, the quotation process is generally handled manually. Underwriters need to manually read various unstructured quotation documents provided by clients, such as electronic application forms and scanned copies of business licenses, extracting key data such as policyholder information, insured information, occupational category, insurance period, and liability limits, and then entering this information into the core business system for quotation calculation. However, this highly manual operation-dependent model has the following technical drawbacks: First, the processing efficiency is low. Manual reading and data entry are time-consuming and difficult to handle a large number of concurrent quotation requests, which restricts business processing capabilities.

[0003] Secondly, due to differences in the understanding of unstructured data among different personnel, errors or omissions are prone to occur during the information extraction process, resulting in inconsistent data standards and affecting the accuracy and consistency of quotations.

[0004] Third, the occupational risk matching process faces significant challenges. There are numerous occupational categories for insurance, and manually searching the internal occupational code database is not only time-consuming but also prone to errors, making it difficult to achieve fast and accurate occupational risk level matching.

[0005] Fourth, the terms and conditions of liability are complex, with a wide variety of liabilities for the main insurance and supplementary insurance. The standard terms and conditions codes and limit rules for manually determining liability present technical barriers, which further increases the difficulty and risk of errors in quotation processing.

[0006] Therefore, there is an urgent need to provide a technical solution to address the above problems. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention provides an insurance quote information processing method, system, electronic device, and storage medium.

[0008] In a first aspect, the present invention provides a method for processing insurance quote information, the technical solution of which is as follows: The system receives an attachment file containing insurance quote information uploaded by the user and supplementary text input by the user. It uses a large language model to parse the attachment file, identifies the attachment type, extracts the initial structured quote information from the attachment content, and combines the supplementary text to generate an initial data object containing policyholder information, insured information, insured information, and insurance plan. The initial data object is subjected to integrity verification of key fields. If the verification fails, guidance information is generated to instruct the user to supplement the missing fields. Occupational description information is extracted from the verified initial data object, and semantic matching is performed in a preset occupational knowledge base to obtain standard occupational information containing occupational codes and risk levels. If the matching result is not unique, an interactive interface is triggered to obtain the user's selection confirmation from multiple candidate occupational information. The insurance liability description information is extracted from the verified initial data object, and semantic matching is performed in the preset main insurance knowledge base and supplementary insurance knowledge base to obtain the clause information containing standard clause codes and calculation rules. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate clause information. The standard occupational information and the terms information are merged with the initial data object, and the corresponding non-standard fields in the initial data object are replaced to form a standardized quotation data object; According to the interface specifications of the core business system, the standardized quotation data object is converted into a quotation request message and output to the core business system.

[0009] The beneficial effects of the insurance quotation information processing method of the present invention are as follows: The method of this invention parses the attachment file using a large language model and combines it with supplementary text to generate an initial data object. After integrity verification, semantic matching is performed in the professional knowledge base and the clause knowledge base to obtain standard professional information and clause information. Finally, the data is merged to form a standardized quotation data object and converted for output. This method solves the technical problems of low efficiency in manual processing of unstructured quotation documents, error-prone data extraction, inaccurate matching of professional risks, and complex correspondence of clause responsibilities, thereby improving the efficiency and accuracy of quotation processing.

[0010] Based on the above solution, the insurance quotation information processing method of the present invention can be further improved as follows.

[0011] In one alternative approach, the steps of parsing the attachment file using a large language model, identifying the attachment type, extracting initial structured quotation information from the attachment content, and combining this with the supplementary text to generate an initial data object containing policyholder information, insured information, insured information, and insurance plan include: Identify the format type of the attachment file, which is an electronic file containing the insurance quote information; The corresponding target parsing strategy is selected according to the format type. The parsing strategy includes an optical character recognition parsing strategy for image format attachments and a text extraction parsing strategy for document format attachments. According to the target parsing strategy, for image-formatted attachments, the optical character recognition model is called to extract the text content in the image, and the large language model is used to perform visual understanding of the non-text elements in the image to obtain additional information related to insurance pricing information; for document-formatted attachments, the document parsing interface is called to extract the text content in the document. The extracted text content, the additional information, and the supplementary text are combined to form a set of text to be parsed that contains the insurance quote information; Using the large language model and based on a preset parsing prompt template, the field values ​​corresponding to the policyholder information, the insured information, the object information, and the insurance plan are identified and extracted from the text set to be parsed; The extracted field values ​​are cleaned and normalized to generate the initial data object containing the policyholder information, the insured information, the object information, and the insurance plan.

[0012] The advantages of adopting the above-mentioned optional methods are as follows: by further distinguishing between image and document formats and calling the corresponding parsing strategies, and combining optical character recognition and the visual understanding capabilities of large language models, the technical problem of unified parsing of multi-format attachments is solved, and the completeness and accuracy of unstructured data extraction are improved.

[0013] In one alternative approach, the initial data object undergoes a key field integrity check. If the check fails, guidance information is generated, instructing the user to complete the missing fields, including: Obtain a list of key fields, which includes at least one field to be verified defined from the policyholder information, the insured information, the object information, and the insurance plan; Iterate through each field to be verified in the list of key fields, check whether there is a field value in the initial data object that corresponds to the field to be verified, and determine whether the field value conforms to the preset field format rules; If all fields to be verified have corresponding field values ​​and the field values ​​all conform to the field format rules, then the integrity verification is deemed to have passed. If at least one field to be validated is missing a corresponding field value or the field value does not conform to the field format rules, then the field to be validated will be marked as a missing field or a field with an incorrect format. The large language model is invoked to generate guidance information for natural language description based on the identifiers of the missing or formatted fields. The guidance information includes supplementary or corrective suggestions for the missing or formatted fields. The guidance information is output to the interactive interface.

[0014] The advantages of adopting the above optional methods are: by further verifying key fields and generating intelligent guidance information, the technical problem of difficulty in quickly locating missing or incorrectly formatted fields in the initial data object is solved, thereby improving the automation level of data integrity verification and the user interaction experience.

[0015] In one alternative approach, occupational description information is extracted from the validated initial data object, semantic matching is performed in a preset occupational knowledge base to obtain standard occupational information containing occupational codes and risk levels, and if the matching results are not unique, an interactive interface is triggered to obtain the user's confirmation of selection from multiple candidate occupational information, including: The occupational description information is extracted from the verified initial data object, wherein the occupational description information is a natural language description of the occupational name or occupational category text; The occupational description information is converted into a first feature vector. The second feature vector corresponding to each occupational record in the occupational knowledge base is traversed. The semantic similarity between the first feature vector and each second feature vector is calculated. The occupational knowledge base contains multiple occupational records. Each occupational record contains at least a standard occupational name, occupational code and risk level. All occupational records are sorted in descending order based on the semantic similarity, and occupational records with semantic similarity higher than a preset threshold are selected as candidate occupational information. If the number of candidate occupation information is one, then the candidate occupation information is determined as the standard occupation information; If there are multiple candidate occupation information entries, the multiple candidate occupation information entries will be presented to the user through the interactive interface, and the user will select and confirm the target occupation information from the multiple candidate occupation information entries, and the target occupation information will be used as the standard occupation information.

[0016] The beneficial effects of adopting the above-mentioned optional methods are as follows: by further vectorizing occupational description information and calculating semantic similarity in the knowledge base, the technical problem of accurately mapping natural language occupational descriptions and standard occupational codes is solved, thereby improving the speed and accuracy of occupational risk level matching.

[0017] In one optional approach, the insured liability description information is extracted from the verified initial data object, and semantic matching is performed in a preset main insurance knowledge base and supplementary insurance knowledge base to obtain clause information containing standard clause codes and calculation rules. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate clause information, including: The insurance liability description information is extracted from the verified initial data object. The insurance liability description information is a liability name or liability clause text described in natural language. The description of the insured liability is converted into a third feature vector, and semantic matching is performed in the main insurance knowledge base and the supplementary insurance knowledge base respectively. Specifically, in the main insurance knowledge base, a first semantic similarity is calculated between the third feature vector and the fourth feature vector corresponding to each main insurance clause record, and in the supplementary insurance knowledge base, a second semantic similarity is calculated between the third feature vector and the fifth feature vector corresponding to each supplementary insurance clause record. The main insurance knowledge base contains multiple main insurance clause records, and each main insurance clause record contains at least a standard main insurance clause code and calculation rules. The supplementary insurance knowledge base contains multiple supplementary insurance clause records, and each supplementary insurance clause record contains at least a standard supplementary insurance clause code and calculation rules. All main insurance clause records with a first semantic similarity higher than a preset threshold are selected from the main insurance knowledge base as the first candidate set, and all supplementary insurance clause records with a second semantic similarity higher than a preset threshold are selected from the supplementary insurance knowledge base as the second candidate set. The first candidate set and the second candidate set are merged to form a candidate clause information set; If the candidate clause information set is empty, a prompt message is generated and returned; If the number of candidate clauses in the candidate clause information set is one, then the one candidate clause information is determined as the standard clause information; If there are multiple candidate clauses in the candidate clause information set, the multiple candidate clauses are presented to the user through the interactive interface, and the user selects and confirms the target clause from the multiple candidate clauses, and the target clause is used as the standard clause.

[0018] The beneficial effects of adopting the above-mentioned optional methods are as follows: by further solving the technical problems of complex multi-source retrieval of liability clauses and standard code correspondence through parallel semantic matching of the description information of insurance liability in the knowledge base of main insurance and supplementary insurance, the comprehensiveness of clause information acquisition and matching efficiency are improved.

[0019] In one alternative approach, the step of merging the standard occupational information, the term information, and the initial data object, and replacing the corresponding non-standard fields in the initial data object to form a standardized quotation data object, includes: Identify a first non-standard field corresponding to the occupational description information and a second non-standard field corresponding to the insurance liability description information from the initial data object; Write the occupational code and risk level from the standard occupational information into the initial data object, replacing the original occupational description information in the first non-standard field; and write the standard clause code and calculation rules from the clause information into the initial data object, replacing the original insurance liability description information in the second non-standard field. Iterate through the remaining fields in the initial data object to check if there are any fields to be supplemented that are related to the standard occupational information or the terms information; If the field to be supplemented exists, then according to the risk level in the standard occupational information or the calculation rule in the clause information, the preset field filling rule is invoked to assign a value to the field to be supplemented. The initial data object, after field replacement and supplementation, is repackaged according to a preset data structure to form the standardized quotation data object.

[0020] The advantages of adopting the above optional method are as follows: by further identifying and replacing non-standard fields and automatically filling in the rules of associated fields, the technical problem of inconsistent field mapping during the fusion of standard information and initial data objects is solved, and the integrity and standardization of the standardized encapsulation of quotation data objects are improved.

[0021] In one alternative approach, the step of converting the standardized quotation data object into a quotation request message according to the interface specification of the core business system and outputting it to the core business system includes: Obtain the interface specifications of the core business system, which include a field mapping table, value conversion rules, and message structure definitions; Based on the field mapping table, the name of each data field in the standardized quotation data object is mapped to the corresponding interface field name in the interface specification; According to the value conversion rules, the value of each mapped data field is converted into a target value that is consistent with the data type and format required by the interface specification; According to the message structure definition, the interface field names and their corresponding target values ​​are assembled into the quotation request message in a preset format; According to the interface call address and call method defined in the interface specification, the interface service provided by the core business system is invoked, and the quotation request message is transmitted to the core business system.

[0022] The advantages of adopting the above optional methods are: by further solving the technical problem of difficulty in adapting standardized data objects to the core business system interface through interface specification mapping, value conversion and message assembly, the compliance and transmission stability of quotation request message generation are improved.

[0023] Secondly, the present invention provides an insurance quotation information processing system, the technical solution of which is as follows: The identification module is used to receive attachment files containing insurance quotation information uploaded by users and supplementary text input by users. It uses a large language model to parse the attachment files, identify the attachment type, extract the initial structured quotation information from the attachment content, and combine the supplementary text to generate an initial data object containing policyholder information, insured information, insured information, and insurance plan. The verification module is used to perform integrity verification on the key fields of the initial data object. If the verification fails, it generates guidance information to instruct the user to supplement the missing fields. The extraction module is used to extract occupational description information from the verified initial data object, perform semantic matching in a preset occupational knowledge base, and obtain standard occupational information containing occupational codes and risk levels. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate occupational information. The matching module is used to extract the description information of the insured liability from the verified initial data object, perform semantic matching in the preset main insurance knowledge base and supplementary insurance knowledge base, and obtain the clause information containing standard clause codes and calculation rules. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate clause information. The fusion module is used to merge the standard occupational information, the terms information and the initial data object, and replace the corresponding non-standard fields in the initial data object to form a standardized quotation data object; The output module is used to convert the standardized quotation data object into a quotation request message according to the interface specifications of the core business system, and output it to the core business system.

[0024] The beneficial effects of the insurance quotation information processing system of the present invention are as follows: The system of this invention parses the attachment file using a large language model and combines it with supplementary text to generate an initial data object. After integrity verification, semantic matching is performed in the professional knowledge base and the clause knowledge base to obtain standard professional information and clause information. Finally, the data is merged to form a standardized quotation data object and converted for output. This solves the technical problems of low efficiency in manual processing of unstructured quotation documents, error-prone data extraction, inaccurate matching of professional risks, and complex correspondence of clause responsibilities, thereby improving the efficiency and accuracy of quotation processing.

[0025] Thirdly, the technical solution of an electronic device according to the present invention is as follows: The invention includes a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the insurance quote information processing method of the present invention.

[0026] Fourthly, the technical solution of a computer-readable storage medium provided by the present invention is as follows: The computer-readable storage medium stores instructions that, when read, cause the computer-readable storage medium to perform the steps of the insurance quote information processing method of the present invention.

[0027] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0028] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart illustrating an embodiment of an insurance quote information processing method according to the present invention. Figure 2 This is the overall flowchart; Figure 3 This is a schematic diagram of the structure of an embodiment of the insurance quotation information processing system of the present invention; Figure 4 This is a schematic diagram of an embodiment of an electronic device according to the present invention. Detailed Implementation

[0029] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0030] Figure 1 This diagram illustrates a flowchart of an embodiment of an insurance quote information processing method provided by the present invention. This method can be executed by an electronic device such as a terminal device or a server. The terminal device can be any fixed or mobile terminal, such as a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, in-vehicle device, or wearable device. The server can be a single server or a server cluster consisting of multiple servers. Any electronic device can implement the insurance quote information processing method by having its processor call computer-readable instructions stored in its memory. Figure 1 As shown, it includes the following steps: S1. Receive the attachment file containing insurance quotation information uploaded by the user and the supplementary text input by the user. Use a large language model to parse the attachment file, identify the attachment type and extract the initial structured quotation information from the attachment content. Combine the supplementary text to generate an initial data object containing policyholder information, insured information, object information and insurance plan.

[0031] The insurance quote information refers to various data related to quote calculation in insurance business, including information on the insured entity, insured object, insurance liability, insurance period, and liability limit. For example, when Technology A insures its employees with employer's liability insurance, the quote information provided includes the company name, employee occupation category, number of insured persons, death liability limit of RMB 1 million, and medical expense liability limit of RMB 100,000. Attached files refer to electronic files uploaded by users containing insurance quote information, in formats including PDF, images, and Word documents. For example, Technology A uploads an electronic insurance application PDF file. Supplementary text refers to additional text information provided by users through the input interface, used to supplement or clarify quote information not included or unclear in the attachments. For example, Technology A manually enters "Total number of employees: 50, including 30 office staff and 20 sales staff" on the quote page.

[0032] Here, "Large Language Model" refers to a natural language processing model trained using deep learning technology, possessing text understanding, generation, extraction, and reasoning capabilities. It's used to parse the content of attachment files and extract structured information; for example, using a large language model to parse text in an electronic insurance application. "Attachment Type" refers to the format category of the attachment file, used to determine the subsequent parsing method; for example, a PDF file is recognized as a document attachment, and a scanned copy of the insurance application is recognized as an image attachment. "Attachment Content" refers to all information contained in the attachment file, including text, tables, seals, and handwritten content; for example, the policyholder's name, insured information, occupational category list, and insurance amount in an electronic insurance application PDF. "Structured Quotation Information" refers to data extracted from unstructured attachment content and organized according to predetermined fields for easy subsequent processing; for example, field values ​​such as "Policyholder: Technology Co., Ltd.", "Occupation: Programmer", and "Liability: Death Liability" extracted from the insurance application.

[0033] The information provided includes: Policyholder information: Identity and contact data related to the policyholder, including the policyholder's name, unified social credit code, address, and contact person; for example, the name of Technology Co., Ltd. A, its unified social credit code is 9123456789ABCDEF, and its address is No. 1, a certain road in a certain district of a certain city. Insured information: Identity and risk data related to the insured, typically the employee's name, ID number, and job title in employer's liability insurance; for example, an employee's name, ID number, and job title are programmer. Insured information: A detailed description of the insured object, mainly the number of employees, work location, and job duties in employer's liability insurance; for example, Technology Co., Ltd. A has 50 employees, and their work location is an office building in a certain district of a certain city. Insurance plan: The combination of insurance types selected by the user and the corresponding liability limits, deductibles, and other underwriting conditions; for example, the main employer's liability insurance includes death liability (limit of 1 million yuan) and medical liability (limit of 100,000 yuan), and the supplementary insurance includes lost wages liability (200 yuan per day).

[0034] The initial data object refers to a structured data set containing extracted fields that is initially generated after parsing by a large language model. The fields may be in non-standard formats. For example, the generated data object contains key-value pairs such as "Company Name: A Technology Co., Ltd.", "Occupation: Programmer", and "Responsibility: Death".

[0035] S2. Perform integrity checks on the key fields of the initial data object. If the check fails, generate guidance information to instruct the user to supplement the missing fields.

[0036] Among them, the integrity verification of key fields refers to checking the predefined required fields in the initial data object to determine whether they exist and are in the correct format; for example, checking whether the "insured person's name" field has a value, and checking whether the value of the "occupation category" field conforms to the standard format.

[0037] The guidance information refers to the prompt text generated when the integrity check fails, which informs the user which fields are missing or incorrect and guides the user to supplement or correct them; for example, the prompt text may display "Occupational description information is incomplete, please supplement the occupational categories of all employees." Missing fields refer to required fields that are not included in the initial data object and are found during the integrity check; for example, if the check finds that the "Number of Insured Persons" field has no value, this field is marked as a missing field.

[0038] S3. Extract occupational description information from the verified initial data object, perform semantic matching in a preset occupational knowledge base to obtain standard occupational information containing occupational codes and risk levels. If the matching result is not unique, trigger the interactive interface to obtain the user's selection confirmation from multiple candidate occupational information.

[0039] Occupational description information refers to the occupational name or category text extracted from the initial data objects and described in natural language; for example, "programmer" or "sales representative". The occupational knowledge base refers to a pre-built database containing standard occupational names, occupational codes, and risk level mappings; for example, the knowledge base records "programmer" as having occupational code C001 and a low risk level, and "sales representative" as having occupational code S002 and a medium risk level.

[0040] Standard occupational information refers to occupational records that include occupational codes and risk levels, determined through semantic matching. For example, the standard occupational information obtained after matching might be "Occupational Code: C001, Risk Level: Low". An occupational code is a unique identifier assigned to each standard occupation in the occupational knowledge base; for example, C001 represents a programmer. A risk level refers to a classification based on the degree of occupational risk, used to calculate insurance premium rates; for example, a programmer's risk level is "low," and a sales representative's risk level is "medium."

[0041] The interactive interface refers to the interface used to display information or receive user input, including the display of guiding information and the selection of candidate lists; for example, a pop-up dialog box displays "Please select from the following candidate occupations: Programmer (C001), Software Engineer (C002)". Candidate occupation information refers to multiple occupation records with a similarity higher than a threshold and a potential match in occupational semantic matching; for example, for the input "Programmer", the candidate occupation information includes "Programmer (C001)" and "Software Engineer (C002)".

[0042] S4. Extract the description information of the insured liability from the verified initial data object, perform semantic matching in the preset main insurance knowledge base and supplementary insurance knowledge base to obtain the clause information containing standard clause codes and calculation rules. If the matching result is not unique, trigger the interactive interface to obtain the user's selection confirmation from multiple candidate clause information.

[0043] The insured liability description information refers to the insurance liability name or clause text extracted from the initial data object and described in natural language; for example, "death liability," "medical expenses," and "lost wages supplementary insurance." The main insurance knowledge base refers to a pre-built database containing main insurance clause records, their standard codes, and calculation rules; for example, the main insurance knowledge base records the standard main insurance clause code EL01 for "Employer's Liability Insurance Main Insurance - Death," with the calculation rule being "number of deaths × limit." The supplementary insurance knowledge base refers to a pre-built database containing supplementary insurance clause records, their standard codes, and calculation rules; for example, the supplementary insurance knowledge base records the standard supplementary insurance clause code ELR01 for "Employer's Liability Insurance Supplementary Insurance - Lost Wages," with the calculation rule being "number of lost work days × daily standard."

[0044] The "term information" refers to the term record, which includes standard term codes and calculation rules, determined after semantic matching. For example, the term information obtained after matching might be "Standard term code: EL01, Calculation rule: Number of insured persons × Limit per person × Premium rate." The standard term code is a unique identifier assigned to each standard term in the knowledge base; for example, EL01 represents the employer's liability death liability main insurance term, and ELR01 represents the lost wages supplementary insurance term. The calculation rule refers to the premium or compensation calculation method corresponding to the term, used for subsequent quotation calculations; for example, the calculation rule for death liability is "Number of insured persons × Limit per person × Premium rate," and the calculation rule for lost wages is "Number of lost work days × 200 yuan per day."

[0045] Among them, candidate clause information refers to multiple clause records with similarity higher than a threshold selected from the knowledge base of main insurance or supplementary insurance in the semantic matching of clauses; for example, for the input "death liability", the candidate clause information includes main insurance clause EL01 and supplementary insurance clause ELR02.

[0046] S5. Merge the standard occupational information, the terms and conditions information, and the initial data object, and replace the corresponding non-standard fields in the initial data object to form a standardized quotation data object.

[0047] Non-standard fields refer to fields in the initial data object that do not use internal standard codes or formatting specifications and need to be replaced by standard information. For example, the "Occupation" field in the initial data object might have the value "Programmer," while the internal standard requires an occupation code; this field is therefore a non-standard field. Standardized quotation data objects refer to data objects where all fields conform to internal specifications after the integration of standard occupation information and clause information. For example, the final object might contain "Policyholder: A Technology Co., Ltd., Occupation Code: C001, Risk Level: Low, Main Insurance Clause Code: EL01, Supplementary Insurance Clause Code: ELR01."

[0048] S6. Based on the interface specifications of the core business system, convert the standardized quotation data object into a quotation request message and output it to the core business system.

[0049] The core business system refers to the insurance company's core business processing system, used to receive quotation requests, calculate premiums, and return quotation results; for example, the insurance company's internal quotation core system. The interface specification refers to the data format, field mappings, transmission protocols, and other requirements defined by the core business system for external interaction; for example, the interface specification requires quotation request messages to be in JSON format, containing fields such as "policyHolder" and "occupationCode". The quotation request message refers to a data packet generated according to the interface specification, containing all standard quotation data, used to send quotation requests to the core business system; for example, the generated JSON message contains {"policyHolder":"A Technology Co., Ltd.", "occupationCode":"C001", ...}.

[0050] The technical solution of this embodiment parses the attachment file using a large language model and combines it with supplementary text to generate an initial data object. After integrity verification, semantic matching is performed in the professional knowledge base and the clause knowledge base to obtain standard professional information and clause information. Finally, the data is merged to form a standardized quotation data object and converted for output. This solves the technical problems of low efficiency in manually processing unstructured quotation documents, error-prone data extraction, inaccurate matching of professional risks, and complex correspondence of clause responsibilities, thereby improving the efficiency and accuracy of quotation processing.

[0051] In one alternative approach, S1 specifically includes: Identify the format type of the attachment file, which is an electronic file containing the insurance quote information.

[0052] The format type refers to the storage format of the attachment file, such as PDF, JPEG, or DOCX; for example, the uploaded electronic insurance application is in PDF format.

[0053] The corresponding target parsing strategy is selected according to the format type. The parsing strategy includes an optical character recognition parsing strategy for image format attachments and a text extraction parsing strategy for document format attachments.

[0054] The target parsing strategy refers to the specific method selected based on the attachment format type for extracting information; for example, a text extraction parsing strategy is selected for PDF files, and an optical character recognition parsing strategy is selected for JPEG images.

[0055] According to the target parsing strategy, for image-formatted attachments, the optical character recognition model is called to extract the text content in the image, and the large language model is used to perform visual understanding of the non-text elements in the image to obtain additional information related to insurance quotes; for document-formatted attachments, the document parsing interface is called to extract the text content in the document.

[0056] Image-format attachments refer to attachments stored in image format, such as scanned copies or photos; for example, a JPEG file uploaded by Technology Co., Ltd. after photographing a paper insurance application. Optical Character Recognition (OCR) models refer to machine learning models used to recognize and extract text content from images; for example, an OCR model extracting text from an insurance application photo. Text content refers to plain text information extracted from attachments through parsing methods; for example, text such as "Insured: Technology Co., Ltd." extracted from a PDF.

[0057] Non-text elements refer to visual elements in the attachments other than text, such as seals, icons, and table lines; for example, the company seal and handwritten signature on the insurance application. Visual understanding refers to using the multimodal capabilities of a large language model to analyze and interpret non-text elements in an image to obtain implicit information; for example, visual understanding can be used to identify that the company name on the seal matches the policyholder's information. Supplementary information refers to supplementary information obtained from non-text elements through visual understanding; for example, the "A Technology Co., Ltd." extracted from the seal can be used to verify the policyholder's information.

[0058] Among these, document format attachments refer to attachments stored in editable document formats, such as PDF, Word, and Excel; for example, an electronic insurance application PDF provided by Technology Co., Ltd. Document parsing interfaces refer to program interfaces used to extract text and structure from document format files; for example, calling a PDF parsing library to extract text blocks from a PDF.

[0059] The extracted text content, the additional information, and the supplementary text are combined to form a set of text to be parsed that contains the insurance quote information.

[0060] Using the large language model and based on the preset parsing prompt template, the field values ​​corresponding to the policyholder information, the insured information, the object information, and the insurance plan are identified and extracted from the text set to be parsed.

[0061] The preset parsing prompt templates refer to pre-designed prompt word templates that guide the large language model to extract specific information; for example, templates such as "Please extract the policyholder's name, insured's information, occupation category, and insurance liability from the following text." Field values ​​refer to the specific data content extracted from the attachments, corresponding to each field; for example, the value of the policyholder's name field is "A Technology Co., Ltd."

[0062] The extracted field values ​​are cleaned and normalized to generate the initial data object containing the policyholder information, the insured information, the object information, and the insurance plan.

[0063] Among the above-mentioned optional methods, by further distinguishing between image and document formats and calling the corresponding parsing strategies, and combining optical character recognition and the visual understanding capabilities of large language models, the technical problem of unified parsing of multi-format attachments is solved, and the completeness and accuracy of unstructured data extraction are improved.

[0064] In one alternative approach, S2 specifically includes: Obtain a list of key fields, which includes at least one field to be verified defined from the policyholder information, the insured information, the object information, and the insurance plan.

[0065] The key field list refers to a predefined set of field names that must be checked during integrity verification; for example, the key field list includes "Insured Person's Name," "Occupation Category," and "Number of Insured Persons." The fields to be verified refer to each specific field in the key field list; for example, the field currently being verified is "Occupation Category."

[0066] Iterate through each field to be verified in the list of key fields, check whether there is a field value in the initial data object that corresponds to the field to be verified, and determine whether the field value conforms to the preset field format rules.

[0067] Among them, field format rules refer to the format requirements for field values, such as date format, number range, and code specifications; for example, the value of the occupation category field should be a standard occupation code with a length not exceeding 10 characters.

[0068] If all fields to be verified have corresponding field values ​​and the field values ​​all conform to the field format rules, then the integrity verification is deemed to have passed.

[0069] If at least one field to be validated is missing a corresponding field value or the field value does not conform to the field format rules, then the field to be validated will be marked as a missing field or a field with an incorrect format.

[0070] Among them, a format error field refers to a field whose value does not conform to the preset format rules; for example, if the occupation category field value is "programmer" instead of standard code, this field is marked as a format error field.

[0071] The large language model is invoked to generate guidance information in natural language description based on the identifiers of the missing or malformed fields. The guidance information includes supplementary or corrective suggestions for the missing or malformed fields.

[0072] The guidance information is output to the interactive interface.

[0073] Among the above optional methods, the technical problem of difficulty in quickly locating missing or formatted fields in the initial data object is solved by further using key field traversal verification and intelligent guidance information generation mechanism, thereby improving the automation level of data integrity verification and user interaction experience.

[0074] In one alternative approach, S3 specifically includes: The occupational description information is extracted from the verified initial data object. The occupational description information is a natural language description of the occupational name or occupational category text.

[0075] Among them, "job title" refers to the occupational designation expressed in natural language; for example, "programmer" or "sales manager". "Job category text" refers to the textual description of the job, which may contain multiple job titles or descriptions; for example, "programmer who works in software development".

[0076] The occupational description information is converted into a first feature vector. The second feature vector corresponding to each occupational record in the occupational knowledge base is traversed. The semantic similarity between the first feature vector and each second feature vector is calculated. The occupational knowledge base contains multiple occupational records. Each occupational record contains at least a standard occupational name, occupational code and risk level.

[0077] The first feature vector refers to the numerical vector converted from occupational description information through an embedding model; for example, the text "programmer" is converted into a 256-dimensional vector. The second feature vector refers to the numerical vector converted from the standard occupational name corresponding to each occupational record in the occupational knowledge base; for example, the standard name "programmer" is converted into a vector. The standard occupational name refers to the occupational name used in a standardized manner within the occupational knowledge base; for example, the standard name for "programmer" in the knowledge base is "computer software engineer".

[0078] All occupational records are sorted in descending order based on the semantic similarity, and occupational records with semantic similarity higher than a preset threshold are selected as candidate occupational information.

[0079] The preset threshold refers to the minimum similarity value used to filter candidate information; only records with a similarity value higher than this value are included in the candidate list. For example, if the threshold is set to 0.8, only records with a similarity greater than 0.8 will become candidate occupational information.

[0080] If the number of candidate occupation information is one, then the candidate occupation information is determined as the standard occupation information.

[0081] If there are multiple candidate occupation information entries, the multiple candidate occupation information entries will be presented to the user through the interactive interface, and the user will select and confirm the target occupation information from the multiple candidate occupation information entries, and the target occupation information will be used as the standard occupation information.

[0082] Among the above-mentioned optional methods, the technical problem of accurately mapping natural language occupational descriptions and standard occupational codes is further solved by vectorizing occupational description information and calculating semantic similarity in the knowledge base, thereby improving the speed and accuracy of occupational risk level matching.

[0083] In one alternative approach, S4 specifically includes: The insurance liability description information is extracted from the verified initial data object. The insurance liability description information is a liability name or liability clause text described in natural language.

[0084] The term "liability name" refers to the insurance liability term expressed in natural language; for example, "death liability" or "medical expense liability". The "liability clause text" refers to a detailed description of the insurance liability; for example, "During the insurance period, if the insured's employee dies due to an accident, the insurer shall pay compensation in accordance with the contract."

[0085] The description of the insured liability is converted into a third feature vector, and semantic matching is performed in the main insurance knowledge base and the supplementary insurance knowledge base respectively. Specifically, in the main insurance knowledge base, a first semantic similarity is calculated between the third feature vector and the fourth feature vector corresponding to each main insurance clause record, and in the supplementary insurance knowledge base, a second semantic similarity is calculated between the third feature vector and the fifth feature vector corresponding to each supplementary insurance clause record. The main insurance knowledge base contains multiple main insurance clause records, and each main insurance clause record contains at least a standard main insurance clause code and calculation rules. The supplementary insurance knowledge base contains multiple supplementary insurance clause records, and each supplementary insurance clause record contains at least a standard supplementary insurance clause code and calculation rules.

[0086] The third feature vector refers to the numerical vector converted from the description of the insured liability; for example, the vector converted from the text "death liability". The fourth feature vector refers to the numerical vector converted from the standard clause name or description corresponding to each main insurance clause record in the main insurance knowledge base; for example, the vector converted from "employer's liability insurance death clause". The first semantic similarity refers to the similarity between the third feature vector and the fourth feature vector of a certain record in the main insurance knowledge base; for example, the similarity with the main insurance clause record is 0.92.

[0087] The fifth feature vector refers to the numerical vector converted from the standard clause name or description corresponding to each supplementary insurance clause record in the supplementary insurance knowledge base; for example, the vector converted from "lost wages supplementary insurance clause". The second semantic similarity refers to the similarity between the third feature vector and the fifth feature vector of a certain record in the supplementary insurance knowledge base; for example, the similarity with the supplementary insurance clause record is 0.88.

[0088] The standard main insurance clause code refers to a unique identifier assigned to a standard main insurance clause in the main insurance knowledge base; for example, EL01 represents the employer's liability insurance death liability main insurance. The standard supplementary insurance clause code refers to a unique identifier assigned to a standard supplementary insurance clause in the supplementary insurance knowledge base; for example, ELR01 represents the employer's liability insurance lost wages supplementary insurance.

[0089] All main insurance clause records with a first semantic similarity higher than a preset threshold are selected from the main insurance knowledge base as the first candidate set, and all supplementary insurance clause records with a second semantic similarity higher than a preset threshold are selected from the supplementary insurance knowledge base as the second candidate set.

[0090] The first candidate set refers to the set of all main insurance clause records selected from the main insurance knowledge base whose first semantic similarity is higher than a preset threshold; for example, the set containing EL01 and EL02. The second candidate set refers to the set of all supplementary insurance clause records selected from the supplementary insurance knowledge base whose second semantic similarity is higher than a preset threshold; for example, the set containing ELR01 and ELR02.

[0091] The first candidate set and the second candidate set are merged to form a candidate clause information set.

[0092] The candidate clause information set refers to the set of all candidate clause records obtained by merging the first candidate set and the second candidate set; for example, the merged set contains four records: EL01, EL02, ELR01, and ELR02.

[0093] If the candidate clause information set is empty, a prompt message is generated and returned.

[0094] If the number of candidate clauses in the candidate clause information set is one, then the one candidate clause information is determined as the standard clause information.

[0095] If there are multiple candidate clauses in the candidate clause information set, the multiple candidate clauses are presented to the user through the interactive interface, and the user selects and confirms the target clause from the multiple candidate clauses, and the target clause is used as the standard clause.

[0096] Among the above-mentioned optional methods, the parallel semantic matching of the insured liability description information in the knowledge base of the main insurance and supplementary insurance further solves the technical problems of multi-source retrieval of liability clauses and the complex correspondence of standard codes, and improves the comprehensiveness of clause information acquisition and matching efficiency.

[0097] In one alternative approach, S5 specifically includes: Identify a first non-standard field corresponding to the occupational description information and a second non-standard field corresponding to the insurance liability description information from the initial data object.

[0098] The first non-standard field refers to the field in the initial data object that stores occupational description information, and its value is in a non-standard format; for example, the value of the "Occupation" field is "Programmer". The second non-standard field refers to the field in the initial data object that stores insurance liability description information, and its value is in a non-standard format; for example, the value of the "Liability" field is "Death".

[0099] The occupational code and risk level from the standard occupational information are written into the initial data object, replacing the original occupational description information in the first non-standard field. The standard clause code and calculation rules from the clause information are also written into the initial data object, replacing the original insurance liability description information in the second non-standard field.

[0100] The original occupational description information refers to the value of the first non-standard field in the initial data object, i.e., the initially extracted occupational text; for example, "programmer". The original insurance liability description information refers to the value of the second non-standard field in the initial data object, i.e., the initially extracted liability text; for example, "death".

[0101] Iterate through the remaining fields in the initial data object to check if there are any fields to be added that are related to the standard occupational information or the terms information.

[0102] Among them, fields to be supplemented refer to fields that are discovered during the integration process, are related to standard occupational information or clause information, but have not yet been assigned values; for example, a "risk coefficient" field may need to be supplemented based on the occupational risk level, and if this field is empty, it becomes a field to be supplemented.

[0103] If the field to be supplemented exists, then according to the risk level in the standard occupational information or the calculation rules in the clause information, the preset field filling rules are invoked to assign values ​​to the field to be supplemented.

[0104] The field filling rules refer to the preset logic used to automatically calculate or assign values ​​to the fields to be filled based on standard information; for example, if the risk level is "low", the risk coefficient is assigned a value of 1.0; if the risk level is "medium", the risk coefficient is assigned a value of 1.2.

[0105] The initial data object, after field replacement and supplementation, is repackaged according to a preset data structure to form the standardized quotation data object.

[0106] Among them, data structure re-encapsulation refers to reorganizing the data objects after replacement and supplementation according to a predefined data structure to form standardized quotation data objects; for example, encapsulating data into JSON objects containing fixed field order and nested structure.

[0107] Among the above optional methods, the technical problem of inconsistent field mapping during the fusion of standard information and initial data objects is further solved by using non-standard field identification and replacement and automatic filling rule calls for associated fields, thereby improving the integrity and standardization of the standardized encapsulation of quotation data objects.

[0108] In one alternative approach, S6 specifically includes: Obtain the interface specifications of the core business system, which include field mapping tables, value conversion rules, and message structure definitions.

[0109] The field mapping table refers to the table defined in the interface specification that maps field names in standardized quotation data objects to field names in the core business system interface. For example, in the mapping table, "Insured Person's Name" corresponds to the interface field "policyHolder," and "Occupation Code" corresponds to "occupationCode." Value conversion rules refer to the rules defined in the interface specification that convert field values ​​to the data types or formats required by the interface. For example, the date format is converted from "January 1, 2026" to "20260101," and the amount field is converted to a string with two decimal places. The message structure definition refers to the overall structure of the quotation request message specified in the interface specification, including hierarchical relationships, field order, and separators. For example, the message structure is a JSON object containing a root node "request," followed by "header" and "body."

[0110] Based on the field mapping table, the name of each data field in the standardized quotation data object is mapped to the corresponding interface field name in the interface specification.

[0111] The interface field name refers to the field name defined in the core business system interface specification and used in the message; for example, "policyHolder" and "occupationCode".

[0112] According to the value conversion rules, the value of each mapped data field is converted into a target value that is consistent with the data type and format required by the interface specification.

[0113] The target value refers to the field value that conforms to the interface specification after being processed by the value conversion rules; for example, the target value of the insured's name is "A Technology Co., Ltd." and the target value of the occupation code is "C001".

[0114] According to the message structure definition, the interface field names and their corresponding target values ​​are assembled into a quotation request message in a preset format.

[0115] The default format refers to the final message format required by the interface specification, such as JSON, XML, or text with a specific delimiter; for example, the default format is JSON.

[0116] According to the interface call address and call method defined in the interface specification, the interface service provided by the core business system is invoked, and the quotation request message is transmitted to the core business system.

[0117] The interface call address refers to the network address provided by the core business system for receiving quotation request messages. The calling method refers to the HTTP method used when sending a request to the core business system, such as POST or GET; for example, sending a quotation request using the POST method. The interface service refers to the functional unit provided by the core business system for receiving and processing quotation requests; for example, the quotation core system receives messages and returns quotation results through an HTTP service.

[0118] Among the above-mentioned optional methods, further interface specification mapping, value conversion and message assembly are used to solve the technical problem of difficulty in adapting standardized data objects to core business system interfaces, thereby improving the compliance and transmission stability of quotation request message generation.

[0119] like Figure 2 As shown, the specific process of the insurance quote information processing method in this embodiment includes: The interactive input module receives attachments containing insurance quote information uploaded by the user, as well as supplementary text entered by the user. Attachments are electronic files containing insurance quote information, such as electronic insurance applications in PDF format or scanned copies of insurance applications in JPEG format. Supplementary text is additional text information provided by the user through the input interface, used to supplement or clarify quote information not included in the attachments or that is unclear, such as the number of employees or occupational categories manually entered by the user.

[0120] The AI ​​vision and semantic parsing module receives attachment files and supplementary text from the interactive input module. Equipped with a large language model, this module parses the attachment files, identifies the attachment type, and extracts initial structured pricing information from the attachment content. The large language model is a natural language processing model trained using deep learning technology, possessing text understanding, generation, extraction, and reasoning capabilities. The attachment type refers to the format category of the attachment file, used to determine the subsequent parsing method. The attachment content refers to all information contained in the attachment file, including text, tables, seals, handwritten content, etc.

[0121] The information processing node processes the extracted initial structured quotation information, extracts JSON format data and splits variables, and combines it with supplementary text to generate an initial data object containing policyholder information, insured information, insured property information, and insurance plan information. Policyholder information refers to the policyholder's identity and contact data, including policyholder name, unified social credit code, address, and contact person. Insured information refers to the insured's identity and risk data, typically employee name, ID number, and job title in employer's insurance. Insured property information refers to the specific description of the insured property, mainly the number of employees, work location, and job duties in employer's insurance. The insurance plan refers to the combination of insurance types selected by the user and the corresponding liability limits, deductibles, and other underwriting conditions. The initial data object is a structured data set containing extracted fields, initially generated after parsing by a large language model; the fields may be in non-standard formats.

[0122] The information integrity verification module performs integrity checks on key fields of the initial data object. This check examines predefined required fields in the initial data object to determine their existence and correct format. If the verification passes, subsequent processing is triggered. If the verification fails, the system proceeds to the incomplete information response module, generating guidance information and returning to the interactive interface, instructing the user to supplement the missing fields. The guidance information is the prompt text generated when the integrity verification fails, informing the user which fields are missing or incorrect and guiding them to supplement or correct them. Missing fields refer to required fields found during the integrity verification that do not contain values ​​in the initial data object.

[0123] For initial data objects that pass verification, the occupational category intelligent matching module extracts occupational description information from them. This occupational description information is the occupational name or category text extracted from the initial data object, described in natural language. The occupational category intelligent matching module then enters a batch occupational category search loop, performing a knowledge base search within a pre-built occupational knowledge base. The occupational knowledge base is a pre-constructed database containing standard occupational names, occupational codes, and risk level mappings. The occupational code is a unique identifier assigned to each standard occupation in the occupational knowledge base. The risk level is a classification based on the degree of occupational risk, used to calculate insurance premium rates.

[0124] During the knowledge base retrieval process, occupational description information is converted into a first feature vector. Then, the second feature vector corresponding to each occupational record in the knowledge base is traversed, and the semantic similarity between the first feature vector and each second feature vector is calculated. Semantic similarity refers to the closeness between two feature vectors, typically calculated using cosine similarity. All occupational records are sorted in descending order based on semantic similarity, and records with semantic similarity higher than a preset threshold are selected as candidate occupational information.

[0125] The AI-powered occupational category decision module evaluates the matching results. If the matching result is unique, the obtained standard occupational information is determined as the final standard occupational information. Standard occupational information is an occupational record containing an occupational code and risk level, determined after semantic matching. If the matching result is not unique, an interactive interface is triggered, requiring the user to select and confirm from multiple candidate occupational information. The interactive interface is used to display information or receive user input, including guiding information display and candidate list selection. Candidate occupational information refers to multiple occupational records with a similarity higher than a threshold and a potential match in occupational semantic matching. After the user confirms their selection, the target occupational information selected by the user is used as the standard occupational information.

[0126] After the occupational category intelligent matching is completed, the clause liability intelligent matching module extracts the insured liability description information from the initial data object. The insured liability description information is the name of the insurance liability or the clause text, described in natural language, extracted from the initial data object. The clause liability intelligent matching module then enters a batch clause retrieval loop, performing knowledge base searches in the pre-set main insurance knowledge base and supplementary insurance knowledge base. The main insurance knowledge base is a pre-built database containing main insurance clause records, their standard codes, and calculation rules. The supplementary insurance knowledge base is a pre-built database containing supplementary insurance clause records, their standard codes, and calculation rules.

[0127] During the knowledge base retrieval process, the description of the insured liability is converted into a third feature vector. The first semantic similarity between this third feature vector and the fourth feature vector corresponding to each main insurance clause record is calculated in the main insurance knowledge base. The second semantic similarity between this third feature vector and the fifth feature vector corresponding to each supplementary insurance clause record is calculated in the supplementary insurance knowledge base. All main insurance clause records with a first semantic similarity higher than a preset threshold are selected from the main insurance knowledge base as the first candidate set, and all supplementary insurance clause records with a second semantic similarity higher than a preset threshold are selected from the supplementary insurance knowledge base as the second candidate set. The first and second candidate sets are then merged to form a candidate clause information set.

[0128] The AI-powered decision-making module evaluates the matching results. If the candidate clause information set is empty, a prompt message is generated and returned. If the candidate clause information set contains only one candidate clause, that candidate clause is designated as the standard clause. The standard clause code is a unique identifier assigned to each standard clause in the knowledge base. The calculation rule is the premium or claim calculation method corresponding to the clause, used for subsequent quotation calculation. Clause information is a clause record determined after semantic matching, containing the standard clause code and calculation rule. If the candidate clause information set contains multiple candidate clauses, these multiple clauses are presented to the user through an interactive interface, requiring the user to select and confirm from among them. Candidate clause information refers to multiple clause records with a similarity higher than a threshold selected from the main or supplementary insurance knowledge base during clause semantic matching. After the user confirms their selection, the target clause information selected by the user is designated as the standard clause information.

[0129] The data mapping and data backfilling module integrates standard occupational information and standard clause information with the initial data object, replacing corresponding non-standard fields in the initial data object to form a standardized quotation data object. Non-standard fields refer to fields in the initial data object that do not use internal standard codes or formatting specifications and need to be replaced by standard information. A standardized quotation data object is a data object where all fields conform to internal specifications after the integration of standard occupational information and clause information.

[0130] The output module receives standardized quotation data objects, converts them into quotation request messages according to the interface specifications of the core business system, and outputs them to the core business system. The core business system is the insurance company's core business processing system, used to receive quotation requests, calculate premiums, and return quotation results. The interface specifications are the data formats, field mappings, transmission protocols, and other requirements defined by the core business system for external interaction with it. The quotation request message is a data packet containing all standard quotation data, generated according to the interface specifications, used to send a quotation request to the core business system.

[0131] After receiving the quotation request message, the core business system calculates the premium and returns the quotation result to the user.

[0132] In another embodiment of the insurance quote information processing method of the present invention, the following steps are specifically included: S10. Receive the attachment file containing insurance quote information uploaded by the user and the supplementary text entered by the user.

[0133] S20. Use a large language model to parse the attachment file, identify the attachment type and extract the initial structured quotation information from the attachment content, and combine it with supplementary text to generate an initial data object containing policyholder information, insured information, object information and insurance plan.

[0134] S30. Perform integrity checks on key fields of the initial data object. If the check fails, generate guidance information to instruct the user to supplement the missing fields.

[0135] S40. Extract occupational description information from the verified initial data object, perform semantic matching in the preset occupational knowledge base to obtain standard occupational information containing occupational code and risk level. If the matching result is not unique, trigger the interactive interface to obtain the user's selection confirmation from multiple candidate occupational information.

[0136] S50. Extract the description information of the insured liability from the initial data object that has passed the verification, perform semantic matching in the preset main insurance knowledge base and supplementary insurance knowledge base, and obtain the clause information containing standard clause codes and calculation rules. If the matching result is not unique, trigger the interactive interface to obtain the user's selection confirmation from multiple candidate clause information.

[0137] S60. Integrate standard occupational information and clause information with the initial data object, replace the corresponding non-standard fields in the initial data object, and form a standardized quotation data object.

[0138] S70. Call the preset association rule base to perform occupation-term association verification on the standardized quotation data object. The association rule base contains rules for allowed combinations, prohibited combinations, and mandatory additions between occupation codes and term codes. If the occupation code in the standard occupation information and the term code in the term information violate the prohibited combination rule, a conflict prompt message is generated and the interactive interface is triggered to allow the user to reselect occupation information or term information. If the occupation code and term code meet the mandatory addition rule but the term information is missing the corresponding mandatory addition clause, the corresponding additional insurance clause information is automatically added to the standardized quotation data object according to the mandatory addition rule.

[0139] The rules for permitted combinations refer to the correspondence between occupation codes and clause codes defined in the association rule base, which are allowed to appear simultaneously in the same insurance plan. For example, occupation code C001 corresponds to the programmer occupation, and the main insurance clause code EL01 (death liability clause) is allowed to be combined with it. The rules for prohibited combinations refer to the correspondence between occupation codes and clause codes defined in the association rule base, which are not allowed to appear simultaneously in the same insurance plan. For example, occupation code S002 corresponds to the sales representative occupation, and the supplementary insurance clause code ELR03 (high-risk sports liability clause) is prohibited from being combined with it. The rules for mandatory supplementary ...

[0140] S80. Based on the interface specifications of the core business system, convert the standardized quotation data object that has passed the occupation-term association verification into a quotation request message and output it to the core business system.

[0141] This embodiment generates an initial data object by parsing the attachment file step by step. After integrity verification, it performs semantic matching in the professional knowledge base and the clause knowledge base to obtain standard information. The data is then merged to form a standardized quotation data object and a professional-clause correlation verification is performed. This solves the technical problems of low efficiency in manually processing unstructured quotation documents, error-prone data extraction, inaccurate matching of professional risks and clause responsibilities, and difficulty in automatically verifying the compliance of professional and clause combinations. This achieves a comprehensive improvement in quotation processing efficiency, data accuracy, and business compliance.

[0142] Figure 3 A schematic diagram of an embodiment of an insurance quote information processing system 200 provided by the present invention is shown. Figure 3 As shown, the insurance quotation information processing system 200 includes: The identification module 201 is used to receive an attachment file containing insurance quotation information uploaded by the user and supplementary text input by the user, parse the attachment file using a large language model, identify the attachment type and extract the initial structured quotation information from the attachment content, and generate an initial data object containing policyholder information, insured information, object information and insurance plan by combining the supplementary text. The verification module 202 is used to perform integrity verification on the key fields of the initial data object. If the verification fails, guidance information is generated to instruct the user to supplement the missing fields. The extraction module 203 is used to extract occupational description information from the verified initial data object, perform semantic matching in a preset occupational knowledge base, and obtain standard occupational information containing occupational codes and risk levels. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate occupational information. The matching module 204 is used to extract the description information of the insured liability from the verified initial data object, perform semantic matching in the preset main insurance knowledge base and supplementary insurance knowledge base, and obtain the clause information containing standard clause codes and calculation rules. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate clause information. The fusion module 205 is used to merge the standard occupational information, the clause information and the initial data object, and replace the corresponding non-standard fields in the initial data object to form a standardized quotation data object; The output module 206 is used to convert the standardized quotation data object into a quotation request message according to the interface specification of the core business system, and output it to the core business system.

[0143] In one alternative embodiment, the identification module 201 is specifically used for: Identify the format type of the attachment file, which is an electronic file containing the insurance quote information; The corresponding target parsing strategy is selected according to the format type. The parsing strategy includes an optical character recognition parsing strategy for image format attachments and a text extraction parsing strategy for document format attachments. According to the target parsing strategy, for image-formatted attachments, the optical character recognition model is called to extract the text content in the image, and the large language model is used to perform visual understanding of the non-text elements in the image to obtain additional information related to insurance pricing information; for document-formatted attachments, the document parsing interface is called to extract the text content in the document. The extracted text content, the additional information, and the supplementary text are combined to form a set of text to be parsed that contains the insurance quote information; Using the large language model and based on a preset parsing prompt template, the field values ​​corresponding to the policyholder information, the insured information, the object information, and the insurance plan are identified and extracted from the text set to be parsed; The extracted field values ​​are cleaned and normalized to generate the initial data object containing the policyholder information, the insured information, the object information, and the insurance plan.

[0144] In one alternative embodiment, the verification module 202 is specifically used for: Obtain a list of key fields, which includes at least one field to be verified defined from the policyholder information, the insured information, the object information, and the insurance plan; Iterate through each field to be verified in the list of key fields, check whether there is a field value in the initial data object that corresponds to the field to be verified, and determine whether the field value conforms to the preset field format rules; If all fields to be verified have corresponding field values ​​and the field values ​​all conform to the field format rules, then the integrity verification is deemed to have passed. If at least one field to be validated is missing a corresponding field value or the field value does not conform to the field format rules, then the field to be validated will be marked as a missing field or a field with an incorrect format. The large language model is invoked to generate guidance information for natural language description based on the identifiers of the missing or formatted fields. The guidance information includes supplementary or corrective suggestions for the missing or formatted fields. The guidance information is output to the interactive interface.

[0145] In an alternative embodiment, the extraction module 203 is specifically used for: The occupational description information is extracted from the verified initial data object, wherein the occupational description information is a natural language description of the occupational name or occupational category text; The occupational description information is converted into a first feature vector. The second feature vector corresponding to each occupational record in the occupational knowledge base is traversed. The semantic similarity between the first feature vector and each second feature vector is calculated. The occupational knowledge base contains multiple occupational records. Each occupational record contains at least a standard occupational name, occupational code and risk level. All occupational records are sorted in descending order based on the semantic similarity, and occupational records with semantic similarity higher than a preset threshold are selected as candidate occupational information. If the number of candidate occupation information is one, then the candidate occupation information is determined as the standard occupation information; If there are multiple candidate occupation information entries, the multiple candidate occupation information entries will be presented to the user through the interactive interface, and the user will select and confirm the target occupation information from the multiple candidate occupation information entries, and the target occupation information will be used as the standard occupation information.

[0146] In an alternative embodiment, the matching module 204 is specifically used for: The insurance liability description information is extracted from the verified initial data object. The insurance liability description information is a liability name or liability clause text described in natural language. The description of the insured liability is converted into a third feature vector, and semantic matching is performed in the main insurance knowledge base and the supplementary insurance knowledge base respectively. Specifically, in the main insurance knowledge base, a first semantic similarity is calculated between the third feature vector and the fourth feature vector corresponding to each main insurance clause record, and in the supplementary insurance knowledge base, a second semantic similarity is calculated between the third feature vector and the fifth feature vector corresponding to each supplementary insurance clause record. The main insurance knowledge base contains multiple main insurance clause records, and each main insurance clause record contains at least a standard main insurance clause code and calculation rules. The supplementary insurance knowledge base contains multiple supplementary insurance clause records, and each supplementary insurance clause record contains at least a standard supplementary insurance clause code and calculation rules. All main insurance clause records with a first semantic similarity higher than a preset threshold are selected from the main insurance knowledge base as the first candidate set, and all supplementary insurance clause records with a second semantic similarity higher than a preset threshold are selected from the supplementary insurance knowledge base as the second candidate set. The first candidate set and the second candidate set are merged to form a candidate clause information set; If the candidate clause information set is empty, a prompt message is generated and returned; If the number of candidate clauses in the candidate clause information set is one, then the one candidate clause information is determined as the standard clause information; If there are multiple candidate clauses in the candidate clause information set, the multiple candidate clauses are presented to the user through the interactive interface, and the user selects and confirms the target clause from the multiple candidate clauses, and the target clause is used as the standard clause.

[0147] In an alternative embodiment, the fusion module 205 is specifically used for: Identify a first non-standard field corresponding to the occupational description information and a second non-standard field corresponding to the insurance liability description information from the initial data object; Write the occupational code and risk level from the standard occupational information into the initial data object, replacing the original occupational description information in the first non-standard field; and write the standard clause code and calculation rules from the clause information into the initial data object, replacing the original insurance liability description information in the second non-standard field. Iterate through the remaining fields in the initial data object to check if there are any fields to be supplemented that are related to the standard occupational information or the terms information; If the field to be supplemented exists, then according to the risk level in the standard occupational information or the calculation rule in the clause information, the preset field filling rule is invoked to assign a value to the field to be supplemented. The initial data object, after field replacement and supplementation, is repackaged according to a preset data structure to form the standardized quotation data object.

[0148] In an alternative embodiment, the output module 206 is specifically used for: Obtain the interface specifications of the core business system, which include a field mapping table, value conversion rules, and message structure definitions; Based on the field mapping table, the name of each data field in the standardized quotation data object is mapped to the corresponding interface field name in the interface specification; According to the value conversion rules, the value of each mapped data field is converted into a target value that is consistent with the data type and format required by the interface specification; According to the message structure definition, the interface field names and their corresponding target values ​​are assembled into the quotation request message in a preset format; According to the interface call address and call method defined in the interface specification, the interface service provided by the core business system is invoked, and the quotation request message is transmitted to the core business system.

[0149] It should be noted that the beneficial effects of the insurance quotation information processing system 200 provided in the above embodiments are the same as those of the insurance quotation information processing method described above, and will not be repeated here. Furthermore, the system provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the system can be divided into different functional modules according to the actual situation to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, and will not be repeated here.

[0150] The insurance quote information processing system 200 of the present invention can be a computer program (including program code) running on a computer device. For example, the insurance quote information processing system 200 of the present invention is an application software that can be used to execute the corresponding steps in the insurance quote information processing method of the present invention.

[0151] In some embodiments, the insurance quote information processing system 200 of the present invention can be implemented in a combination of hardware and software. As an example, the insurance quote information processing system 200 of the present invention can be a processor in the form of a hardware decoding processor, which is programmed to execute the insurance quote information processing method of the present invention. For example, the processor in the form of a hardware decoding processor can be one or more application specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0152] The modules described in the embodiments of this invention can be implemented in software or hardware. The names of the modules are not, in some cases, limiting the scope of the module itself.

[0153] An electronic device according to an embodiment of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-described insurance quote information processing methods. That is, an electronic device according to an embodiment of the present invention may include, but is not limited to: a processor and a memory; the memory is used to store the computer program; the processor is used to execute the insurance quote information processing method shown in any embodiment of the present invention by calling the computer program.

[0154] In one alternative embodiment, an electronic device is provided, such as Figure 4 As shown, Figure 4 The illustrated electronic device 4000 includes a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, for example, via a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 4004 is not limited to one type, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present invention.

[0155] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0156] Bus 4002 may include a path for transmitting information between the aforementioned components. Bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 4002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The bus 4002 is represented by only one thick line, but this does not mean that there is only one bus or one type of bus.

[0157] The memory 4003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0158] The memory 4003 stores application code (computer program) for executing the present invention, and its execution is controlled by the processor 4001. The processor 4001 executes the application code stored in the memory 4003 to implement the content shown in the foregoing method embodiments.

[0159] Among them, electronic devices can also be terminal devices. A terminal device can be any terminal device that can install applications and access web pages through applications, including at least one of smartphones, tablets, laptops, desktop computers, smart speakers, smartwatches, smart TVs, and smart in-vehicle devices.

[0160] It should be noted that, Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0161] An embodiment of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the above-described insurance quote information processing methods.

[0162] Alternatively, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, a floppy disk, and an optical data storage device, etc.

[0163] In an exemplary embodiment, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the aforementioned insurance quote information processing method.

[0164] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0165] It should be understood that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0166] The computer-readable storage medium provided in this invention can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0167] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method shown in the above embodiments.

[0168] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.

[0169] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.

[0170] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this invention can be specifically implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, this invention can also be implemented as a computer program product contained in one or more computer-readable media, which includes computer-readable program code.

[0171] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for processing insurance quote information, characterized in that, include: The system receives an attachment file containing insurance quote information uploaded by the user and supplementary text input by the user. It uses a large language model to parse the attachment file, identifies the attachment type, extracts the initial structured quote information from the attachment content, and combines the supplementary text to generate an initial data object containing policyholder information, insured information, insured information, and insurance plan. The initial data object is subjected to integrity verification of key fields. If the verification fails, guidance information is generated to instruct the user to supplement the missing fields. Occupational description information is extracted from the verified initial data object, and semantic matching is performed in a preset occupational knowledge base to obtain standard occupational information containing occupational codes and risk levels. If the matching result is not unique, an interactive interface is triggered to obtain the user's selection confirmation from multiple candidate occupational information. The insurance liability description information is extracted from the verified initial data object, and semantic matching is performed in the preset main insurance knowledge base and supplementary insurance knowledge base to obtain the clause information containing standard clause codes and calculation rules. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate clause information. The standard occupational information and the terms information are merged with the initial data object, and the corresponding non-standard fields in the initial data object are replaced to form a standardized quotation data object; According to the interface specifications of the core business system, the standardized quotation data object is converted into a quotation request message and output to the core business system.

2. The insurance quote information processing method according to claim 1, characterized in that, The steps of parsing the attachment file using a large language model, identifying the attachment type, extracting initial structured quotation information from the attachment content, and generating an initial data object containing policyholder information, insured information, insured information, and insurance plan information by combining the supplementary text include: Identify the format type of the attachment file, which is an electronic file containing the insurance quote information; The corresponding target parsing strategy is selected according to the format type. The parsing strategy includes an optical character recognition parsing strategy for image format attachments and a text extraction parsing strategy for document format attachments. According to the target parsing strategy, for image-formatted attachments, the optical character recognition model is called to extract the text content in the image, and the large language model is used to perform visual understanding of the non-text elements in the image to obtain additional information related to insurance pricing information; for document-formatted attachments, the document parsing interface is called to extract the text content in the document. The extracted text content, the additional information, and the supplementary text are combined to form a set of text to be parsed that contains the insurance quote information; Using the large language model and based on a preset parsing prompt template, the field values ​​corresponding to the policyholder information, the insured information, the object information, and the insurance plan are identified and extracted from the text set to be parsed; The extracted field values ​​are cleaned and normalized to generate the initial data object containing the policyholder information, the insured information, the object information, and the insurance plan.

3. The insurance quote information processing method according to claim 1, characterized in that, Perform integrity checks on key fields of the initial data object. If the checks fail, generate guidance information instructing the user to complete the missing fields, including: Obtain a list of key fields, which includes at least one field to be verified defined from the policyholder information, the insured information, the object information, and the insurance plan; Iterate through each field to be verified in the list of key fields, check whether there is a field value in the initial data object that corresponds to the field to be verified, and determine whether the field value conforms to the preset field format rules; If all fields to be verified have corresponding field values ​​and the field values ​​all conform to the field format rules, then the integrity verification is deemed to have passed. If at least one field to be validated is missing a corresponding field value or the field value does not conform to the field format rules, then the field to be validated will be marked as a missing field or a field with an incorrect format. The large language model is invoked to generate guidance information for natural language description based on the identifiers of the missing or formatted fields. The guidance information includes supplementary or corrective suggestions for the missing or formatted fields. The guidance information is output to the interactive interface.

4. The insurance quote information processing method according to claim 3, characterized in that, Occupational description information is extracted from the verified initial data object, and semantic matching is performed in a preset occupational knowledge base to obtain standard occupational information containing occupational codes and risk levels. If the matching results are not unique, an interactive interface is triggered to obtain the user's confirmation of selection from multiple candidate occupational information, including: The occupational description information is extracted from the verified initial data object, wherein the occupational description information is a natural language description of the occupational name or occupational category text; The occupational description information is converted into a first feature vector. The second feature vector corresponding to each occupational record in the occupational knowledge base is traversed. The semantic similarity between the first feature vector and each second feature vector is calculated. The occupational knowledge base contains multiple occupational records. Each occupational record contains at least a standard occupational name, occupational code and risk level. All occupational records are sorted in descending order based on the semantic similarity, and occupational records with semantic similarity higher than a preset threshold are selected as candidate occupational information. If the number of candidate occupation information is one, then the candidate occupation information is determined as the standard occupation information; If there are multiple candidate occupation information entries, the multiple candidate occupation information entries will be presented to the user through the interactive interface, and the user will select and confirm the target occupation information from the multiple candidate occupation information entries, and the target occupation information will be used as the standard occupation information.

5. The insurance quote information processing method according to claim 4, characterized in that, The insured liability description information is extracted from the verified initial data object, and semantic matching is performed in the preset main insurance knowledge base and supplementary insurance knowledge base to obtain clause information containing standard clause codes and calculation rules. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate clause information, including: The insurance liability description information is extracted from the verified initial data object. The insurance liability description information is a liability name or liability clause text described in natural language. The description of the insured liability is converted into a third feature vector, and semantic matching is performed in the main insurance knowledge base and the supplementary insurance knowledge base respectively. Specifically, in the main insurance knowledge base, a first semantic similarity is calculated between the third feature vector and the fourth feature vector corresponding to each main insurance clause record, and in the supplementary insurance knowledge base, a second semantic similarity is calculated between the third feature vector and the fifth feature vector corresponding to each supplementary insurance clause record. The main insurance knowledge base contains multiple main insurance clause records, and each main insurance clause record contains at least a standard main insurance clause code and calculation rules. The supplementary insurance knowledge base contains multiple supplementary insurance clause records, and each supplementary insurance clause record contains at least a standard supplementary insurance clause code and calculation rules. All main insurance clause records with a first semantic similarity higher than a preset threshold are selected from the main insurance knowledge base as the first candidate set, and all supplementary insurance clause records with a second semantic similarity higher than a preset threshold are selected from the supplementary insurance knowledge base as the second candidate set. The first candidate set and the second candidate set are merged to form a candidate clause information set; If the candidate clause information set is empty, a prompt message is generated and returned; If the number of candidate clauses in the candidate clause information set is one, then the one candidate clause information is determined as the standard clause information; If there are multiple candidate clauses in the candidate clause information set, the multiple candidate clauses are presented to the user through the interactive interface, and the user selects and confirms the target clause from the multiple candidate clauses, and the target clause is used as the standard clause.

6. The insurance quote information processing method according to claim 5, characterized in that, The steps of merging the standard occupational information, the terms and conditions information, and the initial data object, and replacing the corresponding non-standard fields in the initial data object to form a standardized quotation data object, include: Identify a first non-standard field corresponding to the occupational description information and a second non-standard field corresponding to the insurance liability description information from the initial data object; Write the occupational code and risk level from the standard occupational information into the initial data object, replacing the original occupational description information in the first non-standard field; and write the standard clause code and calculation rules from the clause information into the initial data object, replacing the original insurance liability description information in the second non-standard field. Iterate through the remaining fields in the initial data object to check if there are any fields to be supplemented that are related to the standard occupational information or the terms information; If the field to be supplemented exists, then according to the risk level in the standard occupational information or the calculation rule in the clause information, the preset field filling rule is invoked to assign a value to the field to be supplemented. The initial data object, after field replacement and supplementation, is repackaged according to a preset data structure to form the standardized quotation data object.

7. The insurance quote information processing method according to any one of claims 1 to 6, characterized in that, The steps of converting the standardized quotation data object into a quotation request message and outputting it to the core business system according to the interface specifications of the core business system include: Obtain the interface specifications of the core business system, which include a field mapping table, value conversion rules, and message structure definitions; Based on the field mapping table, the name of each data field in the standardized quotation data object is mapped to the corresponding interface field name in the interface specification; According to the value conversion rules, the value of each mapped data field is converted into a target value that is consistent with the data type and format required by the interface specification; According to the message structure definition, the interface field names and their corresponding target values ​​are assembled into the quotation request message in a preset format; According to the interface call address and call method defined in the interface specification, the interface service provided by the core business system is invoked, and the quotation request message is transmitted to the core business system.

8. An insurance quote information processing system, characterized in that, include: The identification module is used to receive attachment files containing insurance quotation information uploaded by users and supplementary text input by users. It uses a large language model to parse the attachment files, identify the attachment type, extract the initial structured quotation information from the attachment content, and combine the supplementary text to generate an initial data object containing policyholder information, insured information, insured information, and insurance plan. The verification module is used to perform integrity verification on the key fields of the initial data object. If the verification fails, it generates guidance information to instruct the user to supplement the missing fields. The extraction module is used to extract occupational description information from the verified initial data object, perform semantic matching in a preset occupational knowledge base, and obtain standard occupational information containing occupational codes and risk levels. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate occupational information. The matching module is used to extract the description information of the insured liability from the verified initial data object, perform semantic matching in the preset main insurance knowledge base and supplementary insurance knowledge base, and obtain the clause information containing standard clause codes and calculation rules. If the matching result is not unique, the interactive interface is triggered to obtain the user's selection confirmation from multiple candidate clause information. The fusion module is used to merge the standard occupational information, the terms information and the initial data object, and replace the corresponding non-standard fields in the initial data object to form a standardized quotation data object; The output module is used to convert the standardized quotation data object into a quotation request message according to the interface specifications of the core business system, and output it to the core business system.

9. An electronic device, characterized in that, The electronic device includes a processor coupled to a memory storing at least one computer program, which is loaded and executed by the processor to enable the electronic device to implement the insurance quote information processing method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which, when executed by a processor, implements the insurance quote information processing method as described in any one of claims 1 to 7.

Images