Session interaction method and device for insurance service, equipment and medium

Abstract

The embodiment of the invention provides an insurance service session interaction method and device, equipment and a medium, and relates to the technical field of artificial intelligence or the field of insurance. In response to interaction triggered by a user, according to a first robot instruction text and a first user reply text of the current round of interaction, a reply result corresponding to a business process instruction of the current round of interaction in the first user reply text is determined. And when the first reply result represents that the business process instruction is completed, inputting all session contexts including the current round of interaction into the multi-round session large model to obtain a second robot instruction text of the next round of interaction output by the multi-round session large model. According to the method, the interaction instruction text of the robot in the next round is obtained according to the session context of the previous round or multiple rounds, the problems that an existing automatic case reporting system depends on fixed keywords and templates and is poor in flexibility are solved, the flexibility and accuracy of automatic processing of insurance services are improved, and user experience is optimized.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology or the insurance field, and in particular to a conversational interaction method, apparatus, device, and medium for insurance business. Background Art

[0002] With the continuous development of society and the improvement of people's living standards, insurance has gradually become an indispensable part of modern life. As a risk management tool, insurance can provide users with protection against future uncertainties, help reduce potential losses, and provide financial support in the event of an accident.

[0003] Existing insurance methods typically utilize automated reporting systems. These systems, through pre-set keyword or template matching, help users quickly fill out their report and initiate the claims process. These systems allow users to conveniently report incidents by phone or online, saving the time and cost of manual processing.

[0004] However, since the existing automatic reporting system relies on fixed keywords and templates, the system has low flexibility, making it difficult for the system to cope with complex dialogue scenarios and diverse user expressions. Summary of the Invention

[0005] The present application provides a conversational interaction method, apparatus, device and medium for insurance business, which is used to solve the technical problem that the existing automatic reporting system relies on fixed keywords and templates and has poor flexibility.

[0006] In a first aspect, the present application provides a conversational interaction method for insurance business, comprising:

[0007] In an insurance business conversation interaction, in response to a user triggering an interaction, determining a first reply result corresponding to a first business process instruction in the current round of interaction in the first user reply text based on a first robot instruction text and a first user reply text in the current round of interaction, wherein the first robot instruction text includes the first robot interaction text and the corresponding first business process instruction;

[0008] When the first reply result indicates that the first business process instruction has been completed, the entire conversation context including the current round of interaction is input into the multi-round conversation model to obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model, and the second robot instruction text includes the second robot interaction text and the corresponding second business process instruction.

[0009] In one possible implementation, determining, based on the first robot instruction text and the first user reply text of the current round of interaction, a first reply result corresponding to the first business process instruction in the first user reply text includes:

[0010] Identify the first business process instruction in the first robot instruction text, and parse the first user reply text according to the first business process instruction to obtain the first reply result.

[0011] In one possible implementation, parsing the first user reply text according to the first business process instruction to obtain the first reply result includes:

[0012] A prompt word is generated according to the first business process instruction and the first user reply text, and the prompt word is input into a universal large model to obtain the first reply result output by the universal large model.

[0013] In one possible implementation, it also includes:

[0014] Call the insurance application system interface to query the user's policy information, and replace the policy information field in the second robot instruction text with the user's policy information obtained by the query.

[0015] In one possible implementation, it also includes:

[0016] In the case where the second business process instruction is a summary instruction, insurance business information is extracted from all session contexts, and an insurance application system interface is called to generate an insurance business work order based on the insurance business information.

[0017] In one possible implementation, it also includes:

[0018] If the first reply result indicates that the first business process instruction is not completed, output the first robot instruction text again; or

[0019] When the first reply result indicates that the first business process instruction is not completed, a third robot instruction text for the next round of interaction is output, wherein the third robot instruction text includes a third robot interaction text and a corresponding third business process instruction, wherein the third robot interaction text is used to inquire whether to transfer to manual operation, and the third business process instruction indicates whether to determine whether to transfer to manual operation.

[0020] In one possible implementation, it also includes:

[0021] Obtaining historical insurance business conversation data, and marking the historical insurance business conversation data with business process instructions according to the insurance business process;

[0022] The labeled insurance business historical conversation data is used to train the big model to obtain the multi-round conversation big model.

[0023] In a second aspect, the present application provides a conversational interaction device for insurance services, comprising:

[0024] A response module, used to respond to user-triggered interactions during insurance business session interactions;

[0025] a determination module, configured to determine, based on a first robot instruction text and a first user reply text of a current round of interaction, a first reply result corresponding to a first business process instruction in the current round of interaction in the first user reply text, wherein the first robot instruction text includes the first robot interaction text and the corresponding first business process instruction;

[0026] An input module is used to input the entire conversation context including the current round of interaction into the multi-round conversation model when the first reply result indicates that the first business process instruction has been completed, and obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model, where the second robot instruction text includes the second robot interaction text and the corresponding second business process instruction.

[0027] In one possible implementation, the apparatus further includes: an identification module;

[0028] The recognition module is configured to recognize the first business process instruction in the first robot instruction text;

[0029] The device further comprises: a parsing module;

[0030] The parsing module is used to parse the first user reply text according to the first business process instruction to obtain the first reply result.

[0031] In one possible implementation, the apparatus further includes: a generating module;

[0032] The generating module is configured to generate a prompt word according to the first business process instruction and the first user reply text;

[0033] The input module is further used to input the prompt word into the universal large model to obtain the first reply result output by the universal large model.

[0034] In one possible implementation, the apparatus further includes: a calling module;

[0035] The calling module is used to call the insurance application system interface to query the user's insurance policy information;

[0036] The device further comprises: a replacement module;

[0037] The replacement module is used to replace the policy information field in the second robot instruction text with the user policy information obtained by query.

[0038] In one possible implementation, the apparatus further includes: an extraction module;

[0039] The extraction module is configured to extract insurance business information from the entire session context when the second business process instruction is a summary instruction;

[0040] The generating module is further used to call the insurance application system interface and generate an insurance business work order according to the insurance business information.

[0041] In one possible implementation, the device further includes: an output module;

[0042] The output module is configured to output the first robot instruction text again if the first reply result indicates that the first business process instruction is not completed; or

[0043] The output module is also used to output a third robot instruction text for the next round of interaction when the first reply result indicates that the first business process instruction is not completed. The third robot instruction text includes a third robot interaction text and a corresponding third business process instruction. The third robot interaction text is used to inquire whether to transfer to manual labor, and the third business process instruction indicates whether to judge whether to transfer to manual labor.

[0044] In one possible implementation, the apparatus further includes: an acquisition module;

[0045] The acquisition module is used to acquire insurance business historical conversation data;

[0046] The device further comprises: a marking module;

[0047] The annotation module is used to annotate the insurance business history conversation data with business process instructions according to the insurance business process;

[0048] The device further comprises: a training module;

[0049] The training module is used to train the large model using the labeled historical insurance business conversation data to obtain the multi-round conversation large model.

[0050] In a third aspect, an embodiment of the present application provides a conversation interaction device for insurance services, including: a memory, a processor;

[0051] The memory stores computer-executable instructions;

[0052] The processor executes the computer-executable instructions stored in the memory, so that the processor executes the above first aspect and / or various possible implementations of the first aspect.

[0053] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which computer-executable instructions are stored. When the computer-executable instructions are executed by a processor, they are used to implement the first aspect above and / or various possible implementation methods of the first aspect.

[0054] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the above first aspect and / or various possible implementation methods of the first aspect.

[0055] The insurance business conversation interaction method, apparatus, device, and medium provided in the present application, in response to user-triggered interactions, determine the response result corresponding to the business process instruction of the current round of interaction in the first user's response text based on the first robot instruction text and the first user's response text of the current round of interaction. When the first response result indicates that the business process instruction has been completed, the entire conversation context including the current round of interaction is input into a multi-round conversation model to obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model. By obtaining the interaction instruction text of the next round of robots based on the conversation context of one or more previous rounds, this method not only solves the problem of existing automatic reporting systems relying on fixed keywords and templates and having poor flexibility, but also improves the flexibility and accuracy of automated processing of insurance business and optimizes the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0057] Figure 1 Schematic diagram of the process of the conversational interaction method for insurance business provided in this application Figure 1 ;

[0058] Figure 2 Schematic diagram of the process of the conversational interaction method for insurance business provided in this application Figure 2 ;

[0059] Figure 3 A schematic diagram of the structure of the conversation interaction device for insurance business provided in this application;

[0060] Figure 4 A structural diagram of the conversational interaction device for insurance business provided in this application.

[0061] The above drawings illustrate specific embodiments of the present application, which will be described in more detail below. These drawings and the textual description are not intended to limit the scope of the present application in any way, but rather to illustrate the concepts of the present application to those skilled in the art by reference to specific embodiments. DETAILED DESCRIPTION

[0062] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0063] The collection, storage, use, processing, transmission, provision and disclosure of user personal information involved in the technical solution of this application comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0064] With the continuous advancement of technology and the increasing awareness of risk, insurance plays an increasingly important role in modern society. It not only provides users with protection, but also plays a key role in property loss, health protection, and other aspects, helping users effectively deal with various risks.

[0065] Existing insurance methods typically utilize automated reporting systems. These systems, through pre-set keyword or template matching, help users quickly fill out their report and initiate the claims process. These systems allow users to conveniently report incidents by phone or online, saving the time and cost of manual processing.

[0066] However, since the existing automatic reporting system relies on fixed keywords and templates, the system has low flexibility, making it difficult for the system to cope with complex dialogue scenarios and diverse user expressions.

[0067] In response to the above-mentioned problems, the present application provides a conversational interaction method for insurance business. In response to the interaction triggered by the user, based on the first robot instruction text and the first user reply text of the current round of interaction, the reply result corresponding to the business process instruction of the current round of interaction in the first user reply text is determined. When the first reply result indicates that the business process instruction has been completed, the entire conversation context including the current round of interaction is input into the multi-round conversation model to obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model. By obtaining the interaction instruction text of the next round of the robot based on the conversation context of one or more previous rounds, this method not only solves the problem of poor flexibility of the existing automatic reporting system that relies on fixed keywords and templates, but also improves the flexibility and accuracy of automated processing of insurance business and optimizes the user experience.

[0068] The following specific embodiments describe in detail the technical solution of the present application and how the technical solution of the present application solves the above-mentioned technical problems. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described below in conjunction with the accompanying drawings.

[0069] Figure 1 Schematic diagram of the process of the conversational interaction method for insurance business provided in this application Figure 1 .like Figure 1 As shown, the method includes:

[0070] S101. In an insurance business conversation interaction, in response to a user triggering an interaction, a first reply result corresponding to the first business process instruction of this round of interaction in the first user reply text is determined based on the first robot instruction text and the first user reply text of this round of interaction, wherein the first robot instruction text includes the first robot interaction text and the corresponding first business process instruction.

[0071] The first robot interaction text refers to the text content sent by the robot to the user in the first round of interaction. For example, if the first robot interaction text is "Hello, please tell me what you need to do," then in the first round of interaction, the text content that the robot needs to send to the user is "Hello, please tell me what you need to do."

[0072] The first business process instruction refers to a business process task related to the first round of interaction. For example, if the first business process instruction is a "business confirmation instruction," the business process task of the first round is "business confirmation."

[0073] The first user response text refers to the user's response to the first robot interaction text sent by the robot in the first round of interaction. For example, if the user responds "I want to file a car insurance report," the user's response to the first robot instruction text in the first round of interaction is "I want to file a car insurance report."

[0074] The first response indicates whether the first business process instruction has been completed. If the first response indicates that the first business process instruction has been completed, the first round of business process tasks has been completed and the next round can be entered. If the first response indicates that the first business process instruction has not been completed, the first round of business process tasks has not been completed and the current round of interaction tasks needs to be continued.

[0075] The purpose of this step is to judge the execution status of this round of business process based on the user's response content and prepare for the next round of interaction.

[0076] As you can understand, when a user triggers an insurance business conversation, the system responds to the user's trigger and begins the first round of complete business interaction. At this point, the system controls the robot to execute the first robot interaction text of the first round. Simultaneously, the system determines the first business process instructions of the first round. Next, the system determines the first response text provided by the user. Finally, based on the first user response text and the first robot interaction text sent by the robot, the system determines whether the first business process instructions of the first round are complete.

[0077] For example, if it is known that a user has triggered an insurance business session through a mobile phone, then based on the above information, the system will first execute the first round of the first robot interaction text "Hello, please tell me the business you need to handle", and at the same time, determine the first round of the first business process instruction "Business confirmation"; then, determine the first user reply text "I want to report a car insurance case" fed back by the user; finally, based on the first user reply text "I want to report a car insurance case" and the first robot interaction text "Hello, please tell me the business you need to handle", determine whether the first round of the first business process instruction "Business confirmation" is completed.

[0078] S102. When the first reply result indicates that the first business process instruction has been completed, the entire conversation context including the current round of interaction is input into the multi-round conversation model to obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model. The second robot instruction text includes the second robot interaction text and the corresponding second business process instruction.

[0079] The second robot interaction text refers to the text content sent by the robot to the user in the second round of interaction. For example, if the second robot interaction text is "Hello, is the license plate number you want to report the case "{License Plate to be filled in}", then in the second round of interaction, the text content that the robot needs to send to the user is "Hello, is the license plate number you want to report the case "{License Plate to be filled in}".

[0080] The second business process instruction refers to the business process task related to the second round of interaction. For example, if the second business process instruction is "confirm license plate instruction", then the business process task of the second round is "confirm license plate".

[0081] The purpose of this step is to generate the instruction text for the next round of interaction based on the entire conversation context of the previous round.

[0082] It is understandable that the tasks in different rounds will be different. For example, the business process task of the first round is "business determination", the business process task of the second round is "license plate confirmation", and the business process task of the third round is "identifying the time of accident".

[0083] Therefore, if the first response indicates that the first business process instruction has been completed, it means that the current round of business process tasks has been successfully processed and can proceed to the next stage. To ensure that the next round of interaction is based on the previous conversation and user needs, the system needs to pass the entire conversation content of the current round to the multi-turn conversation model. This allows the multi-turn conversation model to understand the context of the entire conversation and accurately generate the instructions required for the next round of interaction, ensuring that the system can naturally and smoothly guide the user to the next business step.

[0084] For example, let's say the first robot interaction text in the first round is "Hello, please state the service you need to handle," the first service process instruction is "Service confirmation instruction," and the first user response text provided by the user is "I want to file a car insurance report." Based on this information, if the first service process instruction is determined to be completed, the entire conversation context, including the current round of interaction, can be input into the multi-round conversation model. The result is the second robot interaction text in the next round being "Hello, the license plate number you want to report is "{license plate to be filled in}" and the corresponding second service process instruction being "Confirm license plate instruction."

[0085] An embodiment of the present application provides a method for conversational interaction in insurance services. In an insurance service conversational interaction, the method, in response to a user-triggered interaction, determines, based on the first robot instruction text and the first user reply text of the current round of interaction, a response result corresponding to the business process instruction in the current round of interaction in the user reply. When the first reply result indicates that the business process instruction has been completed, the entire conversation context, including the current round of interaction, is input into a multi-round conversation model, and the model outputs the instruction text for the next round of interaction, which includes the new robot interaction text and the corresponding business process instruction.

[0086] This method introduces a multi-round conversation model that can generate a new round of robot interaction text and a new round of business process instructions. It not only solves the problem of the existing automatic reporting system relying on fixed keywords and templates and having poor flexibility, but also enables the system to understand customer intentions more accurately and respond flexibly to diverse conversation scenarios, thereby significantly improving the flexibility and versatility of the system, greatly enhancing the system's integration capabilities, and thus improving the user experience during use.

[0087] Figure 2 Schematic diagram of the process of the conversational interaction method for insurance business provided in this application Figure 2 ,like Figure 2 As shown, this embodiment Figure 1 Based on the embodiment, a conversation interaction method for insurance business is described in detail. The method includes:

[0088] S201. In an insurance business session interaction, respond to a user triggering an interaction.

[0089] The explanation of step S201 refers to the explanation of the above embodiment and will not be repeated here.

[0090] S202: Identify a first business process instruction in the first robot instruction text, and parse the first user reply text according to the first business process instruction to obtain a first reply result, wherein the first robot instruction text includes the first robot interaction text and the corresponding first business process instruction.

[0091] The purpose of identifying the first business process instruction in the first robot instruction text is to determine the business process tasks to be executed in the first round. The purpose of parsing the first user reply text based on the first business process instruction to obtain the first reply result is to determine whether the business process tasks in the first round have been completed.

[0092] It can be understood that, first of all, the first robot instruction text is the text content that the system robot needs to send to the user in the first round after the user triggers the insurance business session.

[0093] Secondly, the first user reply text is the reply content made by the user in response to the content corresponding to the instruction text after receiving the instruction text.

[0094] Therefore, by parsing the first user's reply text through the first business process instruction, it can be determined whether the user has completed the business process task that needs to be executed in the first round.

[0095] Optionally, the present application provides a possible implementation method for parsing the first user reply text according to the first business process instruction to obtain a first reply result, including: generating prompt words according to the first business process instruction and the first user reply text, inputting the prompt words into a general big model, and obtaining the first reply result output by the general big model.

[0096] The prompt words are used to guide the general large model to judge different rounds of business process tasks. The prompt words for different rounds are different. For example, if the first business process instruction of the first round of interaction is "business confirmation instruction" and the first user's reply text is "I want to report a crime", the prompt word is "report judgment"; if the first business process instruction of the first round of interaction is "business confirmation instruction" and the first user's reply text is "I want rescue", the prompt word is "rescue judgment"; if the first business process instruction of the first round of interaction is "business confirmation instruction" and the first user's reply text is "I want to complain", the prompt word is "complaint judgment".

[0097] The general large model is used to determine whether business process tasks in different rounds are completed. For example, in the first round of interaction, if the business process instruction is "Business confirmation instruction" and the user's response text is "I want to report a crime," the system-generated prompt word is "Report judgment." The purpose is to allow the general large model to determine whether the user has clearly stated the report request. In the second round of interaction, if the business process instruction is "License plate confirmation instruction" and the user's response text is "Yes," the prompt word generated is "License plate confirmation," which guides the large model to confirm whether the user has provided the license plate information.

[0098] The purpose of this step is to guide the general large model to determine whether each round of business process tasks is completed through prompt words in different rounds.

[0099] S203. When the first reply result indicates that the first business process instruction has been completed, the entire conversation context including the current round of interaction is input into the multi-round conversation model to obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model. The second robot instruction text includes the second robot interaction text and the corresponding second business process instruction.

[0100] The explanation of step S203 is similar to that of the above step S102 and will not be repeated here.

[0101] S204: Call the insurance application system interface to query the user's insurance policy information, and replace the insurance policy information field in the second robot instruction text with the user's insurance policy information obtained by the query.

[0102] Different insurance business scenarios correspond to different user policy information. For example, if the insurance business scenario is "reporting a crime," then it is necessary to obtain user policy information related to the reporting scenario; if the insurance business scenario is "damage assessment," then it is necessary to obtain user policy information related to the damage assessment scenario; if the insurance business scenario is "policy supplementation," then it is necessary to obtain user policy information related to the policy supplementation scenario.

[0103] The purpose of calling the insurance application system interface to query the user's policy information in this step is to replace the policy information field in the second robot instruction text generated by the system according to the detailed policy data related to the user pre-stored by the system.

[0104] Understandably, in an insurance business interaction scenario, when the system generates a second robot instruction text, if the text contains placeholders to be filled in, this indicates that the instruction is a template that needs to be completed. In this case, the system needs to first call the insurance application system interface to query the detailed policy information related to the user, and then replace the placeholders in the instruction text with the query results, thereby generating a unique and complete second robot instruction text for the user.

[0105] For example, assuming that the second robot instruction text is "Hello, is the license plate number you want to report the case "{license plate to be filled in}", and the user's license plate number in the user's insurance policy information is "11111", then based on the above information, the user's second robot instruction text "Hello, is the license plate number you want to report the case "111111" can be generated.

[0106] This step calls the insurance application system interface to query the user's insurance policy information, for example, by querying the information from the insurance application system based on the user's mobile phone number. This application does not impose any special restrictions on this.

[0107] S205: When the second business process instruction is a summary instruction, extract insurance business information from all session contexts, call the insurance application system interface, and generate an insurance business work order according to the insurance business information.

[0108] The summary instruction is used to indicate that the conversation interaction process has ended.

[0109] The purpose of this step is to integrate and process the insurance business information contained in the conversation context of all rounds, and generate a specific insurance business work order by calling the insurance application system interface.

[0110] As you can understand, at specific aggregation stages in the business process, the system performs a comprehensive information integration task. It first consolidates the insurance business information contained in all rounds of conversation context. Then, leveraging its interface with insurance application systems, the system transforms this integrated information into standardized insurance business work orders.

[0111] For example, if the second business process instruction for the case reporting scenario is "summarize case reporting information instruction", then based on the above information, the system will organize all the conversation contexts of the previous rounds to obtain a business insurance work order.

[0112] S206: When the first reply result indicates that the first business process instruction is not completed, output the first robot instruction text again.

[0113] The purpose of this step is to prompt the system to continue executing unfinished tasks until the system confirms that the first round of business process tasks has been completed.

[0114] It is understandable that when the system detects that the first reply result indicates that the first business process instruction is not completed, in order to avoid process interruption or task omission, the system will trigger the operation of outputting the first robot instruction text again to ensure that the business process can be completed accurately according to the predetermined process steps.

[0115] S207. When the first reply result indicates that the first business process instruction is not completed, output a third robot instruction text for the next round of interaction. The third robot instruction text includes a third robot interaction text and a corresponding third business process instruction. The third robot interaction text is used to inquire whether to transfer to manual labor, and the third business process instruction indicates whether to determine whether to transfer to manual labor.

[0116] The purpose of this step is to give users the right to choose, so as to give users the right to judge whether manual customer service intervention is needed based on the specific situation.

[0117] Understandably, if the first response indicates that the first business process instruction is unfinished, the system will generate a third robot instruction text based on pre-set logic. This text not only contains the interactive text asking the user whether to transfer to a human, but also includes the corresponding business process instructions for determining whether to initiate the human service process. This not only ensures that the problem can be resolved promptly and effectively, but also reflects the system's pursuit of business process flexibility.

[0118] S208: Obtain insurance business historical conversation data, and mark the insurance business historical conversation data with business process instructions according to the insurance business process.

[0119] The purpose of acquiring historical insurance business conversation data is to collect conversation records generated during past insurance business exchanges. The purpose of annotating the historical insurance business conversation data with business process instructions according to the insurance business process is to clarify the specific tasks of each process node.

[0120] It is understandable that insurance business historical conversation data records a variety of information such as user consultations, service requests, and problem answers. By collecting these historical conversation data, the system can fully understand the user's behavior and needs during the insurance business process.

[0121] Furthermore, by annotating these historical conversation data with key information according to the insurance business process, such as user requests, system or manual responses, and business process nodes, the originally scattered and disordered conversation content can be converted into a clearly structured and easy-to-analyze data set, which will help the system clarify the specific tasks of each process node.

[0122] S209: Use the labeled insurance business historical conversation data to train the large model to obtain a multi-round conversation large model.

[0123] The purpose of this step is to use the labeled historical insurance business conversation data to train the big model, so as to obtain a big model that can accurately understand the user's complex needs in the insurance business scenario and can smoothly conduct multiple rounds of conversations.

[0124] As you can understand, the annotated historical insurance conversation data presents the conversational flow of insurance business scenarios in a structured form. Therefore, by using this structured data, the big model can learn the conversational logic in insurance business scenarios, the diverse expression patterns of user needs, and the key steps of business processing.

[0125] Therefore, the trained multi-round dialogue model can accurately understand the complex needs of users in insurance business scenarios, and can conduct multi-round dialogues smoothly, providing users with more accurate and personalized services.

[0126] The conversational interaction method for insurance business provided in the embodiment of the present application first identifies and parses the first business process instruction in the first robot instruction text after the user triggers the interaction, and generates a first reply result based on the parsing result. If the first business process instruction has been completed, the entire conversation context is input into the multi-round conversation big model to generate a second robot instruction text containing the second robot interaction text and the second business process instruction, and the insurance application system interface is called to query the user's policy information to replace the policy information field therein. If the second business process instruction is a summary instruction, the insurance business information in the entire conversation context is extracted to generate an insurance business work order. If the first business process instruction is not completed, the first robot instruction text is output again or a third robot instruction text is output asking whether to transfer to manual operation according to the situation. In addition, the method also obtains historical conversation data of insurance business and annotates business process instructions, and uses the annotated data to train the big model to complete the training of the multi-round conversation big model.

[0127] This method, by employing a large, multi-turn conversation model based on instruction enhancement, effectively addresses the technical challenges of existing automated reporting systems, which rely on fixed keywords and templates and suffer from limited flexibility. This not only enables the system to more accurately understand user intent and flexibly respond to various insurance business scenarios, but also improves the user experience. Furthermore, by training the large model using a large amount of historical insurance business conversation data, this method significantly enhances its ability to understand and process insurance business scenarios. This enables the trained large, multi-turn conversation model to more accurately grasp customer intent and needs, providing conversational interactions that better meet user needs.

[0128] Figure 3 A structural diagram of the conversational interaction device for insurance business provided in this application, such as Figure 3 As shown, the conversation interaction device 300 for insurance services provided in this embodiment includes:

[0129] The response module 301 is used to respond to user-triggered interactions during insurance business session interactions;

[0130] A determination module 302 is configured to determine, based on a first robot instruction text and a first user reply text of the current round of interaction, a first reply result corresponding to a first business process instruction in the first user reply text of the current round of interaction, wherein the first robot instruction text includes the first robot interaction text and the corresponding first business process instruction;

[0131] The input module 303 is used to input the entire conversation context including the current round of interaction into the multi-round conversation model when the first reply result indicates that the first business process instruction has been completed, and obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model, wherein the second robot instruction text includes the second robot interaction text and the corresponding second business process instruction.

[0132] In one possible implementation, the apparatus further includes: an identification module 304;

[0133] The recognition module 304 is configured to recognize the first business process instruction in the first robot instruction text;

[0134] The device further includes: a parsing module 305;

[0135] The parsing module 305 is used to parse the first user reply text according to the first business process instruction to obtain the first reply result.

[0136] In one possible implementation, the apparatus further includes: a generating module 306;

[0137] The generating module 306 is configured to generate a prompt word according to the first business process instruction and the first user reply text;

[0138] The input module 303 is further configured to input the prompt word into the general large model to obtain the first reply result output by the general large model.

[0139] In one possible implementation, the apparatus further includes: a calling module 307;

[0140] The calling module 307 is used to call the insurance application system interface to query the user's insurance policy information;

[0141] The apparatus further includes: a replacement module 308;

[0142] The replacement module 308 is used to replace the policy information field in the second robot instruction text with the user policy information obtained by query.

[0143] In one possible implementation, the apparatus further includes: an extraction module 309;

[0144] The extraction module 309 is configured to extract insurance business information from the entire session context when the second business process instruction is a summary instruction;

[0145] The generating module 306 is further configured to call an insurance application system interface and generate an insurance business work order according to the insurance business information.

[0146] In one possible implementation, the apparatus further includes: an output module 310;

[0147] The output module 310 is configured to output the first robot instruction text again if the first reply result indicates that the first business process instruction is not completed; or

[0148] The output module 310 is also used to output a third robot instruction text for the next round of interaction when the first reply result indicates that the first business process instruction is not completed. The third robot instruction text includes a third robot interaction text and a corresponding third business process instruction. The third robot interaction text is used to inquire whether to transfer to manual labor, and the third business process instruction indicates whether to judge whether to transfer to manual labor.

[0149] In one possible implementation, the apparatus further includes: an acquisition module 311;

[0150] The acquisition module 311 is used to acquire insurance business historical conversation data;

[0151] The apparatus further includes: a labeling module 312;

[0152] The annotation module 312 is used to annotate the insurance business history conversation data with business process instructions according to the insurance business process;

[0153] The apparatus further includes: a training module 313;

[0154] The training module 313 is used to train the large model using the annotated insurance business historical conversation data to obtain the multi-round conversation large model.

[0155] The conversational interaction device for insurance business provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail in this embodiment.

[0156] Figure 4 This is a structural diagram of the conversation interaction device for the insurance business provided in this application. Figure 4 As shown, the electronic device 400 provided in this embodiment includes: at least one processor 401 and a memory 402. Optionally, the device 400 further includes a communication component 403. The processor 401, the memory 402 and the communication component 403 are connected via a bus 404.

[0157] In a specific implementation process, at least one processor 401 executes the computer-executable instructions stored in the memory 402, so that the at least one processor 401 performs the above method.

[0158] The specific implementation process of the processor 401 can be found in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here in this embodiment.

[0159] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASICs), etc. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the method disclosed in the present invention may be directly executed by a hardware processor or by a combination of hardware and software modules within the processor.

[0160] The memory may include random access memory (RAM) and may also include non-volatile memory (NVM), such as at least one disk storage.

[0161] A bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus. Buses can be categorized as address buses, data buses, and control buses. For ease of illustration, the buses in the drawings of this application are not limited to just one bus or just one type of bus.

[0162] The present application also provides a computer program product, including a computer program, which implements the above method when executed by a processor.

[0163] The present application also provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the above method is implemented.

[0164] The readable storage medium may be implemented by any type of volatile or non-volatile memory device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium may be any available medium that can be accessed by a general-purpose or special-purpose computer.

[0165] An exemplary readable storage medium is coupled to a processor so that the processor can read information from the readable storage medium and write information to the readable storage medium. Of course, the readable storage medium can also be an integral part of the processor. The processor and the readable storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the readable storage medium can also exist in the device as discrete components.

[0166] The division of units is merely a logical functional division; actual implementations may employ alternative divisions, such as combining or integrating multiple units or components into another system, or omitting or disabling certain features. Furthermore, any direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units, either through an interface, electrical, mechanical, or other means.

[0167] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0168] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0169] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the method of the present invention. The aforementioned storage medium includes various media that can store program code, such as USB flash drives, mobile hard drives, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical disks.

[0170] Those skilled in the art will appreciate that all or part of the steps in the above-described method embodiments can be implemented using hardware associated with program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0171] It should be noted that for the aforementioned method embodiments, for the sake of simplicity, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all optional embodiments, and the actions and modules involved are not necessarily required by this application.

[0172] It should be further noted that, although the various steps in the flowchart are shown in sequence as indicated by the arrows, these steps are not necessarily performed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps may be performed in other orders. Moreover, at least a portion of the steps in the flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily performed at the same time, but may be performed at different times. The execution order of these sub-steps or stages is not necessarily to be performed in sequence, but may be performed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0173] It should be understood that the above-described device embodiments are merely illustrative, and the device of the present application may also be implemented in other ways. For example, the division of units / modules in the above-described embodiments is merely a logical functional division, and actual implementations may employ other division methods. For example, multiple units, modules, or components may be combined or integrated into another system, or some features may be omitted or not implemented.

[0174] In addition, unless otherwise specified, the functional units / modules in the various embodiments of the present application may be integrated into a single unit / module, each unit / module may exist physically separately, or two or more units / modules may be integrated together. The aforementioned integrated units / modules may be implemented in the form of hardware or software program modules.

[0175] If an integrated unit / module is implemented in hardware, the hardware may be digital circuits, analog circuits, etc. The physical implementation of the hardware structure includes, but is not limited to, transistors, memristors, etc. Unless otherwise specified, the processor may be any appropriate hardware processor, such as a CPU, GPU, FPGA, DSP, and ASIC. Unless otherwise specified, the storage unit may be any appropriate magnetic storage medium or magneto-optical storage medium, such as resistive random access memory (RRAM), dynamic random access memory (DRAM), static random access memory (SRAM), enhanced dynamic random access memory (EDRAM), high-bandwidth memory (HBM), hybrid memory cube (HMC), etc.

[0176] If the integrated unit / module is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a memory and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes: U disk, read-only memory (ROM), random access memory (RAM), mobile hard disk, magnetic disk, or optical disk, etc., various media that can store program code.

[0177] In the above embodiments, the description of each embodiment has its own emphasis. For parts not described in detail in a particular embodiment, please refer to the relevant description of other embodiments. The technical features of the above embodiments can be combined in any way. To keep the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0178] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, and the true scope and spirit of the present application are indicated by the following claims.

[0179] It should be understood that the present application is not limited to the exact structure described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A conversational interaction method for insurance business, characterized in that: include: In an insurance business conversation interaction, in response to a user triggering an interaction, determining a first reply result corresponding to a first business process instruction in the current round of interaction in the first user reply text based on a first robot instruction text and a first user reply text in the current round of interaction, wherein the first robot instruction text includes the first robot interaction text and the corresponding first business process instruction; When the first reply result indicates that the first business process instruction has been completed, the entire conversation context including the current round of interaction is input into the multi-round conversation model to obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model, and the second robot instruction text includes the second robot interaction text and the corresponding second business process instruction.

2. The method according to claim 1, characterized in that The determining, based on the first robot instruction text and the first user reply text of the current round of interaction, a first reply result corresponding to the first business process instruction of the current round of interaction in the first user reply text includes: Identify the first business process instruction in the first robot instruction text, and parse the first user reply text according to the first business process instruction to obtain the first reply result.

3. The method according to claim 2, characterized in that The parsing the first user reply text according to the first business process instruction to obtain the first reply result includes: A prompt word is generated according to the first business process instruction and the first user reply text, and the prompt word is input into a universal large model to obtain the first reply result output by the universal large model.

4. The method according to any one of claims 1 to 3, characterized in that Also includes: Call the insurance application system interface to query the user's policy information, and replace the policy information field in the second robot instruction text with the user's policy information obtained by the query.

5. The method according to any one of claims 1 to 3, characterized in that Also includes: In the case where the second business process instruction is a summary instruction, insurance business information is extracted from all session contexts, and an insurance application system interface is called to generate an insurance business work order based on the insurance business information.

6. The method according to any one of claims 1 to 3, characterized in that Also includes: If the first reply result indicates that the first business process instruction is not completed, output the first robot instruction text again; or, When the first reply result indicates that the first business process instruction is not completed, a third robot instruction text for the next round of interaction is output, wherein the third robot instruction text includes a third robot interaction text and a corresponding third business process instruction, wherein the third robot interaction text is used to inquire whether to transfer to manual operation, and the third business process instruction indicates whether to determine whether to transfer to manual operation.

7. The method according to any one of claims 1 to 3, characterized in that Also includes: Obtaining historical insurance business conversation data, and marking the historical insurance business conversation data with business process instructions according to the insurance business process; The labeled insurance business historical conversation data is used to train the big model to obtain the multi-round conversation big model.

8. A conversational interaction device for insurance business, characterized in that: include: A response module, used to respond to user-triggered interactions during insurance business session interactions; a determination module, configured to determine, based on a first robot instruction text and a first user reply text of a current round of interaction, a first reply result corresponding to a first business process instruction in the current round of interaction in the first user reply text, wherein the first robot instruction text includes the first robot interaction text and the corresponding first business process instruction; An input module is used to input the entire conversation context including the current round of interaction into the multi-round conversation model when the first reply result indicates that the first business process instruction has been completed, and obtain the second robot instruction text for the next round of interaction output by the multi-round conversation model, where the second robot instruction text includes the second robot interaction text and the corresponding second business process instruction.

9. An electronic device, characterized in that: include: a processor, and a memory communicatively connected to the processor; The memory stores computer-executable instructions; The processor executes the computer-executable instructions stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer-executable instructions, which are used to implement the method according to any one of claims 1 to 7 when executed by a processor.

Images