Car insurance business question and answer interaction method and device based on large model, medium and equipment

Abstract

The invention discloses a car insurance business question and answer interaction method and device based on a large model, a medium and equipment, and relates to the technical field of artificial intelligence. The method comprises the following steps: inputting a car insurance consultation question of a user into a preset car insurance large model to generate a reply result; if the reply result is a quotation function aiming at the vehicle insurance demand of the user, sending the quotation function to a vehicle insurance quotation system for quotation calculation, writing a quotation calculation result into a cache after the vehicle insurance quotation system completes calculation, and returning a target quotation verbal skill corresponding to the quotation function to the user; when the user inquires whether the quotation succeeds or not in the next round of conversation, whether a quotation calculation result exists in the cache or not is judged; and if the quotation calculation result exists in the cache, replacing the quotation inquiry problem of the user with the quotation calculation result, and inputting the quotation calculation result into a preset vehicle insurance large model to generate quotation detail information so as to reply to the user. According to the invention, the car insurance consultation problem of the user can be accurately answered, and the car insurance business processing efficiency and the success rate of car insurance marketing are improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a method, device, medium and equipment for interactive question and answer of auto insurance business based on a large model. Background Art

[0002] Auto insurance business refers to the insurance business where insurance companies provide vehicle insurance to vehicle owners to compensate for property losses and personal injury caused by accidents. Auto insurance is one of the largest markets in the current insurance market and one of the most important products of insurance companies.

[0003] At present, it is usually the sales staff who accept the users' auto insurance consultation and then sell the auto insurance to the users. However, this manual sales method is too dependent on the sales staff's work experience. Different sales staff have different professional levels and business capabilities, so it is difficult to guarantee the accuracy of the answers to users' questions, which will affect the processing efficiency of auto insurance business and the success rate of auto insurance marketing. Summary of the invention

[0004] In view of this, the present application provides a method, device, medium and equipment for interactive question and answer of auto insurance business based on a large model, the main purpose of which is to accurately respond to users' auto insurance inquiries, thereby improving the processing efficiency of auto insurance business and the success rate of auto insurance marketing.

[0005] According to the first aspect of the present application, a method for interactive question and answer of auto insurance business based on a large model is provided, the method comprising: Obtain the user's auto insurance consultation questions; Inputting the auto insurance consultation question into a preset auto insurance big model to generate a response result corresponding to the auto insurance consultation question; If the reply result is a quotation function for the user's auto insurance demand, the quotation function is sent to the auto insurance quotation system for quotation calculation. After the auto insurance quotation system completes the calculation, the quotation calculation result is written into the cache, and the target quotation script corresponding to the quotation function is returned to the user at the same time; When the user asks whether the quotation is successful in the next round of dialogue, determining whether the quotation calculation result exists in the cache; If the quotation calculation result exists in the cache, the user's quotation inquiry question is replaced with the quotation calculation result and input into the preset auto insurance model to generate quotation details information to reply to the user.

[0006] According to a second aspect of the present application, a large model-based auto insurance business question-answering interactive device is provided, the device comprising: An acquisition unit, used to acquire the user's auto insurance consultation questions; A generating unit, used for inputting the auto insurance consultation question into a preset auto insurance big model to generate a response result corresponding to the auto insurance consultation question; a calculation unit, configured to send the quotation function to the auto insurance quotation system for quotation calculation if the reply result is a quotation function for the user's auto insurance demand, write the quotation calculation result into a cache after the auto insurance quotation system completes the calculation, and return the target quotation script corresponding to the quotation function to the user; A determination unit, configured to determine whether the quotation calculation result exists in the cache when the user asks whether the quotation is successful in the next round of dialogue; The generating unit is further configured to replace the user's quotation inquiry question with the quotation calculation result if the quotation calculation result exists in the cache, input the result into the preset auto insurance model, generate quotation details information, and reply to the user.

[0007] According to a third aspect of the present application, a storage medium is provided, on which a computer program is stored, and when the program is executed by a processor, the above-mentioned large model-based auto insurance business question-and-answer interaction method is implemented.

[0008] According to the fourth aspect of the present application, an electronic device is provided, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein when the processor executes the program, the above-mentioned large model-based auto insurance business question-and-answer interaction method is implemented.

[0009] By means of the above technical scheme, the present application provides a method, device, medium and equipment for interactive question-answering of auto insurance business based on a large model. Compared with the manual sales method, the user's auto insurance consultation question can be input into the preset auto insurance large model to generate the corresponding reply result of the auto insurance consultation question. If the reply result is a quotation function for the user's auto insurance demand, the quotation function is sent to the auto insurance quotation system for quotation calculation. After the auto insurance quotation system completes the calculation, the quotation calculation result is written into the cache. At the same time, the target quotation speech corresponding to the quotation function is returned to the user. When the user asks whether the quotation is successful in the next round of dialogue, the user's quotation inquiry question is replaced with the quotation calculation result in the cache and input into the preset auto insurance large model to generate quotation details information to reply to the user. It can be seen that the present application can automatically complete the marketing task of the whole process by adopting the preset auto insurance large model, accurately answer the user's consultation questions about auto insurance, thereby improving the processing efficiency of auto insurance business and the success rate of auto insurance marketing, and saving human resources and costs. In addition, this application can avoid premature hanging up of the call due to the user waiting for the quote for too long by returning the target quotation words corresponding to the quotation function to the user during quotation calculation, thereby improving the user's call experience and further increasing the transaction rate of auto insurance business.

[0010] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings: Figure 1 A schematic diagram of a process flow of a large model-based auto insurance business question-answering interactive method provided in an embodiment of the present application is shown; Figure 2 The overall process diagram of the auto insurance business question-and-answer interaction provided in the embodiment of the present application is shown; Figure 3 A schematic diagram of the training method flow of a preset auto insurance large model provided in an embodiment of the present application is shown; Figure 4 A schematic diagram of the overall model training process provided by an embodiment of the present application is shown; Figure 5 A structural schematic diagram of a large model-based auto insurance business question-and-answer interactive device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0012] The present application will be described in detail below with reference to the accompanying drawings and in combination with embodiments. It should be noted that the embodiments and features in the embodiments of the present application can be combined with each other without conflict.

[0013] It is difficult for existing technologies to guarantee the accuracy of answers to user questions, which will affect the processing efficiency of auto insurance business and the success rate of auto insurance marketing.

[0014] In order to solve the above problems, the embodiment of the present invention provides a method for interactive question and answer of auto insurance business based on a large model, such as Figure 1 As shown, the method includes: Step 10: Obtain the user's auto insurance consultation questions.

[0015] The user's auto insurance consultation question may be a user's insurance combination quotation requirement. For example, the user's auto insurance consultation question is "calculate the price for me for only third party liability insurance and motor vehicle loss insurance".

[0016] In the embodiment of the present invention, in a real online conversation, a user can consult about car insurance through voice, and the car insurance consultation question is recorded in voice form. Then, the voice-based vehicle consultation question is translated by ASR to obtain the text-based vehicle consultation question, where ASR (Automatic Speech Recognition Transcription) refers to the process of automatically recognizing speech using a computer and recognizing speech signals into text.

[0017] Step 20: Input the auto insurance consultation question into a preset auto insurance model to generate a response result corresponding to the auto insurance consultation question.

[0018] The preset auto insurance big model is a pre-trained big model, which is a big model specially used in the auto insurance field obtained by training on the basis of the big model in the insurance field in the embodiment of the present invention, i.e., the auto insurance marketing customer service robot. The big models in the insurance field specifically include the Waterdrop Water Guardian big model, BERT-Insurance and other big models applied in the insurance field.

[0019] In the field of medical insurance, robots sell single insurance products. Compared with medical insurance, the marketing process of auto insurance often involves a combination of multiple insurance types (compulsory traffic insurance, third-party insurance, vehicle damage insurance, seat insurance, other commercial insurance, etc.). During the entire conversation, users will frequently modify the insurance combination, for example: remove the third-party insurance, only insure vehicle damage and compulsory traffic insurance, remove vehicle damage and seat insurance. Since different insurance combinations have different pricing coefficients, the insurance combination cannot be equivalent to a simple combination of insurance prices. The robot needs to request the auto insurance quotation system to recalculate the price at the right time according to the user's intention. Therefore, a large number of instructions are issued in the entire process of the conversation. Based on this, the embodiment of the present invention trains an auto insurance marketing customer service robot (preset auto insurance big model) based on the insurance field big model to complete the marketing task of the entire process.

[0020] For the embodiment of the present invention, after obtaining the auto insurance consultation question in text form, the auto insurance consultation question can be input into a preset auto insurance big model (auto insurance marketing customer service robot) to generate a response result for the auto insurance consultation question.

[0021] Step 30: If the reply result is a quotation function for the user's auto insurance demand, the quotation function is sent to the auto insurance quotation system for quotation calculation. After the auto insurance quotation system completes the calculation, the quotation calculation result is written into the cache, and the target quotation script corresponding to the quotation function is returned to the user.

[0022] The quotation function includes quotations for all insurance types involved in the user's auto insurance needs. In addition, different quotation functions correspond to different quotation scripts.

[0023] For the embodiment of the present invention, when the user's auto insurance consultation question is a requirement for a combined quote for insurance types, the preset auto insurance big model will generate a corresponding quote function as the answer result for this round of questions, and the quote function covers quotes for all insurance types. For example: User: I have had my car for 5 years. I don’t want to insure against damage to the car, but I want to cover all other insurance. How much will it cost? Agent: <Heuristic speech: OK, thank you for your support of my work. I will calculate the exact price for you according to your requirements. Please wait. > User: OK.

[0024] Seat: <Quotation function: Vehicle damage insurance: not required, third party liability insurance: 3 million, third party medical insurance external medicine: 100,000, driving and riding insurance: required, seat insurance: not required>.

[0025] After the preset auto insurance big model generates a quotation function, the quotation function will be sent to the auto insurance quotation system for quotation calculation. After the calculation is completed, the auto insurance quotation system will write the quotation calculation result into the redis database for reading. At the same time, a quotation script is returned to avoid the user waiting for the quotation for too long and hanging up the call in advance, which affects the user's call experience.

[0026] In order to return appropriate quotation words to the user during the quotation calculation process, the embodiment of the present invention pre-builds a mapping relationship between different quotation functions and different quotation words. When returning the quotation words, Based on the mapping relationship, the target quotation words corresponding to the quotation function are determined, and the target quotation words are returned to the user. An example of the mapping between the quotation function and the quotation words is as follows: Quotation function: <Quotation function: Vehicle damage insurance: not required, third party liability insurance: 3 million, third party medical insurance external medicine: 100,000, driving insurance: required, seat insurance: not required> Quotation script: Well, OK, let me tell you about your insurance plan. The compulsory traffic insurance plus the vehicle and vessel use tax will remain unchanged. Then the three insurances are three million, and the three insurances and external medicines are one hundred thousand. I will calculate the price for you based on this coverage. Here's the approximate price. Wait a moment.

[0027] Furthermore, when the user's auto insurance consultation question does not involve a combination quote for insurance types, the answer result generated by the preset auto insurance big model is not a quote function, and the answer result is directly returned to the user. For example: User: I would like to ask about car insurance.

[0028] Agent: <Okay, go ahead> Step 40: When the user asks whether the quotation is successful in the next round of dialogue, it is determined whether the quotation calculation result exists in the cache.

[0029] The cache may specifically be a redis database.

[0030] For the embodiment of the present invention, when the conversation proceeds to the next round and the user asks whether the quotation is successful, the preset auto insurance big model (auto insurance marketing customer service robot) tries to read the quotation calculation result from the redis database again.

[0031] Step 50: If the quotation calculation result exists in the cache, the user's quotation inquiry question is replaced with the quotation calculation result and input into the preset auto insurance model to generate quotation details information to reply to the user.

[0032] For the embodiment of the present invention, if the reading result of the preset auto insurance big model (auto insurance marketing customer service robot) is not empty, it means that the quotation calculation process has been completed. At this time, the user's quotation inquiry question (such as whether the quotation is successful) will be replaced with the quotation calculation result and input into the preset auto insurance big model to generate quotation details information to reply to the user. For example: Seat: <Quotation function: Vehicle damage insurance: not required, third party liability insurance: 3 million, third party medical insurance external medicine: 100,000, driving insurance: required, seat insurance: not required> User: <Quote calculation results: Third-party liability insurance coverage: 3 million, third-party liability insurance premium: 650 yuan, third-party medical insurance external medicine coverage: 100,000, third-party medical insurance external medicine premium: 150 yuan, driving and passenger insurance coverage: 1.73 million per seat, driving and passenger insurance premium: 286 yuan, vehicle and vessel tax premium: 514 yuan, compulsory traffic insurance premium: 592 yuan, total premium is 2,438 yuan> Agent: <Quote details: Yes, that's it. According to this year's situation, it includes 3 million yuan for third-party liability insurance, 100,000 yuan for medical insurance and external medicine, 1.73 million yuan for each seat for driving insurance, and the required vehicle and vessel tax and compulsory traffic insurance are also included. The total is about 2,438 yuan. What do you think of this price? Can we cooperate for one year? > Furthermore, if the quotation calculation result does not exist in the cache, the user's quotation inquiry question is input into the preset auto insurance big model to generate answer information for the quotation inquiry question.

[0033] Specifically, if the reading result of the preset auto insurance big model (auto insurance marketing customer service robot) is empty, it means that the quotation calculation process has not been completed. At this time, the user's quotation inquiry question (such as whether the quotation is successful) is directly input into the preset auto insurance big model to generate corresponding reply information, such as please wait, the quotation calculation has not been completed. The above-mentioned overall auto insurance business question-and-answer interaction process of the embodiment of the present invention is as follows Figure 2 shown.

[0034] The embodiment of the present invention can automatically complete the marketing task of the whole process by adopting a preset auto insurance big model, accurately answer the user's consultation questions about auto insurance, thereby improving the processing efficiency of auto insurance business and the success rate of auto insurance marketing, and saving human resources and costs. In addition, the embodiment of the present invention can avoid premature hanging up of the call because the user waits for the quotation for too long by returning the target quotation speech corresponding to the quotation function to the user during quotation calculation, thereby improving the user's call experience and further improving the transaction rate of auto insurance business.

[0035] Furthermore, before the embodiment of the present invention uses the preset auto insurance big model (auto insurance marketing customer service robot) to complete the marketing task of the whole process, it is necessary to pre-train the auto insurance big model. Figure 3 As shown, including: Step 60: Obtain the original telephone recording.

[0036] According to the embodiment of the present invention, a large number of original telephone recordings of users inquiring about auto insurance are collected in advance and used as sample data for training the large auto insurance model.

[0037] Step 70: Generate a conversation training sample based on the original telephone recording.

[0038] For the embodiment of the present invention, after obtaining the original telephone recording, text recognition is performed on the original telephone recording to obtain the conversation text, and then the conversation text is processed using the first preset universal generative model (universal generative model one) and the second preset universal generative model (universal generative model two) to obtain a conversation training sample.

[0039] When processing the dialogue text, the first preset universal generation model is used to perform dialogue text role separation, text error correction, data desensitization and data standardization on the dialogue text to obtain an initial dialogue sample; the second preset universal generation model is used to insert quotation information at a key position of the initial dialogue sample to obtain the dialogue training sample.

[0040] Among them, the quotation information includes: quotation function samples, quotation details samples, quotation sending samples and WeChat adding samples, etc. In addition, the first preset universal generation big model and the second preset universal generation big model specifically include Tencent Hunyuan big model, Baichuan big model, Zhipu AI big model, Tongyi Qianwen big model, etc. It should be noted that the first preset universal generation big model and the second preset universal generation big model in the embodiment of the present invention can also be other universal big models, and are not limited to the above examples.

[0041] Specifically, the original telephone recording is first translated into a conversation text by ASR. Then prompt1 is constructed to call the first preset universal generation model to complete the speech text role separation, text error correction, data desensitization and data standardization processing, so as to obtain the initial conversation sample. The result example of calling the first preset universal generation model is as follows: User: Company Address Seat: Our company is located at [company address] User: What are the special offers today? Agent: [User Nickname], today the company has a car insurance promotion [Promotion], and the price is very favorable.

[0042] Then, we construct prompt2 and call the second preset general generation model to insert heuristic speech samples, quotation function samples, quotation calculation result samples, quotation details samples, quotation sending samples, and WeChat account addition samples at key locations to obtain dialogue training samples. The result of calling the second preset general generation model is as follows: User: I have had my car for 5 years. I don’t want to insure against damage to the car, but I want to cover all other insurance. How much will it cost? Agent: <Heuristic speech: OK, thank you for your support of my work. I will calculate the exact price for you according to your requirements. Please wait. > User: OK.

[0043] Seat: <Quotation function: Vehicle damage insurance: not required, third party liability insurance: 3 million, third party medical insurance external medicine: 100,000, driving insurance: required, seat insurance: not required> User: <Details of quotation result: Third-party liability insurance coverage: 3 million, third-party liability insurance premium: 650 yuan, third-party medical insurance external medicine coverage: 100,000, third-party medical insurance external medicine premium: 150 yuan, driving and passenger insurance coverage: 1.73 million per seat, driving and passenger insurance premium: 286 yuan, vehicle and vessel tax premium: 514 yuan, compulsory traffic insurance premium: 592 yuan, total premium is 2,438 yuan> Agent: <Details of the exact quotation: Yes, that's it. According to this year's situation, it includes 3 million yuan for third-party liability insurance, 100,000 yuan for medical insurance and external medicine, 1.73 million yuan for each seat for driving insurance, and the required vehicle and vessel tax and compulsory traffic insurance are also included. The final price is about 2,438 yuan. What do you think of this price? Can we cooperate for one year? > Step 80: Based on the dialogue training samples, the insurance field big model is trained to obtain the preset auto insurance big model.

[0044] For the embodiment of the present invention, after obtaining the dialogue training sample, the dialogue training sample is used to complete the full-parameter SFT based on the insurance field big model, that is, full-parameter training and fine-tuning are performed to obtain the preset auto insurance big model (auto insurance marketing customer service robot). The overall training process of the preset auto insurance big model is as follows: Figure 4 shown.

[0045] The embodiment of the present invention can avoid new nonsense problems during the dialogue process by adopting the full-parameter SFT method, thereby ensuring the dialogue quality with the user.

[0046] Compared with the medical insurance field, the auto insurance field has no fixed process and the conversations often involve specific combinations of different insurance types. Therefore, the task scenarios of intelligent sales robots in the auto insurance field are more complicated. The embodiment of the present invention introduces quotation functions, quotation calculation results, quotation details, etc. in the conversation training samples. Based on the conversation training samples, fine-tuning is performed on the basis of the big model in the insurance field to obtain the auto insurance big model. The auto insurance big model is suitable for multiple rounds of conversations in the auto insurance field and can provide quotation feedback for multiple insurance type combinations.

[0047] The embodiment of the present invention provides a large-model-based auto insurance business question-answering interactive method, which can automatically complete the marketing task of the entire process by adopting a preset auto insurance large model, accurately answer the user's consultation questions about auto insurance, thereby improving the processing efficiency of auto insurance business and the success rate of auto insurance marketing, while saving human resources and costs. In addition, the embodiment of the present invention can avoid premature hanging up of the call because the user waits for the quote for too long by returning the target quotation script corresponding to the quotation function to the user during quotation calculation, thereby improving the user's call experience and further improving the transaction rate of auto insurance business.

[0048] Further, as Figure 1 and Figure 3 The specific implementation of the method shown in the embodiment provides a vehicle insurance business question-answering interactive device based on a large model, such as Figure 5 As shown, the device includes: an acquisition unit 101, a generation unit 102, a calculation unit 103 and a determination unit 104.

[0049] The acquisition unit 101 may be used to acquire the user's car insurance consultation questions.

[0050] The generating unit 102 may be used to input the auto insurance consultation question into a preset auto insurance big model to generate a response result corresponding to the auto insurance consultation question.

[0051] The calculation unit 103 can be used to send the quotation function to the auto insurance quotation system for quotation calculation if the reply result is a quotation function for the user's auto insurance demand, and write the quotation calculation result into the cache after the auto insurance quotation system completes the calculation, and return the target quotation words corresponding to the quotation function to the user.

[0052] The determination unit 104 may be configured to determine whether the quotation calculation result exists in the cache when the user inquires whether the quotation is successful in the next round of dialogue.

[0053] The generating unit 102 may also be used to replace the user's quotation inquiry question with the quotation calculation result if the quotation calculation result exists in the cache, and input it into the preset auto insurance big model to generate quotation details information to reply to the user.

[0054] In some embodiments, the device further comprises: a returning unit.

[0055] The returning unit may be configured to directly return the reply result to the user if the reply result is not a quotation function for the user's auto insurance demand.

[0056] In some embodiments, the apparatus further comprises: a mapping unit.

[0057] The mapping unit can be used to construct a mapping relationship between different quotation functions and different quotation scripts.

[0058] The calculation unit 103 can be specifically used to determine the target quotation language corresponding to the quotation function based on the mapping relationship, and return the target quotation language to the user, wherein the quotation function includes the insurance type quotation involved in the user's auto insurance demand.

[0059] In some embodiments, the generating unit 102 may also be used to input the user's quotation inquiry question into the preset auto insurance big model to generate answer information for the quotation inquiry question if the quotation calculation result does not exist in the cache.

[0060] In some embodiments, the device further comprises: a training unit.

[0061] The acquisition unit 101 may also be used to acquire the original telephone recording.

[0062] The generating unit 102 may also be configured to generate a conversation training sample based on the original telephone recording.

[0063] The training unit can be used to train the insurance field big model based on the dialogue training samples to obtain the preset auto insurance big model.

[0064] In some embodiments, the generating unit 102 includes: an identification module and a processing module.

[0065] The recognition module can be used to perform text recognition on the original telephone recording to obtain the conversation text.

[0066] The processing module can be used to process the dialogue text using the first preset universal generation model and the second preset universal generation model to obtain dialogue training samples.

[0067] In some embodiments, the processing module can be specifically used to use the first preset universal generation model to perform dialogue text role separation, text error correction, data desensitization and data standardization on the dialogue text to obtain an initial dialogue sample; use the second preset universal generation model to insert quotation information at key positions of the initial dialogue sample to obtain the dialogue training sample, wherein the quotation information includes: heuristic speech samples, quotation function samples, quotation calculation result samples and quotation details samples.

[0068] It should be noted that for other corresponding descriptions of the functional units involved in the large model-based auto insurance business question-answering interactive device provided in the embodiment of the present invention, please refer to Figure 1 and Figure 3 The corresponding description in will not be repeated here.

[0069] Based on the above Figure 1 and Figure 3 The method shown in the embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned Figure 1 and Figure 3 The shown method is a question-and-answer interactive method for auto insurance business based on a large model.

[0070] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), including a number of instructions for enabling an electronic device (which can be a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present application.

[0071] Based on the above Figure 1 and Figure 3 The method shown, and Figure 5In order to achieve the above-mentioned purpose, the embodiment of the present application also provides an electronic device, which can be a personal computer, a tablet computer, a server, or other network equipment, etc. The device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 and Figure 3 The shown method is a question-and-answer interactive method for auto insurance business based on a large model.

[0072] Optionally, the above-mentioned physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0073] Those skilled in the art will appreciate that the above-mentioned physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.

[0074] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the above-mentioned physical device, and supports the operation of the information processing program and other software and / or programs. The network communication module is used to realize the communication between the components inside the storage medium, and the communication with other hardware and software in the information processing physical device.

[0075] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform, or by hardware.

[0076] The embodiment of the present invention can automatically complete the marketing task of the whole process by adopting a preset auto insurance big model, accurately answer the user's consultation questions about auto insurance, thereby improving the processing efficiency of auto insurance business and the success rate of auto insurance marketing, and saving human resources and costs. In addition, the embodiment of the present invention can avoid premature hanging up of the call because the user waits for the quotation for too long by returning the target quotation speech corresponding to the quotation function to the user during quotation calculation, thereby improving the user's call experience and further improving the transaction rate of auto insurance business.

[0077] Those skilled in the art will appreciate that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily necessary for implementing the present application. Those skilled in the art will appreciate that the modules in the devices in the implementation scenario can be distributed in the devices of the implementation scenario according to the description of the implementation scenario, or can be changed accordingly and located in one or more devices different from the present implementation scenario. The modules of the above-mentioned implementation scenario can be combined into one module, or can be further split into multiple submodules.

[0078] The above serial numbers of this application are only for description and do not represent the advantages and disadvantages of the implementation scenarios. The above disclosure is only a few specific implementation scenarios of this application, but this application is not limited to them, and any changes that can be thought of by technicians in this field should fall within the scope of protection of this application.

Claims

1. A large model-based auto insurance business question-answering interactive method, characterized in that: include: Obtain the user's auto insurance consultation questions; Inputting the auto insurance consultation question into a preset auto insurance big model to generate a response result corresponding to the auto insurance consultation question; If the reply result is a quotation function for the user's auto insurance demand, the quotation function is sent to the auto insurance quotation system for quotation calculation. After the auto insurance quotation system completes the calculation, the quotation calculation result is written into the cache, and the target quotation script corresponding to the quotation function is returned to the user at the same time; When the user asks whether the quotation is successful in the next round of dialogue, determining whether the quotation calculation result exists in the cache; If the quotation calculation result exists in the cache, the user's quotation inquiry question is replaced with the quotation calculation result and input into the preset auto insurance model to generate quotation details information to reply to the user.

2. The method according to claim 1, characterized in that After inputting the auto insurance consultation question into the preset auto insurance big model to generate a response result corresponding to the auto insurance consultation question, the method further includes: If the reply result is not a quotation function for the user's auto insurance demand, the reply result is directly returned to the user.

3. The method according to claim 1, characterized in that Before returning the target quotation words corresponding to the quotation function to the user, the method further includes: Construct the mapping relationship between different quotation functions and different quotation scripts; The returning the target quotation words corresponding to the quotation function to the user includes: Based on the mapping relationship, the target quotation words corresponding to the quotation function are determined, and the target quotation words are returned to the user, wherein the quotation function includes the quotations for the types of insurance involved in the user's auto insurance demand.

4. The method according to claim 1, characterized in that: After determining whether the quotation calculation result exists in the cache when the user asks whether the quotation is successful in the next round of dialogue, the method further includes: If the quotation calculation result does not exist in the cache, the user's quotation inquiry question is input into the preset auto insurance big model to generate answer information for the quotation inquiry question.

5. The method according to any one of claims 1 to 4, characterized in that: Before inputting the auto insurance consultation question into the preset auto insurance big model to generate a response result corresponding to the auto insurance consultation question, the method further includes: Get the original phone recording; Based on the original telephone recording, generating a conversation training sample; Based on the dialogue training samples, the insurance field big model is trained to obtain the preset auto insurance big model.

6. The method according to claim 5, characterized in that The step of generating a conversation training sample based on the original telephone recording comprises: Performing text recognition on the original telephone recording to obtain a conversation text; The dialogue text is processed using the first preset universal generation large model and the second preset universal generation large model respectively to obtain a dialogue training sample.

7. The method according to claim 6, characterized in that The processing of the dialogue text using the first preset universal generation model and the second preset universal generation model to obtain a dialogue training sample comprises: Using the first preset universal generation model, the dialogue text is subjected to dialogue text role separation, text error correction, data desensitization and data standardization processing to obtain an initial dialogue sample; The second preset universal generation model is used to insert quotation information at key positions of the initial dialogue sample to obtain the dialogue training sample, wherein the quotation information includes: heuristic speech samples, quotation function samples, quotation calculation result samples and quotation details samples.

8. A large model-based auto insurance business question-answering interactive device, characterized in that: include: An acquisition unit, used to acquire the user's auto insurance consultation questions; A generating unit, used for inputting the auto insurance consultation question into a preset auto insurance big model to generate a response result corresponding to the auto insurance consultation question; a calculation unit, configured to send the quotation function to the auto insurance quotation system for quotation calculation if the reply result is a quotation function for the user's auto insurance demand, write the quotation calculation result into a cache after the auto insurance quotation system completes the calculation, and return the target quotation script corresponding to the quotation function to the user; A determination unit, configured to determine whether the quotation calculation result exists in the cache when the user asks whether the quotation is successful in the next round of dialogue; The generating unit is further configured to replace the user's quotation inquiry question with the quotation calculation result if the quotation calculation result exists in the cache, input the result into the preset auto insurance model, generate quotation details information, and reply to the user.

9. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images