Intelligent drilling system and method based on RAG technology, medium, terminal and program product

Abstract

The invention provides an intelligent drilling system and method based on the RAG technology, a medium, a terminal and a program product, and the system comprises a knowledge base construction module which is used for constructing a private domain knowledge base of the insurance industry based on the RAG technology; the exercise mode construction module is used for constructing a progressive exercise mode based on the Bloom teaching target principle, sorting and classifying data corresponding to the constructed exercise mode into a library, and linking the data to a corresponding private domain knowledge base according to the category of the exercise mode; the virtual customer creation module is used for creating a virtual customer in an insurance service scene according to the type of the drilling mode; and the retrieval generation module is used for retrieving the private domain knowledge base based on the RAG technology and according to the dialogue content generated by the virtual customer to generate corresponding prompt content, and inputting the prompt content into the large language model to generate response content corresponding to each category of drilling mode. According to the system, the training efficiency of the insurance agent can be improved, and the training cost is reduced.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an intelligent rehearsal system, method, medium, terminal and program product based on RAG technology. Background Art

[0002] In the current insurance agent training model, insurance agents are required to conduct offline one-on-one drills with their supervisors until they pass the test, so as to achieve the training goals.

[0003] From the perspective of training effectiveness, the uneven qualifications and levels of supervisors, as well as the different standards for passing the course, have affected the passing effect of insurance agents during the actual rehearsal process. From the perspective of training efficiency, supervisors need to spend a fixed amount of time conducting face-to-face simulated interviews with agents. Since a supervisor needs to train multiple insurance agents, this process consumes a lot of supervisors' time and energy; at the same time, insurance agents need to make appointments with supervisors and training locations in advance, which has high communication costs and cannot be conducted anytime, anywhere, making it difficult to rehearse repeatedly. From the perspective of data retention, the rehearsal process is difficult to retain and cannot be used as subsequent data for traceability and analysis.

[0004] Therefore, it is necessary to provide an intelligent rehearsal system, method, medium, terminal and program product based on RAG technology to solve the above-mentioned problems existing in the prior art. Summary of the invention

[0005] In view of the above-mentioned shortcomings of the prior art, the purpose of this application is to provide an intelligent rehearsal system, method, medium, terminal and program product based on RAG technology to solve the technical problems of low efficiency and high cost of the traditional training model for insurance agents.

[0006] To achieve the above-mentioned purpose and other related purposes, the first aspect of the present application provides an intelligent drill system based on RAG technology, the system comprising:

[0007] A knowledge base construction module is used to build a private domain knowledge base for the insurance industry based on RAG technology. The private domain knowledge base includes professional knowledge, sales construction strategies, customer portrait problem solving strategies, and laws and regulations of the insurance industry.

[0008] A drill mode construction module is used to construct a progressive drill mode based on Bloom's teaching goal principle, organize and classify the data corresponding to the constructed drill mode into a library, and link it to the corresponding private domain knowledge base according to the category of the drill mode;

[0009] A virtual customer creation module, used to create a virtual customer in an insurance business scenario according to the category of the exercise mode; the virtual customer is configured to generate dialogue content in the insurance business scenario;

[0010] The retrieval generation module is used to search the private domain knowledge base based on RAG technology and according to the conversation content generated by the virtual customer to generate corresponding prompt content, and input the prompt content into the large language model to generate response content corresponding to each category of drill mode.

[0011] In some embodiments of the first aspect of the present application, the retrieval generation module includes: a context perception unit, used to analyze and understand the problems, needs and intentions of the virtual customer in the current dialogue context; a retrieval unit, used to retrieve relevant knowledge and information in the current dialogue context from the private domain knowledge base based on the problems, needs and intentions of the virtual customer perceived by the context perception unit; an idea and strategy suggestion unit, used to generate ideas and strategy suggestions for students to extract based on the retrieved relevant knowledge and information in the current dialogue context.

[0012] With this design, we can first perceive the real intention behind the questions raised by the virtual customers, and then retrieve relevant knowledge and information from the private domain knowledge base based on the perceived real intention. The ideas and strategic suggestions finally generated can be more in line with the real needs of the virtual customers, providing customers with more professional and personalized services, and also winning recognition from customers for the trainees.

[0013] In some embodiments of the first aspect of the present application, the knowledge base construction module includes: a data acquisition unit, used to collect original data related to various business scenarios of the insurance industry from internal sources and / or external sources; a data processing unit, used to format the acquired original data to convert it into the same file format; a vectorization unit, used to convert the processed text data in the same file format into a vector matrix based on a vector model; a data warehousing unit, used to build an index for the data vectorized based on the vector matrix and write it into a database to form a private domain knowledge base of the insurance industry.

[0014] In some embodiments of the first aspect of the present application, the rehearsal mode includes any one or more of the following: a learning mode, a practice mode, and a test mode; in the learning mode, the retrieval generation module generates standard scripts for students to learn in response to the conversation content generated by the virtual customer; the learning mode is switched to the practice mode, and the practice mode allows students to practice conversations with the virtual customer in various insurance business scenarios; the practice mode is switched to the test mode, and the test mode is used to test students' conversations with the virtual customer in various insurance business scenarios.

[0015] In some embodiments of the first aspect of the present application, the system is further configured with a student interactive interface, and the student interactive interface is configured to display an interactive conversation between the student and the virtual customer.

[0016] In some embodiments of the first aspect of the present application, the system also includes a virtual scene construction module for constructing various insurance business scenarios for dialogues between virtual customers and trainees. The virtual scene construction module includes at least a VR generation device for simulating scenarios based on the age, gender, and occupation of the virtual customer, and outputting simulated practical scenarios.

[0017] To achieve the above-mentioned purpose and other related purposes, the second aspect of the present application provides an intelligent drill method based on RAG technology, the method comprising:

[0018] Build a private domain knowledge base for the insurance industry based on RAG technology, which includes insurance industry expertise, sales development strategies, customer profiling problem solving strategies, and insurance industry laws and regulations;

[0019] Construct a progressive drill model based on Bloom's teaching goal principle, organize and classify the data corresponding to the constructed drill model into a library, and link it to the corresponding private domain knowledge base according to the category of the drill model;

[0020] Creating a virtual customer in an insurance business scenario according to the category of the drill mode; the virtual customer is configured to generate dialogue content in the insurance business scenario;

[0021] Based on the RAG technology and according to the conversation content generated by the virtual customer, the private domain knowledge base is searched to generate corresponding prompt content, and the prompt content is input into the large language model to generate response content corresponding to each category of drill mode.

[0022] To achieve the above-mentioned purpose and other related purposes, the third aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, and the computer program implements the method when executed by a processor.

[0023] To achieve the above-mentioned purpose and other related purposes, the fourth aspect of the present application provides a computer program product, which includes a computer program code. When the computer program code runs on a computer, the computer implements the method.

[0024] To achieve the above-mentioned purpose and other related purposes, the fifth aspect of the present application provides an electronic terminal, including a memory, a processor and a computer program stored in the memory; the processor executes the computer program to implement the method.

[0025] As described above, the intelligent rehearsal system, method, medium, terminal and program product based on RAG technology of the present application have the following beneficial effects:

[0026] By using RAG technology to build a private domain knowledge base for the insurance industry, and based on Bloom's teaching goal principle, a drill model is constructed. The data corresponding to the constructed drill model is sorted and classified into a library, and linked to the corresponding private domain knowledge base according to the category of the drill model. At the same time, multiple virtual customers in the insurance business scenario are created according to the classification category of the drill model. Then, based on RAG technology and according to the dialogue content generated by the virtual customer, the private domain knowledge base is retrieved to generate corresponding prompt content, and the prompt content is input into the large language model to generate the response content corresponding to each category of drill mode. The system can not only improve the training efficiency and quality of insurance agents, but also reduce the cost of training, which is helpful to improve the professional quality and work ability of insurance agents, and realize the comprehensive training of insurance agents from knowledge memory to creative performance. At the same time, by building a private domain knowledge base, the system can provide more comprehensive training content to ensure that insurance agents grasp the latest market dynamics and industry trends and enhance their professional quality and competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 Shown is a framework diagram of an intelligent rehearsal system based on RAG technology in one embodiment of the present application.

[0028] Figure 2 Shown is a schematic diagram of the principles of an intelligent rehearsal system based on RAG technology in one embodiment of the present application.

[0029] Figure 3 Shown is a schematic diagram of an application scenario of an intelligent rehearsal system based on RAG technology in one embodiment of the present application.

[0030] Figure 4 Shown is a flowchart of an intelligent rehearsal method based on RAG technology in one embodiment of the present application.

[0031] Figure 5 Shown is a schematic diagram of the structure of an electronic terminal in one embodiment of the present application. DETAILED DESCRIPTION

[0032] The following describes the embodiments of the present application through specific examples, and those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0033] In the embodiments of the present application, words such as "first" and "second" are used to distinguish the same or similar items with substantially the same functions and effects. For example, the first XX and the second XX are only used to distinguish different XXs, and do not limit their order. Those skilled in the art can understand that words such as "first" and "second" do not limit the quantity and execution order, and words such as "first" and "second" do not necessarily limit them to be different.

[0034] It should be noted that in the embodiments of the present application, words such as "exemplary" or "for example" represent examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in the present application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0035] In the embodiments of the present application, "at least one" refers to one or more, and "plurality" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can represent: a, b, c, ab, ac, bc or abc, where a, b, c can be single or multiple.

[0036] Before further describing the present invention in detail, the nouns and terms involved in the embodiments of the present invention are explained. The nouns and terms involved in the embodiments of the present invention are applicable to the following interpretations:

[0037] <1> RAG (Retrieval-Augmented Generation) technology: is an artificial intelligence technology that combines information retrieval technology with language generation models. This technology retrieves relevant information from an external knowledge base and inputs it as a prompt to a large language model (LLMs) to enhance the model's ability to handle knowledge-intensive tasks such as question answering, text summarization, and content generation.

[0038] <2> Large Language Model (LLM): An AI model that relies on large amounts of text data for training and can perform a variety of tasks, including text summarization, translation, and sentiment analysis. These models typically contain billions of parameters and are able to capture complex patterns in language data. Most LLMs are based on deep learning architectures, such as Transformers, which have demonstrated excellent performance in a variety of natural language processing tasks.

[0039] To facilitate understanding of the embodiments of the present application, first Figure 1 Detailed description. Figure 1 The following is a schematic diagram of a framework of an intelligent drill system based on RAG technology in an embodiment of the present invention. The intelligent drill system 100 based on RAG technology in this embodiment includes:

[0040] A knowledge base construction module 101 is used to construct a private domain knowledge base for the insurance industry based on RAG technology, wherein the private domain knowledge base includes professional knowledge of the insurance industry, sales construction strategies, customer portrait problem solving strategies, solution strategies for preset problems, and laws and regulations of the insurance industry;

[0041] The drill mode construction module 102 is used to construct a progressive drill mode based on Bloom's teaching goal principle, organize and classify the data corresponding to the constructed drill mode into a library, and link to the corresponding private domain knowledge base according to the category of the drill mode;

[0042] A virtual customer creation module 103 is used to create multiple virtual customers in insurance business scenarios according to the categories of the drill mode; the virtual customers are configured to generate dialogue content in the insurance business scenario;

[0043] The retrieval generation module 104 is used to search the private domain knowledge base based on the RAG technology and according to the conversation content generated by the virtual customer to generate corresponding prompt content, and input the prompt content into the large language model to generate response content corresponding to each category of drill mode.

[0044] By using RAG technology to build a private domain knowledge base containing insurance industry professional knowledge, sales construction strategies, customer portrait problem solving strategies and insurance industry laws and regulations, and based on Bloom's teaching goal principle, a drill mode is constructed, the data corresponding to the constructed drill mode is sorted and classified into a library, and the drill mode is classified into learning mode, practice mode and test mode, and linked to the corresponding private domain knowledge base according to the classification category, and at the same time, virtual customers in insurance business scenarios are created according to the classification category of the drill mode. The virtual customers are configured to generate dialogue content in various insurance business scenarios, and then based on RAG technology and according to the dialogue content generated by the virtual customer, the private domain knowledge base is retrieved to generate corresponding prompt content, and the prompt content is input into the large language model to generate the response content corresponding to each category of drill mode. The system can not only improve the training efficiency and quality of insurance agents, but also reduce the cost of training, which is helpful to improve the professional quality and work ability of insurance agents, and realize the comprehensive training of insurance agents from knowledge memory to creative play. At the same time, by building a private domain knowledge base, the system can provide more comprehensive training content to ensure that insurance agents grasp the latest market dynamics and industry trends and enhance their professional quality and competitiveness.

[0045] It should be understood that Bloom's teaching objectives principle is divided into six stages: knowledge, understanding, application, analysis, evaluation and creation, which are also applicable in insurance agent training. The specific examples of applying Bloom's teaching objectives principle to insurance agent training are as follows:

[0046] (1) Knowledge: This stage requires the insurance agent to be able to recall or recognize information. For example, when training on a new product, the insurance agent needs to remember the basic functions and features of the product.

[0047] (2) Understanding: This stage emphasizes the insurance agent's initial understanding of the information and the ability to restate the concept or give a simple explanation in his or her own words. For example, when understanding the company's policy, the insurance agent should be able to explain the main purpose of the policy and its impact on his or her work.

[0048] (3) Application: At this stage, insurance agents need to apply what they have learned to actual work. For example, after learning sales skills, insurance agents should be able to use these skills in the actual sales process to improve sales performance.

[0049] (4) Analysis: This stage requires the agent to analyze information and identify relationships between materials. In agent training, this may involve case studies, where the agent analyzes the best course of action in different situations.

[0050] (5) Synthesis: In this stage, insurance agents are encouraged to integrate multiple information sources or concepts into a new whole, such as developing a marketing plan that combines different techniques and resources to achieve specific marketing goals.

[0051] (6) Evaluation: This is the highest level of learning, requiring insurance agents to evaluate and critically think about information or situations. In insurance agent training, insurance agents’ evaluation abilities can be trained through simulated decision-making scenarios, enabling them to judge and choose the effectiveness of various strategies.

[0052] It should be noted that the data contained in the private domain knowledge base is simulated customer data based on real data structures, and these data are desensitized to ensure customer information security and data compliance. The private domain knowledge base also includes historical successful conversation records, answer strategies for preset questions, etc.

[0053] In some embodiments of the present application, the rehearsal mode includes any one or more of the following: a learning mode, a practice mode, and a test mode; in the learning mode, the retrieval generation module generates standard scripts for students to learn in response to the conversation content generated by the virtual customer; switching from the learning mode to the practice mode, the practice mode allows students to practice conversations with the virtual customer in various insurance business scenarios, such as sales, recruitment, etc.; switching from the practice mode to the test mode, the test mode is used to test students' conversations with the virtual customer in various insurance business scenarios.

[0054] In some specific embodiments, the drill mode creation module 102 creates three drill modes based on Bloom's teaching objective principle, namely, learning mode, practice mode and test mode. It should be noted that the learning mode, practice mode and test mode can be applied in stages, or in a progressive drill mode of learning mode-practice mode-test mode. This progressive drill mode can enable insurance agents to progress from recitation, understanding and internalization to creative performance. Specifically:

[0055] (1) Learning mode: Trainees can learn quickly under the guidance of mechanical knowledge. When talking to virtual customers, when the virtual customers raise certain objections, trainees can repeat and restore the content of the response dialogue according to the system prompts, thereby achieving the effect of strengthening learning. In this drill mode, all dialogues and response dialogues are not pre-set scripts. The purpose is to allow agents to learn the knowledge and dialogues in the private domain knowledge base from simulated dialogue interviews, so as to achieve the purpose of learning dialogues.

[0056] (2) Practice mode: In this practice mode, a virtual customer has a conversation with the trainee in a specific situation. In this process, the drill system provides guidance on solution ideas. During the conversation, the questions raised by the virtual customer are monitored, and solutions to each specific objection are given based on the customer's situation. Problems that arise during the simulated interview are promptly guided and corrected, thereby assisting the trainee and achieving the purpose of practicing conversation skills.

[0057] (3) Test mode: Trainees can freely conduct conversations with specific customers or specific topics to test the training content. In this test mode, trainees can freely guide the conversations, thereby obtaining different results such as virtual customers with judgment ability after RAG retrieval training agreeing to the transaction or rejecting the transaction, thus achieving the purpose of passing the test.

[0058] Through the above-mentioned progressive and systematic training method of learning mode-practice mode-testing mode, students can comprehensively improve their professional quality and work ability, so as to better cope with market challenges and customer needs; at the same time, it also greatly improves the training efficiency of students and reduces training costs. In addition, the rehearsal process between students and virtual customers can be retained and used as subsequent data for traceability and analysis.

[0059] In some embodiments of the present application, Figure 2 As shown, the retrieval generation module 104 includes: a context perception unit 1041, which is used to analyze and understand the problems, needs and intentions of the virtual customer in the current dialogue context; a retrieval unit 1042, which is used to retrieve relevant knowledge and information in the current dialogue context from the private domain knowledge base according to the problems, needs and intentions of the virtual customer perceived by the context perception unit; and a train of thought and strategy suggestion unit 1043, which is used to generate train of thought and strategy suggestions for students to extract based on the relevant knowledge and information in the current dialogue context obtained by retrieval. By designing the retrieval generation module 104 including the context perception unit 1041, the retrieval unit 1042 and the train of thought and strategy suggestion unit 1043, it is realized that in the practice mode, the real intention behind the questions raised by the virtual customer can be first perceived, and then the relevant knowledge and information can be retrieved from the private domain knowledge base according to the perceived real intention, so that the train of thought and strategy suggestions generated in the end can better meet the real needs of the virtual customer and provide customers with more professional and personalized services.

[0060] In some embodiments of the present application, Figure 2As shown, the knowledge base construction module 101 includes: a data acquisition unit 1011, which is used to collect raw data related to various business scenarios of the insurance industry from internal sources and / or external sources; a data processing unit 1012, which is used to format the acquired raw data to convert it into the same file format; a vectorization unit 1013, which is used to convert the processed text data in the same file format into a vector matrix based on a vector model; a data storage unit 1014, which is used to construct an index for the data after vectorization based on the vector matrix and write it into a database to form the private domain knowledge base of the insurance industry. By designing the knowledge base construction module 101 including the data acquisition unit 1011, the data processing unit 1012, the vectorization unit 1013 and the data storage unit 1014, a private domain knowledge base containing internal information and / or knowledge information known to the insurance industry can be formed, thereby providing students with a richer and more accurate database.

[0061] It should be understood that the vector (Embedding) model is a technology that maps complex, high-dimensional or sparse data (such as text, images, classification features, etc.) into a low-dimensional, dense vector space. This vector representation retains the semantic or structural information of the original data and is widely used in machine learning and deep learning to process tasks such as text, images, and user behavior.

[0062] In some embodiments of the present application, the system is further configured with a student interactive interface, and the student interactive interface is configured to display an interactive conversation between the student and the virtual customer. By designing the student interactive interface, a conversation interaction between the student and the virtual customer is achieved.

[0063] In some embodiments of the present application, the system also includes a virtual scene construction module for constructing various insurance business scenarios for dialogues between virtual customers and trainees, and the virtual scene construction module includes at least a VR generation device, which is used to simulate scenarios according to the age, gender, and occupation of the virtual customer, and output simulated practical scenarios. Through the VR generation device, it is possible to simulate a real working scenario in which a virtual user and a trainee are conducting an insurance business dialogue. This allows trainees to truly engage in this dialogue training. It should be noted that the virtual scene construction module can also use other implementation methods to simulate real business scenarios, which are not specifically limited here.

[0064] like Figure 3As shown, the intelligent rehearsal system based on RAG technology of the embodiment of the present application is exemplified by a specific insurance business scenario. First, a specific business insurance scenario of a dialogue between a virtual customer and a trainee is constructed. In this specific business scenario, the virtual customer and the trainee have a dialogue and communication. This training method that simulates a real work scenario is intended to enable trainees to improve their ability to deal with various complex situations in a safe and stress-free environment. The rehearsal system not only plays the other party in the dialogue, but also assumes the role of monitoring, analysis and guidance to ensure that trainees can grow in every interaction. For example, a business scenario description is as follows: Ms. Li, a 28-year-old young woman, newly married, with a stable career, is relatively unfamiliar with the concept of personal pension insurance, and believes that retirement is a very distant thing, and there is no need to consider insurance in this regard at this stage.

[0065] In the learning mode, student A starts the conversation first: Hello, Ms. Li, I heard that you just got married, congratulations; the virtual customer Ms. Li responds: Oh, Xiao A, you are really well-informed, but I don’t want to hear about insurance, please don’t recommend insurance to me. In this learning mode, since the drill system integrates RAG technology, RAG technology gives the drill system the ability to deeply understand and respond immediately. Therefore, when students encounter challenges when discussing professional content such as the above dialogue scenarios and pension insurance clauses, RAG technology will analyze customers’ questions through advanced natural language processing technology and accurately capture the core of the problem. Subsequently, the technology quickly searches the private domain knowledge base for the most matching clause explanation or related case. Once the most relevant information is found, the retrieval generation module generates an accurate and easy-to-understand answer draft for students to refer to or directly adopt, that is, the standard speech: Student A: Ms. Li, you misunderstood. Last time I heard that you just got married, I thought that you might need new family protection needs, so I chatted with you. It was completely in the mood of being a friend to help you consider it. The process not only accelerates trainees’ understanding of complex concepts, but also teaches them how to convey information in a customer-friendly manner, enhancing their communication skills.

[0066] In practice mode, the drill system does not directly provide standard words, but focuses on guiding students on how to think and find solutions and ideas to solve customer problems, thereby improving their professional quality and adaptability. For example, student A first starts the conversation: Hello, Ms. Li, I heard that you just got married, congratulations; the virtual customer Ms. Li responds: Oh, Xiao A, you are really well-informed, but I don’t want to hear about insurance, so don’t recommend insurance to me. Faced with this business scenario, the context perception unit 1041 enables the drill system to deeply understand the content of the conversation, including the customer’s problems, needs, and the emotions and intentions behind them. Through natural language processing technology, the drill system can identify key information and potential problem points in the conversation. The retrieval unit 1042 retrieves knowledge and information related to the current dialogue context from the private domain knowledge base. This knowledge may include insurance terms, product features, market trends, previous successful or failed cases, etc., so that it can quickly find the knowledge that best matches the needs of the virtual customer and provide background information and support for students. Based on the retrieved relevant knowledge and the understanding of the current dialogue situation, the training system generates solution ideas and strategy suggestions in the thinking and strategy suggestion unit 1043. These suggestions guide trainees on how to think about problems from different angles and how to formulate solutions based on the needs of virtual customers and the characteristics of insurance products. For example, the answer idea generated by the thinking and strategy suggestion unit 1043 is: Xiao A should get closer to Ms. Li at this time, and not directly promote sales. He can talk about Ms. Li’s recent marriage life, and then naturally transition to the protection role of insurance, especially long-term planning such as pensions, emphasizing that this is not promotion, but caring about her financial security from the perspective of a friend. In the practice mode, the training system focuses on guiding trainees to form independent thinking and problem-solving abilities, rather than simply providing standard speech. By analyzing the content of the dialogue in real time, retrieving relevant knowledge, providing solution ideas and strategy suggestions, as well as instant feedback and improvement suggestions, it helps trainees to continuously learn and grow in actual operations and eventually become independent insurance experts. In addition, the training system will also conduct a quality assessment of the trainees’ answers, and use the feedback mechanism of RAG technology to immediately point out the shortcomings in the answers, such as unclear logic, missing information, or improper expression. Based on these specific feedbacks, the drill system will provide targeted improvement suggestions to guide students on how to adjust their strategies and optimize their expressions to achieve a higher level of professionalism.

[0067] In the test mode, trainees are placed in an environment that highly simulates real work scenarios. They think independently and respond to various challenges without any preset words or thought prompts. For example, trainee A: Hello, Ms. Li, I heard that you just got married, congratulations; the virtual customer Ms. Li responded: Oh, Xiao A, you are really well-informed, but I don’t want to hear about insurance, don’t recommend insurance to me. In the test mode, the trainee faces Ms. Li completely independently and answers Ms. Li’s questions. The setting of the test mode is to test whether the trainees have truly mastered the knowledge they have learned and their ability to apply this knowledge in actual work. In addition, in this process, trainees may face a variety of different types of customers, each of which has its own unique needs and problems, which requires trainees to be able to flexibly apply the sales skills and communication strategies they have learned to effectively interact with different types of customers; at the same time, trainees may also need to conduct in-depth discussions around specific topics or issues, which helps to examine their understanding and mastery of professional knowledge and fully meet the strict requirements of training clearance. In order to ensure the fairness and effectiveness of the test, the test mode usually includes a series of evaluation criteria and indicators. These criteria may include the trainee's communication skills, problem-solving skills, emotional processing skills, etc. Through these comprehensive assessments, the trainee's actual ability and level can be judged more accurately.

[0068] It should be noted that the insurance business scenarios and dialogue contents constructed in the learning mode, practice mode and test mode may be different, and are only shown here for illustration.

[0069] In some embodiments of the present application, the large language model is a GPT large model.

[0070] It should be understood that the division of modules in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present application may be integrated into a processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0071] The embodiment of the present application also provides an intelligent drill method based on RAG technology. Figure 4 The method flow chart is shown. The intelligent drill method based on RAG technology in this embodiment includes the following steps:

[0072] Step S41: Build a private domain knowledge base for the insurance industry based on RAG technology, where the private domain knowledge base includes professional knowledge, sales development strategies, customer portrait problem solving strategies, and laws and regulations of the insurance industry.

[0073] Step S42: construct a progressive drill model based on Bloom's teaching objective principle, organize and classify the data corresponding to the constructed drill model into a library, and link it to the corresponding private domain knowledge base according to the category of the drill model.

[0074] Step S43: creating a virtual customer in an insurance business scenario according to the category of the drill mode; the virtual customer is configured to generate dialogue content in the insurance business scenario.

[0075] Step S44: Based on the RAG technology and according to the conversation content generated by the virtual customer, the private domain knowledge base is searched to generate corresponding prompt content, and the prompt content is input into the large language model to generate response content corresponding to each category of drill mode.

[0076] It should be noted that the intelligent rehearsal method based on RAG technology in this embodiment can realize the functions of the intelligent rehearsal system based on RAG technology mentioned above, which will not be described in detail here.

[0077] Figure 5 is a schematic block diagram of an electronic terminal provided in an embodiment of the present application. Figure 5 As shown, the electronic terminal 500 includes: at least one processor 501, a memory 502, at least one network interface 503 and a user interface 505. The various components in the electronic terminal 500 are coupled together through a bus system 504. It can be understood that the bus system 504 is used to realize the connection and communication between these components. In addition to the data bus, the bus system 504 also includes a power bus, a control bus and a status signal bus. However, for the sake of clarity, the bus system 504 is not shown in FIG. Figure 5 In the specification, various buses are labeled as bus systems.

[0078] The user interface 505 may include a display, a keyboard, a mouse, a trackball, a click gun, keys, buttons, a touch pad or a touch screen.

[0079] It is understood that the memory 502 can be a volatile memory or a non-volatile memory, and can also include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM). The memory described in the embodiments of the present invention is intended to include but is not limited to these and any other suitable categories of memory.

[0080] The memory 502 in the embodiment of the present invention is used to store various categories of data to support the operation of the electronic terminal 500. Examples of these data include: any executable program for operating on the electronic terminal 500, such as an operating system 5021 and an application 5022; the operating system 5021 includes various system programs, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks. The application 5022 may include various applications, such as a media player (Media Player), a browser (Browser), etc., for implementing various application services. The intelligent rehearsal method based on RAG technology provided in the embodiment of the present invention may be included in the application 5022.

[0081] The method disclosed in the above embodiment of the present invention can be applied to the processor 501, or implemented by the processor 501. The processor 501 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit in the processor 501 or the instruction in the form of software. The above processor 501 may be a general processor, a digital signal processor (DSP, Digital Signal Processor), or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. The processor 501 can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiment of the present invention. The general processor 501 may be a microprocessor or any conventional processor, etc. In combination with the steps of the accessory optimization method provided in the embodiment of the present invention, it can be directly embodied as a hardware decoding processor to execute, or it can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium, which is located in a memory, and the processor reads the information in the memory and completes the steps of the above method in combination with its hardware.

[0082] In an exemplary embodiment, the electronic terminal 500 may be implemented by one or more application specific integrated circuits (ASIC), DSP, programmable logic device (PLD), complex programmable logic device (CPLD) to execute the aforementioned method.

[0083] According to the intelligent rehearsal method based on RAG technology provided in the embodiment of the present application, the present application also provides a computer program product, which includes: a computer program code, when the computer program code is run on a computer, the computer executes Figure 4 A method according to any one of the embodiments shown.

[0084] According to the intelligent rehearsal method based on RAG technology provided in the embodiment of the present application, the present application also provides a computer-readable storage medium, which stores a program code. When the program code is run on a computer, the computer executes Figure 4 A method according to any one of the embodiments shown.

[0085] The terms "component", "module", "system", etc. used in this specification are used to represent computer-related entities, hardware, firmware, a combination of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program and / or a computer. By way of illustration, both applications running on a computing device and a computing device can be components. One or more components may reside in a process and / or an execution thread, and a component may be located on a computer and / or distributed between two or more computers. In addition, these components may be executed from various computer-readable media having various data structures stored thereon. Components may, for example, communicate through local and / or remote processes according to signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system and / or a network, such as the Internet interacting with other systems through signals).

[0086] Those of ordinary skill in the art will appreciate that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0087] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0088] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0089] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] In addition, each functional unit in each embodiment of the present application 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.

[0091] In the above embodiments, the functions of each functional unit can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions (programs). When loading and executing computer program instructions (programs) on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media integrations. Available media may be magnetic media (e.g., floppy disks, hard disks, tapes), optical media (e.g., high-density digital video discs (DVDs), or semiconductor media (e.g., solid state disks (SSDs), etc.).

[0092] If the function is implemented in the form of 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 application, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program codes.

[0093] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0094] In summary, since traditional agent training methods often rely on face-to-face teaching and simulated dialogues, this method is not only costly but also inefficient. With the development of artificial intelligence technology, it has become possible to use RAG technology for intelligent analysis and guidance, so that artificial intelligence can understand and respond to user needs more accurately, especially in the intelligent human-machine training scenarios in the insurance industry. Therefore, the present invention provides an intelligent training system, method, medium, terminal and program product based on RAG technology. By building a powerful private domain knowledge base containing professional knowledge, product details and industry rules of the insurance industry based on RAG technology, when virtual customers generate dialogue content, the training system can provide the private domain knowledge base with intelligent training information. In real time, the relevant information can be retrieved from the chat room, and combined with the specific needs of the virtual customer, so as to generate richer and more accurate response content; in addition, RAG technology can also optimize the conversation experience according to the user's portrait and chat history. By analyzing the user's behavior and reviewing the historical conversations, the rehearsal system can better understand the user's intentions and preferences, thereby providing services that are more in line with the user's needs. For example, if the rehearsal system detects that the user is interested in a certain type of insurance product, it can retrieve the relevant product information through RAG technology and actively provide this information in the conversation to enhance the user experience; at the same time, the rehearsal system can handle complex conversation scenarios more intelligently and provide users with more professional and personalized services. Therefore, this application effectively overcomes the various shortcomings of the prior art and has a high industrial utilization value.

[0095] The above embodiments are merely illustrative of the principles and effects of the present application and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. An intelligent drill system based on RAG technology, characterized in that: The system comprises: A knowledge base construction module is used to build a private domain knowledge base for the insurance industry based on RAG technology. The private domain knowledge base includes professional knowledge, sales construction strategies, customer portrait problem solving strategies, and laws and regulations of the insurance industry. A drill mode construction module is used to construct a progressive drill mode based on Bloom's teaching goal principle, organize and classify the data corresponding to the constructed drill mode into a library, and link it to the corresponding private domain knowledge base according to the category of the drill mode; A virtual customer creation module, used to create a virtual customer in an insurance business scenario according to the category of the drill mode; The virtual customer is configured to generate conversation content in an insurance business scenario; The retrieval generation module is used to search the private domain knowledge base based on RAG technology and according to the conversation content generated by the virtual customer to generate corresponding prompt content, and input the prompt content into the large language model to generate response content corresponding to each category of drill mode.

2. The intelligent drill system based on RAG technology according to claim 1 is characterized in that: The retrieval generation module comprises: A context awareness unit, which is used to analyze and understand the virtual customer's questions, needs, and intentions in the current conversation context; A retrieval unit, configured to retrieve relevant knowledge and information in the current conversation context from the private domain knowledge base according to the questions, needs and intentions of the virtual customer perceived by the context perception unit; The idea and strategy suggestion unit is used to generate ideas and strategy suggestions for students to extract based on the relevant knowledge and information retrieved in the current dialogue context.

3. The intelligent drill system based on RAG technology according to claim 1 is characterized in that: The knowledge base building module includes: A data acquisition unit for collecting raw data related to various business scenarios of the insurance industry from internal sources and / or external sources; A data processing unit, used for formatting the acquired raw data to convert it into the same file format; A vectorization unit, used for converting the processed text data in the same file format into a vector matrix based on a vector model; The data storage unit is used to construct an index for the data vectorized based on the vector matrix and write it into a database to form a private domain knowledge base of the insurance industry.

4. The intelligent drill system based on RAG technology according to claim 1, characterized in that: The practice mode includes any one or more of the following: learning mode, practice mode and test mode; in the learning mode, the retrieval generation module generates standard dialogues for students to learn in response to the dialogue content generated by the virtual customer; switching from the learning mode to the practice mode, the practice mode allows students to practice dialogues with the virtual customer in various insurance business scenarios; Switch from the practice mode to the test mode, where the test mode is used to test the trainee's dialogue with the virtual customer in various insurance business scenarios.

5. The intelligent drill system based on RAG technology according to claim 1, characterized in that: The system is also configured with a student interactive interface, and the student interactive interface is configured to display an interactive dialogue between the student and the virtual customer.

6. The intelligent drill system based on RAG technology according to claim 1, characterized in that: The system also includes a virtual scenario construction module for constructing various insurance business scenarios for dialogues between virtual customers and trainees. The virtual scenario construction module at least includes a VR generation device for simulating scenarios based on the age, gender, and occupation of the virtual customers, and outputting simulated practical scenarios.

7. An intelligent drill method based on RAG technology, characterized in that: The method comprises: Build a private domain knowledge base for the insurance industry based on RAG technology, which includes insurance industry expertise, sales development strategies, customer profiling problem solving strategies, and insurance industry laws and regulations; Construct a progressive drill model based on Bloom's teaching goal principle, organize and classify the data corresponding to the constructed drill model into a library, and link it to the corresponding private domain knowledge base according to the category of the drill model; Creating a virtual customer in an insurance business scenario according to the category of the drill mode; the virtual customer is configured to generate dialogue content in the insurance business scenario; Based on the RAG technology and according to the conversation content generated by the virtual customer, the private domain knowledge base is searched to generate corresponding prompt content, and the prompt content is input into the large language model to generate response content corresponding to each category of drill mode.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program implements the method of claim 7 when executed by a processor.

9. A computer program product, characterized in that The computer program product comprises a computer program code, and when the computer program code is run on a computer, the computer is caused to implement the method as claimed in claim 7.

10. An electronic terminal comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the method in claim 7.

Images