Data processing method and device based on artificial intelligence, computer equipment and medium

Abstract

The invention belongs to the technical field of artificial intelligence, and relates to a data processing method and device based on artificial intelligence, computer equipment and a storage medium, and the method comprises the steps: extracting product name information from a received product drilling request triggered by an agent in a specified page, retrieving specified product content corresponding to the product name information from a product database; generating initial input based on the specified product content and sending the initial input to the target large language model for processing; after preparation completion information returned by the target large language model is received, confirmation information is sent to the seat; executing product drilling processing corresponding to the specified product content on the seat based on the target large language model; and obtaining all target verbal skills input by the seat in the product drilling process, and generating a drilling score of the seat based on the target verbal skills. In addition, the drill score may be stored in a blockchain. Based on the use of the target large language model, the processing efficiency of product drilling is improved, and the cost of product drilling is reduced.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence development technology and financial technology, and in particular to artificial intelligence-based data processing methods, devices, computer equipment and storage media. Background Art

[0002] In the insurance industry, telemarketing, as a marketing method with a long history, has long played the role of a bridge between insurance companies and a large number of potential customers. Its advantage is that it can efficiently and widely reach unfamiliar user groups, and compared with other marketing methods, the cost of telemarketing is relatively low, making it a marketing channel favored by many insurance companies. However, with the increasing prosperity of the insurance market and the continuous enrichment of product lines, the limitations of the telemarketing model in coping with the rapid promotion of new products have gradually emerged.

[0003] When new insurance products are launched, traditional agent channels usually adopt a hierarchical information transmission method, that is, new product information is transmitted from the upper management level to the front-line telemarketing seats step by step. In this process, not only is the information transmission chain long and time-consuming, but it is also easily interfered by human factors, resulting in information distortion or omission. More importantly, telemarketers need to spend a lot of time to understand and familiarize themselves with the insurance liability, selling points and corresponding sales techniques of new products, which greatly slows down the promotion of new products in the market.

[0004] At present, product training for telemarketing agents still relies on the traditional lecture-based teaching model. In this model, supervisors or trainers, as knowledge transmitters, impart product knowledge and sales skills to telemarketing agents through oral explanations, PPT presentations, etc. However, this one-way indoctrination teaching method has the following problems: First, the training efficiency is low, and telemarketing agents passively accept information, lack of interaction and practice opportunities, and it is difficult to deeply understand and internalize the knowledge they have learned; second, the training cost is high, including time cost, labor cost and material cost, etc., especially for large-scale training activities, these costs cannot be ignored. Summary of the invention

[0005] The purpose of the embodiments of the present application is to propose a data processing method, device, computer equipment and storage medium based on artificial intelligence to solve the technical problems of low training efficiency and high training cost in the existing product drill training method for telemarketing agents.

[0006] In order to solve the above technical problems, the embodiment of the present application provides a data processing method based on artificial intelligence, which adopts the following technical solutions:

[0007] Determine whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries product name information of the specified product;

[0008] If yes, extract the product name information from the product rehearsal request, and retrieve the specified product content corresponding to the product name information from a preset product database;

[0009] Generate a corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing;

[0010] After receiving the preparation completion information returned by the target large language model, sending a confirmation message to the agent;

[0011] Based on the target large language model, the agent performs a product rehearsal process corresponding to the specified product content, obtains a specified speech by encapsulating the received speech input by the agent, and sends the specified speech to the target large language model;

[0012] Generate a designated question corresponding to the designated speech based on the target large language model, display the designated question on the designated page, receive an answer corresponding to the designated question input by the agent, and repeatedly perform the steps of product rehearsal processing until a preset rehearsal end condition is met, and end the current dialogue session with the target large language model;

[0013] All target words input by the agent during the product rehearsal are obtained, and a rehearsal score of the agent is generated based on the target words.

[0014] Furthermore, the step of generating the agent's practice score based on the target speech specifically includes:

[0015] Matching the target speech with the sentences in the specified product content to obtain corresponding matching results;

[0016] Determining the number of target sentences related to the designated product based on the matching results;

[0017] Obtain the total number of sentences corresponding to the specified product content;

[0018] Calculating the number of sentences and the total number of sentences based on a preset calculation formula to obtain a corresponding calculation result;

[0019] The calculation result is used as the training score of the agent.

[0020] Furthermore, after the step of determining the number of target sentences related to the designated product based on the matching results, the method further includes:

[0021] Perform sentence elimination processing on the specified product content based on the target sentence to obtain corresponding unmentioned sentences;

[0022] Get the preset drill result template;

[0023] Filling the drill result template based on the drill score, the target sentence, and the unmentioned sentence to obtain a corresponding target drill result;

[0024] The target drill result is displayed in the designated page.

[0025] Furthermore, the step of encapsulating the received speech input by the agent to obtain the specified speech, and sending the specified speech to the target large language model specifically includes:

[0026] Receiving the speech input by the agent;

[0027] Get the preset agent script template;

[0028] Encapsulating the speech based on the agent speech template to obtain the corresponding specified speech;

[0029] The specified speech is sent to the target large language model.

[0030] Furthermore, after the step of generating the training score of the agent based on the target speech, the step further includes:

[0031] Collect relevant data of the agent during the product drill;

[0032] Analyze the relevant data to obtain corresponding data analysis results;

[0033] The target large language model is optimized based on the data analysis result.

[0034] Furthermore, the relevant data at least includes the target speech, model feedback data and the drill score; the step of analyzing the relevant data to obtain the corresponding data analysis results specifically includes:

[0035] Analyze the target speech to obtain a corresponding first analysis result;

[0036] Analyzing the model feedback data to obtain a corresponding second analysis result;

[0037] Analyzing the drill score to obtain a corresponding third analysis result;

[0038] The data analysis result is generated based on the first analysis result, the second analysis result, and the third analysis result.

[0039] Furthermore, before the step of retrieving the designated product content corresponding to the product name information from the preset product database, the method further includes:

[0040] Get detailed information on various products;

[0041] Call the preset initial database;

[0042] Storing the detailed information in the initial database to obtain a corresponding first database;

[0043] Performing search optimization processing on the first database to obtain a corresponding second database;

[0044] The second database is used as the product database.

[0045] In order to solve the above technical problems, the embodiment of the present application further provides a data processing device based on artificial intelligence, which adopts the following technical solution:

[0046] A judgment module, used to judge whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries product name information of a specified product;

[0047] A retrieval module, for extracting the product name information from the product rehearsal request, and retrieving the specified product content corresponding to the product name information from a preset product database;

[0048] A first sending module, configured to generate a corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing;

[0049] A second sending module is used to send a confirmation message to the agent after receiving the preparation completion information returned by the target large language model;

[0050] A first processing module is used to perform a product rehearsal process corresponding to the specified product content on the agent based on the target large language model, obtain a specified speech by encapsulating the received speech input by the agent, and send the specified speech to the target large language model;

[0051] A second processing module is used to generate a specified question corresponding to the specified speech based on the target large language model, display the specified question on the specified page, receive an answer corresponding to the specified question input by the agent, and repeatedly perform product rehearsal processing steps until a preset rehearsal end condition is met, and end the current dialogue session with the target large language model;

[0052] A generation module is used to obtain all target words input by the agent during the product rehearsal process, and generate a rehearsal score for the agent based on the target words.

[0053] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0054] Determine whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries product name information of the specified product;

[0055] If yes, extract the product name information from the product rehearsal request, and retrieve the specified product content corresponding to the product name information from a preset product database;

[0056] Generate a corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing;

[0057] After receiving the preparation completion information returned by the target large language model, sending a confirmation message to the agent;

[0058] Based on the target large language model, the agent performs a product rehearsal process corresponding to the specified product content, obtains a specified speech by encapsulating the received speech input by the agent, and sends the specified speech to the target large language model;

[0059] Generate a designated question corresponding to the designated speech based on the target large language model, display the designated question on the designated page, receive an answer corresponding to the designated question input by the agent, and repeatedly perform the steps of product rehearsal processing until a preset rehearsal end condition is met, and end the current dialogue session with the target large language model;

[0060] All target words input by the agent during the product rehearsal are obtained, and a rehearsal score of the agent is generated based on the target words.

[0061] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0062] Determine whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries product name information of the specified product;

[0063] If yes, extract the product name information from the product rehearsal request, and retrieve the specified product content corresponding to the product name information from a preset product database;

[0064] Generate a corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing;

[0065] After receiving the preparation completion information returned by the target large language model, sending a confirmation message to the agent;

[0066] Based on the target large language model, the agent performs a product rehearsal process corresponding to the specified product content, obtains a specified speech by encapsulating the received speech input by the agent, and sends the specified speech to the target large language model;

[0067] Generate a designated question corresponding to the designated speech based on the target large language model, display the designated question on the designated page, receive an answer corresponding to the designated question input by the agent, and repeatedly perform the steps of product rehearsal processing until a preset rehearsal end condition is met, and end the current dialogue session with the target large language model;

[0068] All target words input by the agent during the product rehearsal are obtained, and a rehearsal score of the agent is generated based on the target words.

[0069] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0070] The present application first determines whether a product rehearsal request triggered by an agent in a specified page is received; if so, the product name information is extracted from the product rehearsal request, and the specified product content corresponding to the product name information is retrieved from a preset product database; then a corresponding initial input is generated based on the specified product content, and the initial input is sent to a preset target large language model for processing; after receiving the preparation completion information returned by the target large language model, a confirmation message is sent to the agent; then, the product rehearsal processing corresponding to the specified product content is performed on the agent based on the target large language model, and the process is completed. The received speech input by the agent is packaged to obtain the designated speech, and the designated speech is sent to the target large language model; subsequently, a designated question corresponding to the designated speech is generated based on the target large language model, the designated question is displayed on the designated page, and the answer corresponding to the designated question input by the agent is received, and the steps of product drill processing are repeated until the preset drill end conditions are met, and the current dialogue session with the target large language model is ended; finally, all target speech input by the agent during the product drill is obtained, and the agent's drill score is generated based on the target speech. This application can efficiently and accurately assist agents in product drills based on the use of a target large language model, effectively improve the processing efficiency of product drills, reduce the cost of product drills, and improve the drill effect of product drills. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0072] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0073] Figure 2 A flowchart of an embodiment of an artificial intelligence-based data processing method according to the present application;

[0074] Figure 3 is a schematic structural diagram of an embodiment of an artificial intelligence-based data processing device according to the present application;

[0075] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0076] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0077] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0078] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0079] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0080] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0081] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to the laptop computer 1011, tablet computer 1012 or mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0082] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0083] It should be noted that the artificial intelligence-based data processing method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the artificial intelligence-based data processing device is generally set in the server / terminal device.

[0084] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0085] Continue to refer Figure 2 , showing a flow chart of an embodiment of an artificial intelligence-based data processing method according to the present application. According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted. The artificial intelligence-based data processing method provided in the embodiment of the present application can be applied to any scenario that requires data processing, and the artificial intelligence-based data processing method can be applied to products in these scenarios, for example, product drills in the financial insurance field. The artificial intelligence-based data processing method comprises the following steps:

[0086] Step S201, determining whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries product name information of a specified product.

[0087] In this embodiment, the data processing method based on artificial intelligence is executed on the electronic device (e.g. Figure 1The server / terminal device shown in the figure) can obtain the product name information through a wired connection or a wireless connection. It should be noted that the above-mentioned wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future. The executor of this application may be a product rehearsal system, or simply referred to as the system. The system is a pre-built product rehearsal system including a front-end user interface and back-end processing logic, and the system is connected to the large language model (LLM) API to ensure that the system can call the LLM in real time for dialogue processing. Among them, the above-mentioned seat may specifically be an insurance company's telemarketing seat. The above-mentioned designated page is the front-end user interface of the system. The seat triggers the above-mentioned product rehearsal request by entering or selecting the name of the insurance product to be rehearsed in the front-end user interface, that is, the product name information of the above-mentioned designated product.

[0088] Step S202: If yes, extract the product name information from the product rehearsal request, and retrieve the specified product content corresponding to the product name information from a preset product database.

[0089] In this embodiment, the product rehearsal request may be parsed to extract the product name information carried in the product rehearsal request, wherein the system may retrieve the corresponding designated product content from a preset product database according to the extracted product name information.

[0090] Step S203: Generate corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing.

[0091] In this embodiment, the system fills the specified product content into the matching position in the predefined product content template according to the retrieved specified product content to generate the corresponding initial input. For example: the above product content template may include: "{productContent} The above is the relevant content of {productName}. If you already understand, just reply OK". The generated initial input is then used as the input of the target large language model and sent to the target large language model through the API for processing. The target large language model will generate a corresponding reply based on the input content and feed it back to the system. Among them, there is no specific limitation on the model selection of the target large language model, and any large language model (LLM) can be used.

[0092] Step S204: after receiving the preparation completion information returned by the target large language model, sending a confirmation message to the agent.

[0093] In this embodiment, the system receives the reply from the target large language model and determines whether the reply content is "OK". If the reply from the target large language model is "OK", it means that the preparation completion information has been received, that is, the target large language model is ready and the product drill can be started. In addition, a confirmation message is sent to the agent through a communication method such as API or WebSocket, indicating that the product drill can be started. After receiving the confirmation message, the front-end user interface displays the corresponding prompt information to the agent.

[0094] Step S205, based on the target large language model, the agent performs a product rehearsal process corresponding to the specified product content, obtains the specified speech by encapsulating the received speech input by the agent, and sends the specified speech to the target large language model.

[0095] In this embodiment, the specific implementation process of encapsulating the received speech input by the agent to obtain the specified speech and sending the specified speech to the target large language model will be described in further detail in subsequent specific embodiments of the present application and will not be elaborated on here.

[0096] Step S206: Generate a specified question corresponding to the specified speech based on the target large language model, display the specified question on the specified page, receive an answer corresponding to the specified question input by the agent, and repeat the steps of product rehearsal processing until the preset rehearsal end conditions are met, and end the current dialogue session with the target large language model.

[0097] In this embodiment, the process of generating the designated question corresponding to the designated speech based on the target large language model, displaying the designated question on the designated page, and receiving the answer corresponding to the designated question input by the agent includes: the target large language model understands and analyzes the input designated speech in combination with the previously loaded designated product content, and then generates corresponding feedback or questions, such as: "Excuse me, what is the main coverage of the XX insurance you mentioned?" or "Can you give a detailed introduction to the claims process of XX insurance?" After receiving the question fed back by the target large language model, the system displays it to the agent. The agent will answer or further introduce the insurance product based on the question fed back by the target large language model.

[0098] Subsequently, the system again listens to the agent's input and encapsulates it into a format that the target large language model can understand. The system then sends the encapsulated agent input to the target large language model for processing, and the target large language model generates new feedback or questions based on the new input. The system receives the feedback or questions from the target large language model and displays them to the agent. Repeat the above steps until the agent chooses to end the conversation or reaches the scheduled rehearsal time, then it is determined that the rehearsal end conditions are met, and the current conversation session with the target large language model is ended to release related resources. In addition, the system will automatically update the agent's rehearsal status to "ended" and record the timestamp of the end of the call.

[0099] Step S207, obtaining all target words input by the agent during the product rehearsal process, and generating a rehearsal score for the agent based on the target words.

[0100] In this embodiment, during the conversation between the agent and the target large language model, the system has recorded in real time every word input by the agent and the feedback of the target large language model. These records are saved in a log file. All target words input by the agent during the product rehearsal can be extracted from the log file. Among them, the specific implementation process of generating the agent's rehearsal score based on the target words will be further described in detail in the subsequent specific embodiments of this application, and will not be elaborated on here.

[0101] The present application first determines whether a product rehearsal request triggered by an agent in a specified page is received; if so, the product name information is extracted from the product rehearsal request, and the specified product content corresponding to the product name information is retrieved from a preset product database; then a corresponding initial input is generated based on the specified product content, and the initial input is sent to a preset target large language model for processing; after receiving the preparation completion information returned by the target large language model, a confirmation message is sent to the agent; then, the product rehearsal processing corresponding to the specified product content is performed on the agent based on the target large language model, and the process is completed. The received speech input by the agent is packaged to obtain the designated speech, and the designated speech is sent to the target large language model; subsequently, a designated question corresponding to the designated speech is generated based on the target large language model, the designated question is displayed on the designated page, and the answer corresponding to the designated question input by the agent is received, and the steps of product drill processing are repeated until the preset drill end conditions are met, and the current dialogue session with the target large language model is ended; finally, all target speech input by the agent during the product drill is obtained, and the agent's drill score is generated based on the target speech. This application can efficiently and accurately assist agents in product drills based on the use of a target large language model, effectively improve the processing efficiency of product drills, reduce the cost of product drills, and improve the drill effect of product drills.

[0102] In some optional implementations, the step of generating the agent's practice score based on the target speech in step S207 includes the following steps:

[0103] The target speech is matched with sentences in the specified product content to obtain corresponding matching results.

[0104] In this embodiment, the target speech can be matched with the sentences in the above-mentioned specified product content by using natural language processing technology, such as keyword matching or semantic similarity calculation, and the corresponding matching results can be obtained. The matching results may include a successful sentence match or a failed sentence match.

[0105] The number of target sentences related to the designated product is determined based on the matching results.

[0106] In this embodiment, the matching results may be analyzed to count the successfully matched sentences, and the successfully matched sentences may be used as target sentences related to the designated product, and the number of corresponding target sentences may be counted.

[0107] Get the total number of sentences corresponding to the specified product content.

[0108] In this embodiment, the total number of sentences corresponding to the specified product content can be obtained by performing sentence statistics on the specified product content.

[0109] The number of sentences and the total number of sentences are calculated based on a preset calculation formula to obtain a corresponding calculation result.

[0110] In this embodiment, the above calculation formula can be specifically: score = number of sentences mentioned / total number of sentences in the product content × 100%, wherein the score calculation method can be adjusted as needed, such as whether to consider factors such as the importance and length of the sentence. The number of sentences mentioned above is the number of sentences mentioned, and the total number of sentences is / the total number of sentences in the product content. The number of sentences and the total number of sentences can be substituted into the above calculation formula for calculation, thereby obtaining the corresponding calculation result, i.e., the exercise score.

[0111] The calculation result is used as the training score of the agent.

[0112] The present application obtains a corresponding matching result by matching the target speech with the sentences in the specified product content; then determines the number of target sentences related to the specified product based on the matching result; then obtains the total number of sentences corresponding to the specified product content; subsequently, calculates and processes the number of sentences and the total number of sentences based on a preset calculation formula to obtain a corresponding calculation result, and uses the calculation result as the rehearsal score of the agent. The present application obtains a matching result by matching the target speech with the sentences in the specified product content, and determines the number of target sentences related to the specified product based on the matching result, and then calculates and processes the number of target sentences and the total number of sentences in the specified product content based on the use of a calculation formula, thereby automatically and accurately generating the rehearsal score of the agent, improving the generation efficiency of the rehearsal score, ensuring the accuracy of the obtained rehearsal score, and facilitating accurate evaluation of the rehearsal efficiency of the agent based on the obtained rehearsal score.

[0113] In some optional implementations of this embodiment, after the step of determining the number of target sentences related to the designated product based on the matching results, the method further includes:

[0114] Sentence elimination processing is performed on the designated product content based on the target sentence to obtain corresponding unmentioned sentences.

[0115] In this embodiment, the target sentences can be eliminated from the above-mentioned specified product content to obtain the eliminated content, and the eliminated content is used as the sentences not mentioned by the user during the product practice process, that is, the above-mentioned unmentioned sentences.

[0116] Get the preset drill result template.

[0117] In this embodiment, the drill result template is a template file constructed according to the actual drill result display requirements, and the drill result template includes at least a scoring area, a mentioned sentence area, and an unmentioned sentence area. Exemplarily, the content of the drill result template includes: "The drill score is XXX, the mentioned sentences include YYY, and the unmentioned sentences include ZZZ". Among them, XXX represents the scoring area, YYY represents the mentioned sentence area, and ZZZ represents the unmentioned sentence area.

[0118] The drill result template is filled based on the drill score, the target sentence and the unmentioned sentence to obtain a corresponding target drill result.

[0119] In this embodiment, the corresponding target drill result can be generated by filling the drill score, the target sentence and the unmentioned sentence into the score area, the mentioned sentence area and the unmentioned sentence area in the above-mentioned drill result template respectively.

[0120] The target drill result is displayed in the designated page.

[0121] In this embodiment, a display position may be set in advance on the designated page, and then the target exercise result may be displayed in the display position on the designated page.

[0122] This application performs sentence elimination processing on the content of the designated product based on the target sentence to obtain the corresponding unmentioned sentences; then obtains a preset drill result template; then fills the drill result template based on the drill score, the target sentence and the unmentioned sentence to obtain the corresponding target drill result; and subsequently displays the target drill result in the designated page. After determining the number of target sentences related to the designated product based on the matching results, this application will automatically perform sentence elimination processing on the content of the designated product based on the target sentence to obtain the unmentioned sentences, and then fills the preset drill result template based on the drill score, the target sentence and the unmentioned sentence to obtain the corresponding target drill result, and then displays the target drill result in the designated page to intelligently display the local product drill results to the agents, so that the agents can clearly understand their product drill performance in a timely manner, which is conducive to improving the agent's user experience.

[0123] In some optional implementations, step S205 includes the following steps:

[0124] Receive the speech input by the agent.

[0125] In this embodiment, the above-mentioned words can be the initial words related to the product drill that the agent inputs after completing the self-introduction.

[0126] Get the preset agent script template.

[0127] In this embodiment, according to the rehearsal requirements of the agents, speech templates for generating the input of the target large language model are pre-designed and configured. These templates may include product content templates (for displaying insurance product information) and agent speech templates (for simulating the agent's opening remarks, questions, etc.). By configuring these templates in the system, it is ensured that the system can automatically generate input templates that meet the requirements based on the selected product name and policy content. Among them, these templates can be stored in configuration files, database tables, or constants in the code, so that they can be called later.

[0128] The speech is packaged based on the agent speech template to obtain the corresponding specified speech.

[0129] In this embodiment, when the agent completes the self-introduction and inputs the speech, the speech will be captured. Then, according to the pre-configured agent speech template, the agent input speech is packaged into a format that can be understood by the target large language model, thereby obtaining the corresponding specified speech. Among them, the above-mentioned agent speech template may specifically include: "{salesInput} Please judge whether this sentence is related to which sentence in the {productName} content, and ask questions or respond accordingly."

[0130] The specified speech is sent to the target large language model.

[0131] In this embodiment, the packaged specified speech can be sent to the target large language model for processing.

[0132] The present application receives the speech input by the agent; then obtains a preset agent speech template; then encapsulates the speech based on the agent speech template to obtain the corresponding specified speech; and subsequently sends the specified speech to the target large language model. After receiving the speech input by the agent, the present application can quickly and accurately complete the encapsulation of the speech based on the use of the agent speech template, thereby effectively improving the accuracy and standardization of the generated specified speech.

[0133] In some optional implementations, after step S207, the electronic device may further perform the following steps:

[0134] Collect relevant data of the agent during the product drill.

[0135] In this embodiment, the above-mentioned relevant data at least includes the target speech, model feedback data and the rehearsal score. The above-mentioned model feedback data refers to the feedback data output by the target large language model during the product rehearsal process.

[0136] The relevant data are analyzed to obtain corresponding data analysis results.

[0137] In this embodiment, the above-mentioned specific implementation process of analyzing the relevant data to obtain the corresponding data analysis results will be further described in detail in subsequent specific embodiments of the present application, and will not be elaborated in detail here.

[0138] The target large language model is optimized based on the data analysis result.

[0139] In this embodiment, the process of optimizing the target large language model based on the data analysis results may specifically include: model evaluation, data enhancement, model fine-tuning and effect evaluation processing. Among them, model evaluation includes: based on the data analysis results, evaluating the current large language model and identifying which aspects of the model need to be improved. Data enhancement includes: collecting more training data, including richer insurance product information, more diverse dialogue scenarios, etc., in view of the shortcomings of the model in understanding and generating speech. Model fine-tuning includes: fine-tuning the model using new training data, optimizing the parameters and structure of the model, and improving the accuracy and generalization ability of the model. Effect evaluation includes: testing on the fine-tuned model, evaluating the improvement of the model in understanding and generating speech, and ensuring that the performance of the optimized model meets expectations. Exemplarily, in view of the shortcomings of the target large language model in understanding complex insurance product information, more training data such as insurance product manuals and customer consultation records are collected. Then, the model is fine-tuned according to the training data, the model's ability to understand insurance product information is optimized, and the accuracy and relevance of the generated feedback are improved.

[0140] This application collects relevant data from the agent during the product rehearsal process, analyzes the relevant data to obtain corresponding data analysis results, and then optimizes the target large language model based on the data analysis results. This application collects relevant data from the agent during the product rehearsal process, analyzes the relevant data to obtain data analysis results, and then intelligently optimizes the target large language model based on the data analysis results, thereby effectively optimizing the target large language model's ability to understand product information, thereby helping to improve the accuracy and relevance of the feedback generated by the target large language model.

[0141] In some optional implementations of this embodiment, the relevant data includes at least the target speech, model feedback data, and the drill score; and analyzing the relevant data to obtain corresponding data analysis results includes the following steps:

[0142] The target speech is analyzed to obtain a corresponding first analysis result.

[0143] In this embodiment, the target speech input by the agent can be analyzed to identify commonly used speech, effective speech, invalid speech, and errors or deficiencies in the speech, thereby obtaining a corresponding first analysis result.

[0144] The model feedback data is analyzed to obtain a corresponding second analysis result.

[0145] In this embodiment, the above-mentioned model feedback data refers to the feedback data output by the target large language model during the product rehearsal process. By analyzing the model feedback data, the accuracy and relevance of the model feedback can be evaluated, and the deficiencies of the target large language model in understanding and generating speech can be identified, thereby obtaining the corresponding second analysis result.

[0146] The drill score is analyzed to obtain a corresponding third analysis result.

[0147] In this embodiment, by performing statistical analysis on the drill scores, the weak links in the agent's familiarity with the specified products can be identified, such as which specified products or aspects of information the agent does not have sufficient grasp of, thereby obtaining the corresponding third analysis results.

[0148] The data analysis result is generated based on the first analysis result, the second analysis result, and the third analysis result.

[0149] In this embodiment, the first analysis result, the second analysis result and the third analysis result can be integrated to obtain an integrated result, and the integrated result can be used as a data analysis result. In addition, the progress and changes of the agents in the product rehearsal process and the improvement of system performance can be observed through time series analysis, and the obtained data analysis results can be further optimized to improve the diversity and accuracy of the data analysis results.

[0150] This application obtains the corresponding first analysis result by analyzing the target speech, and obtains the corresponding second analysis result by analyzing the model feedback data, and obtains the corresponding third analysis result by analyzing the drill score, and subsequently generates the data analysis result based on the first analysis result, the second analysis result, and the third analysis result. This application obtains the corresponding first analysis result, the second analysis result, and the third analysis result by analyzing the target speech, the model feedback data, and the drill score, respectively, and then generates the final data analysis result based on the obtained first analysis result, the second analysis result, and the third analysis result, effectively improving the richness and accuracy of the generated data analysis results.

[0151] In some optional implementations of this embodiment, before step S202, the electronic device may further perform the following steps:

[0152] Get detailed information on various products.

[0153] In this embodiment, in the business scenario of the financial field, the various products mentioned above may include at least financial products, claims products, insurance products, etc. The detailed information may include product name, product content (policy content), premium, coverage, etc.

[0154] Call the preset initial database.

[0155] In this embodiment, an initial database may be established by using a database management system, and a reasonable table structure may be designed to store detailed information of the product.

[0156] The detailed information is stored in the initial database to obtain a corresponding first database.

[0157] In this embodiment, the collected detailed information of various products can be stored in the initial database, thereby obtaining a corresponding first database.

[0158] Perform search optimization processing on the first database to obtain a corresponding second database.

[0159] In this embodiment, the above retrieval optimization process may include: creating an index for the database table, optimizing the query statement, and ensuring efficient retrieval of relevant product information. In addition, full-text indexing, combined indexing and other technologies may be further used to improve the retrieval speed.

[0160] The second database is used as the product database.

[0161] The present application obtains detailed information of various products; then calls a preset initial database; then stores the detailed information in the initial database to obtain a corresponding first database; subsequently performs search optimization processing on the first database to obtain a corresponding second database, and uses the second database as the product database. The present application obtains detailed information of various products in a preset initial database to obtain a first database, and then performs search optimization processing on the first database, so that the required product database can be automatically and intelligently constructed, effectively ensuring that the constructed product database can efficiently retrieve relevant product information and improve the retrieval speed of the product database.

[0162] In some optional implementations, the user information obtained is subject to the user's consent and complies with relevant laws and policies.

[0163] In addition, the non-Company software tools or components that appear in the embodiments of the present application are merely examples and do not represent actual use.

[0164] In addition, based on the product drill method proposed in this application, agents can choose the products they want to train for familiarity drills anytime and anywhere according to their personal needs, and support dynamic configurability, that is, after a new insurance product is launched, it can be quickly transmitted to the telemarketing agents for drills, greatly reducing the training costs of each telemarketing agent by the supervisor. In addition, this application can also score the drills of each telemarketing agent according to the process of the product drill, helping the telemarketing agents to find their own deficiencies in familiarity with a certain product during the product drill, so as to improve them specifically.

[0165] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0166] It should be emphasized that in order to further ensure the privacy and security of the above-mentioned drill scores, the above-mentioned drill scores can also be stored in a node of a blockchain.

[0167] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0168] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0169] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0170] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0171] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0172] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a data processing device based on artificial intelligence. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0173] like Figure 3As shown, the artificial intelligence-based data processing device 300 described in this embodiment includes: a judgment module 301, a retrieval module 302, a first sending module 303, a second sending module 304, a first processing module 305, a second processing module 306 and a generation module 307.

[0174] in:

[0175] The judgment module 301 is used to judge whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries the product name information of the specified product;

[0176] Retrieval module 302, for extracting the product name information from the product rehearsal request, and retrieving the specified product content corresponding to the product name information from a preset product database;

[0177] A first sending module 303 is used to generate a corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing;

[0178] The second sending module 304 is used to send a confirmation message to the agent after receiving the preparation completion information returned by the target large language model;

[0179] The first processing module 305 is used to perform a product rehearsal process corresponding to the specified product content on the agent based on the target large language model, obtain the specified speech by encapsulating the speech input by the agent, and send the specified speech to the target large language model;

[0180] The second processing module 306 is used to generate a specified question corresponding to the specified speech based on the target large language model, display the specified question on the specified page, receive an answer corresponding to the specified question input by the agent, and repeatedly perform the steps of product rehearsal processing until a preset rehearsal end condition is met, and end the current dialogue session with the target large language model;

[0181] The generation module 307 is used to obtain all target words input by the agent during the product rehearsal process, and generate a rehearsal score for the agent based on the target words.

[0182] In this embodiment, the operations performed by the above modules or units correspond one-to-one to the steps of the artificial intelligence-based data processing method in the aforementioned implementation mode, and will not be repeated here.

[0183] In some optional implementations of this embodiment, the generating module 307 includes:

[0184] A matching submodule, used to match the target speech with the sentences in the specified product content to obtain corresponding matching results;

[0185] A first determination submodule, configured to determine the number of target sentences related to the designated product based on the matching result;

[0186] A first acquisition submodule is used to acquire the total number of sentences corresponding to the specified product content;

[0187] A calculation submodule, used to calculate the number of sentences and the total number of sentences based on a preset calculation formula to obtain a corresponding calculation result;

[0188] The second determination submodule is used to use the calculation result as the training score of the agent.

[0189] In some optional implementations of this embodiment, the generating module 307 further includes:

[0190] A second acquisition submodule is used to perform sentence elimination processing on the specified product content based on the target sentence to obtain corresponding unmentioned sentences;

[0191] The third acquisition submodule is used to obtain a preset drill result template;

[0192] A filling submodule, used for filling the training result template based on the training score, the target sentence and the unmentioned sentence to obtain a corresponding target training result;

[0193] The display submodule is used to display the target exercise result in the specified page.

[0194] In some optional implementations of this embodiment, the first processing module 305 includes:

[0195] A receiving submodule, used for receiving the speech input by the agent;

[0196] The fourth acquisition submodule is used to obtain a preset agent speech template;

[0197] The encapsulation submodule is used to encapsulate the speech based on the agent speech template to obtain the corresponding specified speech;

[0198] The sending submodule is used to send the specified speech to the target large language model.

[0199] In some optional implementations of this embodiment, the artificial intelligence-based data processing device further includes:

[0200] A collection module, used to collect relevant data of the agent during the product rehearsal process;

[0201] An analysis module, used to analyze the relevant data to obtain corresponding data analysis results;

[0202] The first optimization module is used to optimize the target large language model based on the data analysis result.

[0203] In some optional implementations of this embodiment, the analysis module includes:

[0204] A first analysis submodule is used to analyze the target speech to obtain a corresponding first analysis result;

[0205] A second analysis submodule is used to analyze the model feedback data to obtain a corresponding second analysis result;

[0206] A third analysis submodule is used to analyze the drill score to obtain a corresponding third analysis result;

[0207] A generating submodule is used to generate the data analysis result based on the first analysis result, the second analysis result and the third analysis result.

[0208] In some optional implementations of this embodiment, the artificial intelligence-based data processing device further includes:

[0209] The acquisition module is used to obtain detailed information of various products;

[0210] A calling module is used to call a preset initial database;

[0211] A storage module, used for storing the detailed information in the initial database to obtain a corresponding first database;

[0212] A second optimization module, used to perform search optimization processing on the first database to obtain a corresponding second database;

[0213] A determination module is used to use the second database as the product database.

[0214] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0215] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0216] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0217] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk equipped on the computer device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions of data processing methods based on artificial intelligence, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0218] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run computer-readable instructions stored in the memory 41 or process data, such as computer-readable instructions for running the artificial intelligence-based data processing method.

[0219] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0220] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the artificial intelligence-based data processing method as described above.

[0221] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0222] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A data processing method based on artificial intelligence, characterized in that: The steps include: Determine whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries product name information of the specified product; If yes, extract the product name information from the product rehearsal request, and retrieve the specified product content corresponding to the product name information from a preset product database; Generate a corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing; After receiving the preparation completion information returned by the target large language model, sending a confirmation message to the agent; Based on the target large language model, the agent performs a product rehearsal process corresponding to the specified product content, obtains a specified speech by encapsulating the received speech input by the agent, and sends the specified speech to the target large language model; Generate a designated question corresponding to the designated speech based on the target large language model, display the designated question on the designated page, receive an answer corresponding to the designated question input by the agent, and repeatedly perform the steps of product rehearsal processing until a preset rehearsal end condition is met, and end the current dialogue session with the target large language model; All target words input by the agent during the product rehearsal are obtained, and a rehearsal score of the agent is generated based on the target words.

2. The data processing method based on artificial intelligence according to claim 1, characterized in that: The step of generating the agent's practice score based on the target speech specifically includes: Matching the target speech with the sentences in the specified product content to obtain corresponding matching results; Determining the number of target sentences related to the designated product based on the matching results; Obtain the total number of sentences corresponding to the specified product content; Calculating the number of sentences and the total number of sentences based on a preset calculation formula to obtain a corresponding calculation result; The calculation result is used as the training score of the agent.

3. The data processing method based on artificial intelligence according to claim 2, characterized in that: After the step of determining the number of target sentences related to the designated product based on the matching results, the method further includes: Perform sentence elimination processing on the specified product content based on the target sentence to obtain corresponding unmentioned sentences; Get the preset drill result template; Filling the drill result template based on the drill score, the target sentence, and the unmentioned sentence to obtain a corresponding target drill result; The target drill result is displayed in the designated page.

4. The data processing method based on artificial intelligence according to claim 1, characterized in that: The step of encapsulating the received speech input by the agent to obtain the specified speech, and sending the specified speech to the target large language model specifically includes: Receiving the speech input by the agent; Get the preset agent script template; Encapsulating the speech based on the agent speech template to obtain the corresponding specified speech; The specified speech is sent to the target large language model.

5. The data processing method based on artificial intelligence according to claim 1, characterized in that: After the step of generating the training score of the agent based on the target speech, the method further includes: Collect relevant data of the agent during the product drill; Analyze the relevant data to obtain corresponding data analysis results; The target large language model is optimized based on the data analysis result.

6. The artificial intelligence-based data processing method according to claim 5, characterized in that: The relevant data at least includes the target speech, model feedback data and the drill score; the step of analyzing the relevant data to obtain the corresponding data analysis results specifically includes: Analyze the target speech to obtain a corresponding first analysis result; Analyzing the model feedback data to obtain a corresponding second analysis result; Analyzing the drill score to obtain a corresponding third analysis result; The data analysis result is generated based on the first analysis result, the second analysis result, and the third analysis result.

7. The data processing method based on artificial intelligence according to claim 1, characterized in that: Before the step of retrieving the designated product content corresponding to the product name information from the preset product database, the method further includes: Get detailed information on various products; Call the preset initial database; Storing the detailed information in the initial database to obtain a corresponding first database; Performing search optimization processing on the first database to obtain a corresponding second database; The second database is used as the product database.

8. A data processing device based on artificial intelligence, characterized in that: include: A judgment module, used to judge whether a product drill request triggered by an agent in a specified page is received; wherein the product drill request carries product name information of a specified product; A retrieval module, for extracting the product name information from the product rehearsal request, and retrieving the specified product content corresponding to the product name information from a preset product database; A first sending module, configured to generate a corresponding initial input based on the specified product content, and send the initial input to a preset target large language model for processing; A second sending module is used to send a confirmation message to the agent after receiving the preparation completion information returned by the target large language model; A first processing module is used to perform a product rehearsal process corresponding to the specified product content on the agent based on the target large language model, obtain a specified speech by encapsulating the received speech input by the agent, and send the specified speech to the target large language model; A second processing module is used to generate a specified question corresponding to the specified speech based on the target large language model, display the specified question on the specified page, receive an answer corresponding to the specified question input by the agent, and repeatedly perform product rehearsal processing steps until a preset rehearsal end condition is met, and end the current dialogue session with the target large language model; A generation module is used to obtain all target words input by the agent during the product rehearsal process, and generate a rehearsal score for the agent based on the target words.

9. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the artificial intelligence-based data processing method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the artificial intelligence-based data processing method according to any one of claims 1 to 7.

Images