Question guidance method, server and storage medium

Abstract

The present application belongs to the field of artificial intelligence, and specifically relates to a question guidance method, server and storage medium. The question guidance method includes: obtaining the recent consultation question set and recent behavior data set of each user terminal; based on the recent consultation question set, determining the target question type that each user terminal is concerned about; based on the recent behavior data set, determining the target product type that each user terminal is concerned about; based on the target product type and the target question type, determining the product issues that each user terminal is concerned about, and associating each user terminal identifier with the product issues that each user terminal is concerned about and storing them in a cache database; after detecting that the user terminal has entered the customer service system, sending the product issues that the user terminal is concerned about to the user terminal. The method of the present application first analyzes the product issues that the user is concerned about based on big data technology and provides guidance before the user asks a question, thereby improving customer service efficiency and satisfaction.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence, and specifically relates to a question guidance method, a server, and a storage medium. Background Art

[0002] Currently, many product sales platforms have online customer service systems, allowing users to access these systems to inquire about products. Because the same question can be expressed in a variety of languages, if there is a significant discrepancy between the user's description and the knowledge in the customer service knowledge base, the similarity between the two will be low. This can result in the customer service knowledge base being unable to recall the correct answer, leading to incorrect responses, significantly reducing the user experience. Furthermore, many users don't carefully consider their question descriptions and simply ask a few short words. For example, a simple question like "unable to charge" doesn't specify the product, model, or charging stage. If the customer service system responds directly based on the question, there could be 100 potential answers, making it difficult to precisely select the one the user is looking for. By providing prompts and guidance based on the user's vague expression, allowing the user to provide more information, it is more likely to filter out the desired answer from these 100 possible answers.

[0003] To solve the above problems, most existing customer service systems are optimized after users submit questions. This makes it completely uncontrollable for users to ask questions, and will increase additional time and satisfaction costs to clarify user intentions. Summary of the Invention

[0004] The embodiments of the present application provide a question guidance method, a server, and a storage medium, which can improve customer service efficiency and satisfaction.

[0005] In a first aspect, an embodiment of the present application provides a question guidance method, the method comprising:

[0006] Obtain the recent consultation question set and recent behavior dataset of each user terminal;

[0007] Based on the recent consultation question set, determining the target question type that each user terminal is concerned about;

[0008] Based on the recent behavior data set, determining the target product type that each user terminal is interested in;

[0009] Based on the target product type and the target question type, determining the product questions that each user terminal is concerned about, and associating each user terminal identifier with the product question that each user terminal is concerned about and storing it in a cache database;

[0010] After detecting that the user terminal has entered the customer service system, the product issues that the user terminal is concerned about are sent to the user terminal, so that the user terminal generates a predicted question list based on the product issues that the user terminal is concerned about for display.

[0011] Optionally, the method further includes:

[0012] Input the consulting questions collected by the customer service system into the trained question bias classification model for type labeling to obtain the question type of the consulting question;

[0013] Determining the target question type that the user is concerned about based on the recent consultation question set includes:

[0014] Statistics are collected on each consulting question in the recent consulting question set according to question type, and N question types with the highest number of questions are selected as target question types of interest to the user terminal, where N is an integer greater than or equal to 1.

[0015] Optionally, the determining, based on the target product type and the target question type, product questions that each user terminal is concerned about includes:

[0016] Count the proportion of each target question type in the recent consultation question set;

[0017] Obtain a set of product consultation questions corresponding to the target product type;

[0018] According to the N target question types and the proportion of the number of questions corresponding to each target question type, a preset first number of product consulting questions are screened from the product consulting question set as the product questions that each user terminal is concerned about.

[0019] Optionally, the method further includes:

[0020] Obtaining customer service question-and-answer data and product inquiry data, merging the customer service question-and-answer data and the product inquiry data based on the questions to generate complete question-and-answer data, and storing the complete question-and-answer data in the cache database and the vector database, respectively. The customer service question-and-answer data includes multiple customized questions and corresponding customized answers, and the product inquiry data includes the inquiry questions and corresponding responses of each user terminal based on each product type;

[0021] Clustering is performed on the consulting questions in the product consulting data to generate product frequently asked questions, and the product frequently asked questions and corresponding reply answers are formed into product frequently asked question and answer data and stored in the cache database.

[0022] Optionally, the method further includes:

[0023] When detecting that the user terminal manually inputs a question in the customer service system, determining a completion mode based on the manual input method and / or the manual input content;

[0024] If the completion mode is level one completion, keywords are extracted from the manually input content, the keywords are matched with the question list data in the cache database, and a level one completion question list is generated based on a number of successfully matched questions and sent to the user terminal;

[0025] If the completion mode is secondary completion, the manually input content is matched with the question list data in the vector database for similarity, and a secondary completion question list is generated based on a number of successfully matched questions and sent to the user terminal;

[0026] If the completion mode is three-level completion, a number of top-ranked questions are selected from the commodity FAQ data in the cache database to generate a three-level completion question list and send it to the user terminal.

[0027] Optionally, determining the completion mode based on the manual input method and / or the manually input content includes:

[0028] When detecting that the pause duration of the manual input is longer than the first preset pause duration, determining the completion mode to be the first-level completion;

[0029] When it is detected that the manually input pause duration is longer than a second preset pause duration and the first-level completion question list is empty, determining that the completion mode is the second-level completion mode, and the second preset pause duration is longer than the first preset pause duration;

[0030] When it is detected that the manually input content does not contain the product type and the second-level completion question list is empty, the completion mode is determined to be the third-level completion.

[0031] Optionally, performing similarity matching between the manually input content and the question list data in the vector database, and generating a secondary completed question list based on a number of successfully matched questions and sending the list to the user terminal includes:

[0032] The manually input content is matched with the question list data in the vector database for similarity, and a number of questions whose similarity values ​​exceed a preset similarity threshold and whose similarity values ​​are ranked high are selected to generate a secondary completion question list.

[0033] Optionally, the method further includes:

[0034] Count the number of conversations and messages before and after the implementation of the question-guided strategy;

[0035] The efficiency index before and after the implementation of the question-guiding strategy is calculated based on the number of conversations and messages before and after the implementation of the question-guiding strategy;

[0036] The effectiveness of the question-guiding strategy was evaluated based on the changes in efficiency indicators before and after the implementation of the question-guiding strategy.

[0037] In a second aspect, an embodiment of the present application provides a customer service server, comprising at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described above.

[0038] In a third aspect, an embodiment of the present application provides a computer storage medium, which stores instructions or programs. When the instructions or programs are executed by at least one processor, the at least one processor executes the method described above.

[0039] In an embodiment of the present application, a question-guiding method is provided, which first determines the target question type that the user terminal is concerned about based on the user terminal's recent consulting question set, and determines the target commodity type that the user terminal is concerned about based on the user terminal's recent behavior data set; then, based on the target commodity type and the target question type, determines the commodity issues that each user terminal is concerned about, associates each user terminal identifier with the commodity issues that each user terminal is concerned about and stores them in a cache database; finally, when it is detected that the user terminal has entered the customer service system, the commodity issues that the user terminal is concerned about are sent to the user terminal, so that the user terminal generates a predicted question list based on the commodity issues that the user terminal is concerned about for display. The method of the present application first analyzes the commodity issues that the user is concerned about based on big data technology before the user asks a question, intercepts the question at the beginning of the question, saves the user's time in organizing the question, and improves customer service efficiency and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The following is an example of the overall process architecture of a customer service system's question guidance.

[0041] Figure 2 A flowchart of a method for guiding questions before inputting a question is exemplified;

[0042] Figure 3 A flowchart of a method for guiding questions in inputting questions is exemplified;

[0043] Figure 4 An example diagram of input box question completion is shown;

[0044] Figure 5 The hardware structure diagram of the customer service server is shown as an example. DETAILED DESCRIPTION

[0045] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0046] The technical terms used in this application are explained as follows:

[0047] 1. Before asking a question: refers to all operations performed by the user before clicking the Submit Question button.

[0048] 2. Post-questioning: This refers to the background processing operations after the user clicks the Submit Question button.

[0049] 3. Redis: An open-source, in-memory data storage system, it's often used as a cache, message queue, and database. It's a NoSQL database that supports key-value pairs, but also offers a variety of complex data structures, such as strings, hashes, lists, sets, sorted sets, and bitmaps.

[0050] 4. Vector Database: A database system specifically designed for storing, managing, and retrieving high-dimensional vector data. In many applications, particularly in machine learning and artificial intelligence, vectors (typically arrays of floating-point numbers) are used to represent the characteristics or semantic information of data such as text, images, audio, and video.

[0051] 5. Energy storage equipment: A device or system used to store electrical energy, typically consisting of battery cells, inverters, and control systems.

[0052] Please refer to Figure 1 , Figure 1 The customer service system architecture including the question guidance function is shown. The customer service system includes a data module 1, a prediction module 2, a completion module 3 and a feedback module 4. Among them, the data module 1 is the core component of the customer service system, which is used to collect, store and manage different types of data so as to perform efficient retrieval and rapid response in the subsequent customer service process. The prediction module 2 uses big data to analyze user behavior in real time, predicts the content of the user's question before the user is ready to ask a question, and outputs a list of predicted questions so that the user can get the desired answer after clicking. The completion module 3 is used to provide an intelligent automatic completion function when the user enters a question, helping the user to ask questions in a more standardized way, thereby improving the accuracy of question and answer matching and the efficiency of the response. The feedback module 4 is used to compare and analyze the efficiency indicators before and after the execution of the question guidance strategy, so as to optimize the question guidance strategy and improve the efficiency of the customer service system.

[0053] Specifically, the data sources of data module 1 include product consultation data, user behavior data and customer service question and answer data. Among them, the product consultation data includes the consultation questions and corresponding reply answers of each user terminal based on each product type. User behavior data includes various operation data of each user terminal in the online mall, such as purchasing products, adding to shopping carts, searching for products, browsing products, collecting products, clicking on products, etc., which can be recorded through point-of-sale. Customer service question and answer data includes multiple custom questions preset by the customer service team and corresponding custom answers. Among them, the custom questions and corresponding custom answers are set by the customer service team based on business experience around the parameters, characteristics, faults and other aspects of each product type. For example: Custom question 1: Does AC180 support waterproofing? Custom answer 1: Support.

[0054] In the prior art, after obtaining the content of the question submitted by the user, a direct search is generally performed in the product consultation data, user behavior data and customer service question and answer data based on the question content. After matching similar content, the corresponding answer is returned. This method not only has low retrieval efficiency, but also cannot guarantee that the answer that matches the user's expression can be given. To solve this problem, the present application guides users to ask questions in various ways before and during the user's question, and obtains a variety of processed data by processing the product consultation data, user behavior data and customer service question and answer data, and stores each processed data in different databases, thereby quickly providing the required data for various guided questions and improving the efficiency of question guidance.

[0055] First, data module 1 obtains user-product-related question-and-answer data based on product inquiry data and user behavior data, and stores the user-product-related question-and-answer data in a Redis cache database. When the prediction module 2 detects that a user terminal has entered the customer service system, it can quickly obtain the user terminal's product questions from the Redis cache database and send them to the user terminal. Specifically, data module 1 obtains the user terminal's recent inquiry question set (derived from product inquiry data) and recent behavior data set (derived from user behavior data) based on each user terminal. Based on the recent inquiry question set, it determines the target question type of the user terminal, and based on the recent behavior data set, it determines the target product type of the user terminal. Then, based on the target product type and target question type, it determines the product questions of interest to each user terminal, and associates each user terminal identifier with the product questions and responses to each user terminal and stores them in the cache database. The user terminal identifier may include one or more information such as the user terminal ID, the user terminal IP address, and the user terminal's geographic location. Storing multiple types of user terminal information in association with the product questions and responses to the user can achieve expanded recommendations for predicted questions. For example, when a user terminal enters the customer service system for consultation for the first time, product issues associated with other user terminals in the same geographical location as the user terminal can be pushed to the user terminal.

[0056] Secondly, Data Module 1 uses a clustering algorithm to extract representative questions from the product inquiry data as frequently asked questions (FAQs). It then stores these FAQs and their responses in a Redis cache database as FAQ data. When Completion Module 3 enters three-level completion mode, it generates a three-level list of completed questions based on the FAQ data and sends it to the user, providing question completion functionality when specific product information is unavailable.

[0057] Third, data module 1 merges customer service Q&A data and product inquiry data based on questions to generate complete Q&A data, which is then stored in a Redis cache database and a vector database, respectively. Specifically, data module 1 extracts all custom questions from the customer service Q&A data and all inquiry questions from the product inquiry data. After merging all custom questions and all inquiry questions, it performs deduplication and clustering to generate a complete question list. It then generates complete Q&A data based on the complete question list and the answers to each question in the complete question list. When completion module 3 enters the first-level completion mode, it extracts keywords from the input content and quickly matches them against all Q&A data in the Redis cache database to generate a first-level completed question list and send it to the user. When completion module 3 enters the second-level completion mode, it extracts all input content and performs similarity matching against all Q&A data in the vector database to generate a second-level completed question list and send it to the user. In one embodiment, the vector database uses the HNSW (Hierarchical Navigable Small World) algorithm to index the customer service Q&A data, thereby accelerating nearest neighbor searches in multidimensional space.

[0058] In one embodiment, the data module 1 regularly updates the user-product related Q&A data, product FAQ data, and all Q&A data. For example, at 11:00 PM every night, the data module 1 obtains various questions asked by users that day and updates the above data.

[0059] In one embodiment, the prediction module 2 is used to detect that the user terminal has entered the customer service system, and then search for product issues that the user terminal is concerned about in the user and product associated question and answer data in the Redis cache database based on the identification of the user terminal, and send the product issues that the user terminal is concerned about to the user terminal, so that the user terminal generates a predicted question list based on the product issues that the user terminal is concerned about for display.

[0060] Completion module 3 is used to provide intelligent automatic completion function when the user enters the question. Specifically, completion module 3 includes three levels of completion process:

[0061] (1) First-level completion

[0062] When a user enters a question, the customer service system will instantly search the entire Q&A database in Redis for questions that match the user's input keywords, generate a list of first-level completion questions, and present them to the user for selection. This completion method primarily relies on fuzzy matching of keywords.

[0063] (2) Secondary completion

[0064] If keyword matching fails to find a suitable question, the customer service system will further search the entire Q&A database for questions with high similarity to the user's input based on the user's input, generate a secondary list of completed questions, and present it to the user. This process uses a vector model for matching, ensuring that even when the user enters complex or uncommon questions, the system can still provide highly relevant candidate questions.

[0065] (3) Three-level completion

[0066] If the second-level completion function fails to find a relevant question, the customer service system generates a third-level completion list based on the product FAQ data in the Redis cache database and displays it to the user, helping them quickly find relevant answers. This level of completion is mainly used when the user's question is too vague or unrelated to the product.

[0067] Feedback module 4 is used to compare and analyze the efficiency indicators before and after the execution of the question guidance strategy to optimize the question guidance strategy and improve the efficiency of the customer service system. Specifically, the number of conversations and the number of messages before and after the execution of the question guidance strategy are first monitored and counted; then, the efficiency indicators before and after the execution of the question guidance strategy are calculated based on the number of conversations and the number of messages before and after the execution of the question guidance strategy, and the effectiveness of the question guidance strategy is evaluated based on the changes in the efficiency indicators before and after the execution of the question guidance strategy. The number of conversations refers to the number of conversation rounds between the user end and the customer service system. For example, a conversation round ends when the conversation interval exceeds 1 hour. The number of messages refers to the total number of messages between the user end and the customer service system in each conversation round.

[0068] In one embodiment, the efficiency index E is defined as the number of messages / the number of conversations. Assume that the efficiency index after executing the question-guiding strategy is , the efficiency index before executing the question-guiding strategy is ,when When , that is, after the question-guiding strategy is executed, the number of messages in each round of dialogue on the user side is reduced compared with that before the question-guiding strategy, it means that the question-guiding strategy is effective and the communication efficiency of the customer service system is improved; when When , that is, the number of messages in each round of conversation on the user side increases after the question guidance strategy is executed compared with that before the question guidance, it means that the question guidance strategy fails to effectively improve efficiency and needs further optimization.

[0069] In other embodiments, the feedback module 4 is further used to count the number of times the prediction question list, the first-level completion question list, the second-level completion question list and the third-level completion question list are triggered, and to count the number of times the questions in the prediction question list, the first-level completion question list, the second-level completion question list and the third-level completion question list are selected, and to evaluate and optimize the screening strategy for questions in each type of list by calculating the ratio of the number of times the questions in each type of list are selected to the number of times the list is triggered within a preset time period (for example, one month).

[0070] Typically, the customer service system consultation process includes the following steps: Step 1: The user generates a consultation request for a product; Step 2: The user enters a question in the customer service system's dialog window; Step 3: The user clicks the Submit button; Step 4: The customer service system displays the response in the dialog window. This application's method focuses on providing guidance and assistance during step 2, after the user enters the customer service system and before clicking the Submit button (both before and during question entry). This can also be called "pre-question" guidance and optimization.

[0071] Please refer to Figure 2 , Figure 2 A flowchart of a method for guiding questions before inputting a question is shown, the method comprising:

[0072] Step S201: Obtain a recent consultation question set and a recent behavior data set of each user terminal.

[0073] In one embodiment, the customer service system collects and processes the consultation questions and responses of each user terminal based on each product type based on big data technology to obtain product consultation data, and collects and processes various operation data of each user terminal in the online mall (for example, through the method of embedding points to record the user's purchase of products, addition to shopping carts, search for products, browsing products, collection of products, click on products, etc.) to obtain user behavior data.

[0074] In one embodiment, the product consultation data records the initiation time of each consultation question. Based on the initiation time, the consultation questions initiated within the most recent preset time period can be filtered out from the product consultation data to form a recent consultation question set. The user behavior data records the operation time of each behavior. Based on the operation time, the user behavior that occurred within the most recent preset time period can be filtered out from the user behavior data to form a recent behavior data set. For example, the recent consultation question set includes consultation questions initiated by the user terminal within the past three months, and the recent behavior data set includes the user terminal's behavior data of purchasing products, adding to shopping carts, searching for products, browsing products, adding products to favorites, and clicking on products in the online mall within the past three months.

[0075] Step S202: Based on the recent consultation question set, the target question type of each user terminal is determined.

[0076] In one embodiment, the consulting questions collected by the customer service system are input into a trained question bias classification model for type labeling to obtain the question type of the consulting question. The stored content corresponding to each consulting question in the product consulting data may include the question content, answer content, product type, user terminal identifier, question type, and initiation time. The specific implementation process of step S202 is: each consulting question in the recent consulting question set is counted by question type, and the top N question types ranked by number of questions are selected as the target question types of the user terminal, where N is an integer greater than or equal to 1.

[0077] Among them, the training methods of the problem-biased classification model include:

[0078] Step 1: Label the types of questions in the training samples. For example, the labeled question types include:

[0079] 1) Technical preference: Provides technical specifications, product features, setup guides, and other frequently asked questions.

[0080] 2) Price-sensitive: Display discount information, promotions, price comparisons, and other related questions.

[0081] 3) High-frequency purchasing type: providing after-sales service, product warranty, replacement parts and other issues.

[0082] 4) Casual browsing: Provides questions related to product design, user experience, brand story, etc.

[0083] 5) Evaluation-driven: Provides product reviews, user feedback, rating questions, etc.

[0084] 6) Health / environmental concerns: Provide information on material sources, environmental certification, and whether they are harmless.

[0085] 7) Brand Loyalty: Provide questions such as brand history, user loyalty program, brand activities, etc.

[0086] Annotation example:

[0087] 1) Question: "Does this headset support noise cancellation?" Labeled as "Technical preference"

[0088] 2) Question: "Is there a discount on this phone?" Labeled as "price sensitive"

[0089] 3) Question: "How do the headphones feel?" Labeled as "Review-driven"

[0090] 4) Question: "Can I return or exchange my phone?" (Marked as "High-Frequency Purchaser")

[0091] Step 2: Use the naive Bayes algorithm to train the model of the labeled questions to obtain a trained question-biased classification model.

[0092] Step S203: Based on the recent behavior data set, the target product type that each user terminal is interested in is determined.

[0093] In one embodiment, the user's recently purchased items are first determined based on the product purchase data in the recent behavior dataset. If the item exists, it is identified as the target product type of the user's interest. If not, the recent behavior dataset is searched for the user's most searched items. If so, it is identified as the target product type of the user's interest. If not, the recent behavior dataset is searched for the user's most browsed items. If so, it is identified as the target product type of the user's interest. If not, the recent behavior dataset is searched for the user's most browsed items. If so, it is identified as the target product type of the user's interest. If not, the recent behavior dataset is searched for items that the user has added to the shopping cart, clicked on, or saved. If the target product type of the user's interest is not found in the recent behavior dataset, the search time range is expanded, and the above search logic is continued to search for the target product type of the user's interest in older behavior datasets (for example, if the default recent behavior dataset is behavior datasets within three months, the older behavior dataset can be expanded to behavior datasets within six months) until the target product type of the user's interest is found. In one embodiment, the user's interest in one or more target product types is one or more.

[0094] Step S204 : Based on the target product type and the target question type, determine the product questions that each user terminal is concerned about, and associate each user terminal identifier with the product questions that each user terminal is concerned about and store them in a cache database.

[0095] Specifically, first count the proportion of the number of questions of each target question type in the recent consultation question set, then search in the product consultation data based on the target product type to obtain the product consultation question set corresponding to the target product type, and then filter out a preset first number of product consultation questions from the product consultation question set according to the N target question types and the proportion of the number of questions corresponding to each target question type as the product questions that each user terminal is concerned about. The first number is the number of questions included in the predicted question list displayed on the user terminal. For example, assuming that the first number is 10, the target product types are A and B, and the target question types are X, Y, and Z respectively; according to statistics, the proportion of the number of questions of the target question types X, Y, and Z in the recent consultation question set is 5:3:2; the product consultation question set corresponding to the target product type A or B contains a total of 100 consultation questions, then 5 consultation questions of question type X, 3 consultation questions of question type Y, and 2 consultation questions of question type Z are selected from the 100 consultation questions as the product questions that the user terminal is concerned about.

[0096] In one embodiment, each user identifier is associated with the product questions each user is interested in, forming a database of user-product question and answer data, which is stored in a cache database, such as a Redis cache database. The user identifier may include one or more pieces of information, such as the user ID, the user's IP address, and the user's geographic location. Storing these multiple user identifiers in association with the product questions and responses to the user's questions enables expanded recommendations based on anticipated questions.

[0097] In other embodiments, each user terminal identifier is associated with each user terminal's product questions and responses, forming user-product related question and answer data that is stored in a cache database. Storing the responses to the user terminal's product questions in association allows the user to quickly retrieve the response from the cache database and return it to the user terminal when they click on a question in the predicted question list.

[0098] Step S205 , after detecting that the user terminal has entered the customer service system, the product issues that the user terminal is concerned about are sent to the user terminal, so that the user terminal generates a predicted question list based on the product issues that the user terminal is concerned about for display.

[0099] Specifically, when a user has a need to consult about a product and opens the customer service system dialogue window, a monitoring event of the front-end button will be triggered. This event will carry the user terminal identifier in the context (such as the user terminal ID, user terminal IP, user terminal geographic location information, etc.). The user terminal initiates a pre-judgment query request to the customer service system based on the user terminal identifier. The customer service system queries the user and product-related question and answer data in the Redis database based on the user terminal identifier, obtains the product questions that the user terminal is concerned about, and sends it to the user terminal. The user terminal generates a pre-judgment question list based on the product questions that the user terminal is concerned about and displays it in the drop-down list of the dialogue window. For example, the pre-judgment question list formed by the product questions that a user is concerned about is as follows:

[0100] Question 1: Does phone A support wireless charging?

[0101] Question 2: Can I replace the memory card after purchasing phone A?

[0102] Question 3: How to transfer contacts from old phone to new phone A?

[0103] Question 4: How long is the battery life of mobile phone A?

[0104] Question 5: Does phone A have dual SIM dual standby function?

[0105] Question 6: Does earphone B support Bluetooth 5.0?

[0106] Question 7: Is the B wireless headset waterproof?

[0107] Question 8: Does the charging box of earphone B support fast charging?

[0108] Question 9: Does earphone B support noise reduction?

[0109] Question 10: How to reset the B wireless earphones?

[0110] After the user sees the predicted question list, if there is a question he wants to ask in the predicted question list, he can click on the question to display the question in the chat dialog box and give a prompt that it is being queried. After that, the customer service system obtains the corresponding reply answer from the user and product related question and answer data in the cache database based on the question and returns it to the chat window. In this way, the user's question is predicted and responded in advance, and the user's problem is solved in an efficient manner.

[0111] When there is no question that the user wants to ask in the predicted question list, the user will continue to manually enter the question content in the input box. In order to further guide the customer to ask questions, the embodiment of the present application also provides a method for guiding the question in the input question, such as Figure 3 As shown, the question guiding method includes:

[0112] Step S301: When it is detected that the user terminal manually inputs a question in the customer service system, the manual input method and / or manual input content are obtained.

[0113] Among them, the manual input method mainly includes the pause duration of manual input.

[0114] For the stage of inputting questions, this application provides an intelligent automatic completion function. The user's input content is completed according to three levels of completion modes based on the user's manual input method and manual input content in the input box. In order to cope with the data required for different levels of completion modes, the method of this application also performs the following data preparation: (1) Use a clustering algorithm to extract representative consulting questions from the consulting questions contained in the product consulting data as product FAQs, and then store the product FAQs and replies to form product FAQ data in the cache database; (2) Merge the customer service question and answer data and the product consulting data based on the questions to generate all question and answer data, and store all question and answer data in the cache database and the vector database respectively. Specifically, all custom questions are extracted from the customer service question and answer data and all consulting questions are extracted from the product consulting data. After all custom questions and all consulting questions are merged, they are deduplicated, clustered, and processed to obtain a list of all questions. All question and answer data are generated based on the list of all questions and the answers corresponding to each question in the list of all questions.

[0115] Step S302: When the manually input pause duration is greater than the first preset pause duration, the first level completion mode is entered.

[0116] When the pause duration of the manual input is longer than the first preset pause duration, the current completion mode is determined to be the first-level completion mode, and the first-level completion mode is entered. In the first-level completion mode, the customer service system extracts keywords from the manually input content, matches the keywords with all the question and answer data in the cache database, and generates a first-level completion question list based on a number of successfully matched questions. Assuming that the first-level completion question list contains at most a preset second number of questions, when matching in the cache database through keywords, if the number of successfully matched questions is less than the second number, a first-level completion question list is generated based on all successfully matched questions; otherwise, a second number of questions are selected from the successfully matched questions to generate a first-level completion question list.

[0117] For example, when the user enters "EP800 charging" in the input box and pauses for 0.5 seconds, the list of first-level completion questions generated based on the keyword "EP800 charging" is as follows:

[0118] How fast does the EP800 charge?

[0119] How to use the EP800 charger correctly?

[0120] What safety precautions should I take when charging the EP800?

[0121] Can I use a power source while charging the EP800?

[0122] What voltage range does the EP800 charger support?

[0123] Will the battery life of the EP800 be affected when charging?

[0124] Will the EP800 automatically shut down after charging is complete?

[0125] What should I do if the EP800 overheats during charging?

[0126] How long does it take to fully charge the EP800?

[0127] Does the EP800 support fast charging technology?

[0128] Based on the above example, we can see that all questions in the first-level completion question list contain the words "EP800 charging" entered by the user.

[0129] Step S303, determine whether the first-level completion question list is empty, if so, proceed to step S304, if not, send the first-level completion question list to the user end.

[0130] When the first-level completion question list is not empty, the first-level completion question list is sent to the user end, and the user end displays it in the drop-down list of the input box. Figure 4 As shown, after the customer enters "AC18" in the input box, the customer service system background matches the keyword "AC18" to obtain two completion questions: "How to charge AC180?" and "Can AC180 be used with B230?" The user terminal displays the two completion questions in the drop-down list of the input box. If the drop-down list contains the question the user wants to ask, the user can select and click on the question. After clicking, the original content in the input box is cleared, and the selected question is filled in the input box. After the user clicks Submit, the customer service system searches for the corresponding answer based on the question in all the question and answer data in the cache database and returns it to the user terminal. Since the questions in the first-level completion question list are all from the questions in the entire question and answer data, the correct answer can usually be accurately recalled, and then the correct answer is displayed in the chat window with the user.

[0131] Step S304: When the manually input pause duration is greater than the second preset pause duration, the secondary completion mode is entered.

[0132] When the customer service system lacks questions that match the keywords, or the user inputs too much content at one time (no pause, no triggering of the first-level completion condition), the first-level completion question list will be empty when the keyword method is used. In this case, the second-level completion mode is used. The conditions that must be met to enter the second-level completion mode include: (1) detecting that the user manually enters a question in the customer service system (i.e., the input box contains input content); (2) the first-level completion question list is empty; (3) the pause duration of the manual input is greater than the second preset pause duration, and the second preset pause duration is greater than the first preset pause duration. For example, the first preset pause duration is 0.5s, and the second preset pause duration is 1s.

[0133] In the secondary completion mode, the customer service system obtains the manual input content of the user in the input box, matches the manual input content with all the question and answer data in the vector database for similarity, and selects several questions with similarity values ​​exceeding the preset similarity threshold (for example, the similarity threshold is 30%) and with high similarity ranking to generate a secondary completion question list.

[0134] For example, when the user enters "EP800 charging time" in the input box and pauses for 1 second, the list of secondary completion questions generated based on the input content "EP800 charging time" is as follows:

[0135] How long does it take to fully charge the EP800?

[0136] What is the maximum charging time of the EP800?

[0137] How long does it take to fully charge the EP800?

[0138] How long does it take to charge the EP800?

[0139] How long does it take to fully charge the EP800 under standard conditions?

[0140] How long does it take to charge the EP800?

[0141] How long does it take to charge the EP800 from 0% to 100%?

[0142] How long does it take to fully charge the EP800 battery?

[0143] What factors affect the charging time of EP800?

[0144] What is the shortest charging time for EP800?

[0145] Based on the above example, we can see that the questions in the secondary completion question list and the user's input content may not be exactly the same in terms of text, but they express similar meanings.

[0146] Step S305 , determining whether the secondary completion question list is empty, if so, proceeding to step S306 , if not, sending the secondary completion question list to the user end.

[0147] When the secondary completion question list is not empty, it is sent to the user and displayed in the drop-down list of the input box. The user determines whether they have a question based on the secondary completion question list. If so, they select the question with the mouse. After selection, the complete question is filled in the input box. After the user clicks Submit, the customer service system searches the entire Q&A data in the vector database for the corresponding answer based on the question and returns it to the user. Since the questions in the secondary completion question list are all from the entire Q&A data, the correct answer can usually be accurately recalled and then displayed in the chat window with the user.

[0148] Step S306: When the manually input content does not contain the product type, the third-level completion mode is entered.

[0149] When both the first-level completion mode and the second-level completion mode fail to find relevant questions, the third-level completion mode can be used. The conditions for entering the third-level completion mode include: (1) detecting that the user manually enters a question in the customer service system (i.e., the input box contains input content); (2) the second-level completion question list is empty; (3) the manually entered content does not contain the product type.

[0150] In the three-level completion mode, the customer service system queries the product FAQ data in the cache database, obtains several top-ranked questions, generates a three-level completion question list, and sends it to the user end.

[0151] For example, the list of three-level completion questions is as follows:

[0152] How to improve the service life of energy storage batteries?

[0153] What suggestions are there for the maintenance of energy storage batteries?

[0154] What are the common faults of energy storage systems and how to troubleshoot them?

[0155] What types of energy storage equipment are there?

[0156] What should I pay attention to when purchasing energy storage products?

[0157] What is the difference between energy storage batteries and traditional batteries?

[0158] How to determine whether the energy storage device needs to replace the battery?

[0159] What should be done when energy storage equipment fails?

[0160] What are the precautions for safe use of energy storage equipment?

[0161] How to store energy for a long time?

[0162] Based on the above example, we can see that the questions in the three-level completion question list are independent of product type and are universal questions applicable to all products. Based on the three-level completion question list, the user determines whether they have a question they would like to ask. If so, they select the question with the mouse. After selection, the complete question is entered into the input box. After the user clicks Submit, the customer service system searches the cached product FAQ database for the corresponding answer based on the question and returns it to the user. Since the questions in the three-level completion question list are all from the product FAQ database, the correct answer can usually be accurately recalled and then displayed in the chat window with the user.

[0163] If the user still cannot find the question they want to ask in the third-level completion question list, they can ignore these question lists and continue to enter the custom question in the input box and submit it.

[0164] In one embodiment, the question guiding method of the embodiment of the present application further includes: counting the number of conversations and the number of messages before and after the execution of the question guiding strategy, calculating the efficiency index before and after the execution of the question guiding strategy based on the number of conversations and the number of messages before and after the execution of the question guiding strategy, and evaluating whether the question guiding strategy is effective based on the change in the efficiency index before and after the execution of the question guiding strategy. Specifically, the efficiency index E is defined as the number of messages / the number of conversations. Assume that the efficiency index after the execution of the question guiding strategy is , the efficiency index before executing the question-guiding strategy is ,when When , that is, after the question-guiding strategy is executed, the number of messages in each round of dialogue on the user side is reduced compared with that before the question-guiding strategy, it means that the question-guiding strategy is effective and the communication efficiency of the customer service system is improved; when When , that is, the number of messages in each round of conversation on the user side increases after the question guidance strategy is executed compared with that before the question guidance, it means that the question guidance strategy fails to effectively improve efficiency and needs further optimization.

[0165] In other embodiments, the question guidance method of the embodiment of the present application also includes: counting the number of times the predicted question list, the first-level completed question list, the second-level completed question list and the third-level completed question list are triggered, and counting the number of times the questions in the predicted question list, the first-level completed question list, the second-level completed question list and the third-level completed question list are selected, and evaluating and optimizing the screening strategy for questions in each category of lists based on the number of times the questions in each category of lists are selected and the number of times the list is triggered within a preset time period (for example, one month).

[0166] The question guiding method provided in the embodiment of the present application first determines the target question type that the user terminal is concerned about based on the user terminal's recent consulting question set, and determines the target commodity type that the user terminal is concerned about based on the user terminal's recent behavior data set; then, based on the target commodity type and target question type, determines the commodity issues that each user terminal is concerned about, associates each user terminal identifier with the commodity issues that each user terminal is concerned about and stores them in a cache database; finally, when it is detected that the user terminal has entered the customer service system, the commodity issues that the user terminal is concerned about are sent to the user terminal, so that the user terminal generates a predicted question list based on the commodity issues that the user terminal is concerned about for display. The method of the present application first analyzes the commodity issues that the user is concerned about based on big data technology before the user asks a question, intercepts the question at the beginning of the question, saves the user's time in organizing the question, and improves customer service efficiency and satisfaction.

[0167] According to an embodiment of the present application, a customer service server is provided, such as Figure 5 FIG. 1 is a schematic diagram of the hardware structure of a customer service server provided in an embodiment of the present application. The customer service server 100 includes a processor 10, a memory 20, and a communication interface 30. The processor 10, the memory 20, and the communication interface 30 are connected via a line. Figure 5 In the illustrated embodiment, the processor 10 , the memory 20 , and the communication interface 30 are communicatively connected to each other via a bus.

[0168] The memory 20 is used to store software programs, computer-executable program instructions, etc. The memory 20 may include a program storage area and a data storage area. The program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the customer service server.

[0169] The memory 20 may be a read-only memory (ROM), or other types of static storage devices that can store static information and instructions; a random access memory (RAM), or other types of dynamic storage devices that can store information and instructions; or an electrically erasable programmable read-only memory (EEPROM), and the specifics are not limited here.

[0170] For example, the memory 20 may be a double data rate synchronous dynamic random access memory (DDR SDRAM). The memory 20 may exist independently but be connected to the processor 10. Alternatively, the memory 20 may be integrated with the processor 10, for example, within one or more chips.

[0171] In some embodiments, the memory 20 may include a memory remote from the processor 10, which may be connected to the customer service server via a network. Examples of the network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0172] The processor 10 uses various interfaces and lines to connect various parts of the entire customer service server 100, and executes various functions and processes data of the customer service server by running or executing software programs stored in the memory 20, and calling data stored in the memory 20, such as implementing the method described in any embodiment of the present application.

[0173] The processor 10 may be a field programmable gate array (FPGA), a digital signal processor (DSP), a central processing unit (CPU), or the like.

[0174] The processor 10 can be a single-core processor or a multi-core processor. For example, the processor 10 can be composed of multiple FPGAs or multiple DSPs. In addition, the processor 10 can refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions). The processor 10 can be a separate semiconductor chip or integrated into a semiconductor chip together with other circuits. For example, it can form a system on a chip (SoC) with other circuits (such as codec circuits, hardware acceleration circuits, or various bus and interface circuits). Alternatively, it can be integrated into an application-specific integrated circuit (ASIC) as a built-in processor of the ASIC. The ASIC with the integrated processor can be packaged separately or together with other circuits.

[0175] The communication interface 30 may use a transceiver device such as a transceiver to implement communication between the customer service server and other devices or a communication network.

[0176] An embodiment of the present application further provides a non-volatile computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, which are executed by one or more processors, for example, to execute the steps of the question-guiding method described above.

[0177] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Based on the concept of the present application, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the present application as described above. For the sake of simplicity, they are not provided in detail. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A question guiding method, characterized in that: The method comprises: Obtain the recent consultation question set and recent behavior data set of each user terminal; Based on the recent consultation question set, determining the target question type that each user terminal is concerned about; Based on the recent behavior data set, determining the target product type that each user terminal is interested in; Based on the target product type and the target question type, determining the product questions that each user terminal is concerned about, and associating each user terminal identifier with the product question and reply answer to form user-product associated question and answer data and storing it in a cache database; After detecting that the user terminal has entered the customer service system, the product issues that the user terminal is concerned about are sent to the user terminal, so that the user terminal generates a list of predicted issues based on the product issues that the user terminal is concerned about for display; Obtaining customer service question-and-answer data and product inquiry data, merging the customer service question-and-answer data and the product inquiry data based on the questions to generate complete question-and-answer data, and storing the complete question-and-answer data in the cache database and the vector database, respectively. The customer service question-and-answer data includes multiple customized questions and corresponding customized answers, and the product inquiry data includes the inquiry questions and corresponding responses of each user terminal based on each product type; Clustering the consultation questions in the product consultation data to generate product FAQs, and storing the product FAQs and corresponding answers into product FAQ data in the cache database When detecting that the user terminal manually inputs a question in the customer service system, determining a completion mode based on the manual input method and / or the manual input content; If the completion mode is level one completion, keywords are extracted from the manually input content, the keywords are matched with all question and answer data in the cache database, and a level one completion question list is generated based on a number of successfully matched questions and sent to the user terminal; If the completion mode is secondary completion, the manually input content is matched with all question and answer data in the vector database for similarity, and a secondary completion question list is generated based on a number of successfully matched questions and sent to the user terminal; If the completion mode is three-level completion, a number of top-ranked questions are selected from the product FAQ data in the cache database to generate a three-level completion question list and send it to the user terminal; Determining the completion mode based on the manual input method and / or the manual input content includes: When detecting that the pause duration of the manual input is longer than the first preset pause duration, determining the completion mode to be the first-level completion; When it is detected that the manually input pause duration is longer than a second preset pause duration and the first-level completion question list is empty, determining that the completion mode is the second-level completion mode, and the second preset pause duration is longer than the first preset pause duration; When it is detected that the manually input content does not contain the product type and the second-level completion question list is empty, the completion mode is determined to be the third-level completion.

2. The method according to claim 1, characterized in that The method further comprises: Input the consulting questions collected by the customer service system into the trained question bias classification model for type labeling to obtain the question type of the consulting question; Determining the target question type that the user is concerned about based on the recent consultation question set includes: Statistics are collected on each consulting question in the recent consulting question set according to question type, and N question types with the highest number of questions are selected as target question types of interest to the user terminal, where N is an integer greater than or equal to 1.

3. The method according to claim 2, characterized in that The determining of the product issues that each user terminal is concerned about based on the target product type and the target issue type includes: Count the proportion of each target question type in the recent consultation question set; Obtain a set of product consultation questions corresponding to the target product type; According to the N target question types and the proportion of the number of questions corresponding to each target question type, a preset first number of product consulting questions are screened from the product consulting question set as the product questions that each user terminal is concerned about.

4. The method according to claim 1, wherein The performing similarity matching between the manually input content and all question-answer data in the vector database, and generating a secondary completion question list based on a number of successfully matched questions and sending the list to the user terminal comprises: The manually input content is matched with all question and answer data in the vector database for similarity, and a number of questions whose similarity values ​​exceed a preset similarity threshold and whose similarity values ​​are ranked high are selected to generate a secondary completion question list.

5. The method according to any one of claims 1 to 4, characterized in that The method further comprises: Count the number of conversations and messages before and after the implementation of the question-guided strategy; The efficiency index before and after the implementation of the question-guiding strategy is calculated based on the number of conversations and messages before and after the implementation of the question-guiding strategy; The effectiveness of the question-guiding strategy was evaluated based on the changes in efficiency indicators before and after the implementation of the question-guiding strategy.

6. A customer service server, characterized in that: The method comprises at least one processor and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

7. A computer storage medium, characterized in that The computer storage medium stores instructions or programs, and when the instructions or programs are executed by at least one processor, the at least one processor is caused to perform the method according to any one of claims 1 to 5.

Images