Intelligent customer service management method and system based on multi-channel integration

Abstract

The present invention relates to the related technical field of customer service management, and disclosed are an intelligent customer service management method and system based on multi-channel integration. The method comprises: receiving order failure information of an order system, querying an order failure log, and extracting and transmitting order information to a task queue; monitoring the task queue, executing text information capture, and performing homogeneity analysis to generate heterogeneous information; matching a transmission information library, and determining transmission information; determining an information transmission type, and calling a transmission interface; and processing the transmission information by connecting a collaborative system and the transmission interface. The present invention solves the technical problems in existing customer service management that multi-channel data cannot be effectively integrated, the matching degree between abnormal data and a communication scheme is poor, and a solution cannot be rapidly generated due to the absence of unified management, leading to low customer service management efficiency.

Description

Customer service intelligent management method and system based on multi-channel integration TECHNICAL FIELD

[0001] The present application relates to the technical field of customer service management, in particular to a customer service intelligent management method and system based on multi-channel integration. BACKGROUND

[0002] With the rapid development of the Internet and mobile communication technology, the interaction between users and enterprises has become diversified, including telephone, email, social media, instant messaging and other channels. The multi-channel integrated customer service intelligent management system is particularly important. Multi-channel integrated customer service intelligent management provides the ability to integrate various channel data for enterprises. However, traditional customer service management systems often only support a single or a few communication channels, resulting in information silos, poor communication, and difficult data analysis, and many other challenges, making the entire system's workflow cumbersome and inefficient.

[0003] Therefore, at the present stage, there are technical problems in the related technology of customer service management, such as the inability to effectively integrate multi-channel data, poor matching of abnormal data and communication solutions, the inability to uniformly manage and quickly generate solutions, and thus low efficiency of customer service management. TECHNICAL PROBLEM

[0004] The present application provides a customer service intelligent management method and system based on multi-channel integration, which uses text capture, homogeneity analysis, interface calling and information processing techniques to solve the technical problems of existing customer service management, such as the inability to effectively integrate multi-channel data, poor matching of abnormal data and communication solutions, the inability to uniformly manage and quickly generate solutions, and thus low efficiency of customer service management. TECHNICAL SOLUTION

[0005] The present application provides a customer service intelligent management method based on multi-channel integration, which is applied to a customer service intelligent management system based on multi-channel integration. The system is in communication connection with an order system, an abnormal order system, a collaboration system and a customer service system. The method comprises: receiving order failure information of the order system, querying order failure logs, extracting order information and sending it to a task queue; based on the abnormal order system, listening to the task queue and performing text information capture, performing homogeneity analysis on the text information through a text analysis channel, and generating heterogeneous information; based on the heterogeneous information, matching a sending information database to determine sending information; determining the type of information sending according to the sending information, calling a sending interface; and processing the sending information through the sending interface of the collaboration system.

[0006] In a possible implementation, the text information is analyzed homogeneously through a text analysis channel, and the following processing is performed: the text analysis channel includes a first text analysis branch, a second text analysis branch, and a fully connected layer text analysis branch connected in parallel; text span analysis is performed on the text information through the first text analysis branch to obtain P text span sets, where P is a positive integer; the second text analysis branch receives the P text span sets returned by the first text analysis branch, performs time window configuration of the text span on the task queue, and based on the configured time window, P text span thickness sets are retrieved based on the task queue; through the fully connected layer text analysis branch, span information homogeneity analysis is performed based on the P text span thickness sets to generate homogeneous information and the heterogeneous information.

[0007] In a possible implementation, the span information homogeneity analysis is performed based on the P text span thickness sets through the fully connected layer text analysis branch to generate homogeneous information and the heterogeneous information, and the following processing is performed: a first text span thickness set is obtained based on the P text span thickness sets; the first text span thickness set is received by the fully connected layer text analysis branch to perform thickness characteristic homogeneity judgment, if the first text span thickness set satisfies thickness characteristic homogeneity, the first text span thickness set is stored in the homogeneous information; if the first text span thickness set does not satisfy thickness characteristic homogeneity, asynchronous thickness collection is performed, and the heterogeneous information is determined based on the asynchronous thickness.

[0008] In a possible implementation, the heterogeneous information is determined based on the asynchronous thickness, and the following processing is further performed: a heterogeneous threshold is preset; the asynchronous thickness is compared with the heterogeneous threshold, if the asynchronous thickness is less than or equal to the heterogeneous threshold, the homogeneous information is confirmed by analogy for the first text span thickness set; if the asynchronous thickness is greater than the heterogeneous threshold, the heterogeneous information is confirmed for the first text span thickness set.

[0009] In a possible implementation, the sending information is determined based on the matching of the heterogeneous information and a sending information library, and the following processing is further performed: the sending information library is retrieved, including an association relationship between a heterogeneous information span and a sending information span; the sending information span is determined based on matching of the heterogeneous information span in the sending information library based on the input of the heterogeneous information; and the sending information is determined by filling the sending information span based on the homogeneous information confirmation by analogy.

[0010] In a possible implementation, in addition to training the full-connection layer text analysis branch, the following processing is also performed: constructing a network architecture of the full-connection layer text analysis branch, receiving output data of the first text analysis branch and the second text analysis branch connected in parallel; obtaining sample training data, including a plurality of sample text span thickness sets, sample homogeneous information and sample heterogeneous information; and based on the sample training data, training the full-connection layer text analysis branch.

[0011] In a possible implementation, in addition to processing the sending information through the sending interface by means of the collaborative system, the following processing is also performed: receiving the sending information by means of the collaborative system through the customer service system; based on the sending information, solving the call, generating a solution processing scheme for a call receiver; and establishing a customer communication interface according to the solution processing scheme, and sending the solution processing scheme.

[0012] The application further provides a customer service intelligent management system based on multi-channel integration, which is in communication connection with an order system, an abnormal order system, a collaborative system and a customer service system, and comprises:

[0013] An order information extraction module is configured to receive order failure information of the order system, query order failure logs, extract order information and send the order information to a task queue.

[0014] A text information homogeneous analysis module is configured to listen to the task queue based on the abnormal order system, perform text information capture, perform homogeneous analysis on the text information through a text analysis channel and generate heterogeneous information.

[0015] A sending information database matching module is configured to match a sending information database based on the heterogeneous information and determine sending information.

[0016] An information sending type judgment module is configured to judge an information sending type according to the sending information and call a sending interface.

[0017] A sending information processing module is configured to process the sending information through the sending interface by means of the collaborative system.

[0018] The multi-channel integrated customer service intelligent management method and system provided in the application receives failure order information of an order system, queries order failure logs, extracts order information and sends the order information to a task queue; listens to the task queue, performs text information capture, performs notification analysis, generates heterogeneous information, and matches a sending information database to determine sending information; judges the information sending type, calls a sending interface, and processes the sending information through a collaborative system to interface the sending interface, thereby solving the technical problems of the existing customer service management, such as the inability to effectively integrate multi-channel data, the poor matching degree of abnormal data and communication solutions, the inability to uniformly manage and quickly generate solutions, and the low efficiency of customer service management, and achieving the technical effects of intelligently and effectively integrating multi-channel data, providing better data analysis, and improving the work efficiency of customer service management. Advantages

[0019] The multi-channel integrated customer service intelligent management method and system provided in the application receives failure order information of an order system, queries order failure logs, extracts order information and sends the order information to a task queue; listens to the task queue, performs text information capture, performs notification analysis, generates heterogeneous information, and matches a sending information database to determine sending information; judges the information sending type, calls a sending interface, and processes the sending information through a collaborative system to interface the sending interface, thereby solving the technical problems of the existing customer service management, such as the inability to effectively integrate multi-channel data, the poor matching degree of abnormal data and communication solutions, the inability to uniformly manage and quickly generate solutions, and the low efficiency of customer service management, and achieving the technical effects of intelligently and effectively integrating multi-channel data, providing better data analysis, and improving the work efficiency of customer service management. BRIEF DESCRIPTION OF DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present disclosure, the drawings of the embodiments of the present disclosure will be briefly introduced below. In the present application, a flowchart is used to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0021] FIG. 1 is a flowchart of a multi-channel integrated customer service intelligent management method according to an embodiment of the present application;

[0022] FIG. 2 is a structural diagram of a multi-channel integrated customer service intelligent management system according to an embodiment of the present application.

[0023] Reference signs: order information extraction module 10, text information homogeneity analysis module 20, sending information database matching module 30, information sending type judgment module 40, and sending information processing module 50. DETAILED DESCRIPTION

[0024] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described.

[0025] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the drawings. The described embodiments should not be regarded as limiting the present application, and all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0026] In the following description, "some embodiments" are related to a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict. The term "first\second" referred to is only to distinguish similar objects, and does not represent a specific order of the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0027] The embodiments of the present application provide a customer service intelligent management method based on multi-channel integration, which is applied to a customer service intelligent management system based on multi-channel integration, as shown in FIG. 1, and the method comprises the following steps.

[0028] The customer service intelligent management system based on multi-channel integration is in communication connection with an order system, an abnormal order system, a collaboration system and a customer service system, wherein the order system generates order failure data; the abnormal order system is used for receiving abnormal data and judging types, calling interfaces; the system is used for sending messages to customer service personnel; and the customer service system actively contacts technical personnel and consumers to solve complaint problems. In the order system, if an abnormality occurs in the ordering process, the order information and related template configuration data are asynchronously sent to an asynchronous task queue, the abnormal order system listens to the asynchronous queue, once there is abnormal data, the abnormal order system actively receives the abnormal data, then parses the message template, the message type and the message receiver from the abnormal message, and then calls the corresponding sending message interface, and then calls the collaboration system to send a message to notify the customer service personnel, and the customer service personnel actively contacts the technical personnel to handle the abnormal order and contacts the consumer to solve the complaint problem after receiving the reminder message.

[0029] Step S100, receiving order system order failure information, querying order failure log, extracting order information and sending to task queue. The order failure information is received from the order system, and the order failure log is queried to extract the order information and send it to the task queue. Specifically, when the order system fails, for example, payment cancellation, payment failure, insufficient inventory, etc., the system will generate order failure information in the form of logs, message queues, etc. between systems. Query the order failure log to find the corresponding failure order log record, including order number, failure reason, timestamp, etc. Among them, the order failure log is a file that records exceptions or errors that occur during order processing, containing detailed error information and processing flow. By analyzing the order failure log, relevant order information such as order number, customer information, order content, failure reason, order amount, etc. is extracted from the order failure log. The extracted order information is encapsulated into a task and sent to the task queue as a new task for subsequent processing. The task queue is a message middleware or asynchronous task queue used to store and distribute data that needs to be processed, enabling asynchronous processing of order failure information and ensuring quick response and timely processing of failed orders.

[0030] After the order information is sent to the task queue, step S200 is executed, the abnormal order system listens to the task queue based on the abnormal order system, and performs text information capture, and performs homogeneous analysis on the text information through the text analysis channel. Generate heterogeneous information. Specifically, the abnormal order system continuously monitors the task queue and monitors whether new tasks or abnormal order data arrive in real time. When the abnormal order system listens to the task queue and finds a task with abnormal order data, it will capture the text information in the task, which usually includes order details, customer messages, and abnormal description information. The captured text information is processed through the text analysis channel, such as word segmentation, semantic analysis, entity recognition, etc. The abnormal order system performs homogeneous analysis on the text information processed by the text analysis channel. Homogeneous analysis refers to classifying, clustering, etc. Similar text information is classified into a category, and heterogeneous information is generated. Heterogeneous information refers to text information of different categories, such as summaries of order processing problems, abnormal processing suggestions, and consumer feedback. By listening to the task queue through the abnormal order system, the relevant text information of the abnormal order is obtained, and the text analysis channel is analyzed and processed to generate heterogeneous information for subsequent processing.

[0031] In a possible implementation, the step S200 performs homogeneous analysis on the text information through a text analysis channel, further comprising a step S210, wherein the text analysis channel comprises a first text analysis branch, a second text analysis branch and a fully connected layer text analysis branch connected in parallel. The text analysis channel is constructed by connecting the first text analysis branch, the second text analysis branch and the fully connected layer text analysis branch in parallel, which are respectively used for text analysis processing at different levels. The first text analysis branch is the first branch in the text analysis channel, responsible for performing basic text analysis tasks, such as rule-based syntax analysis, semantic analysis and keyword extraction, etc. The second text analysis branch is the second branch in the text analysis channel, which adopts different text analysis techniques to capture higher-level semantic features and context analysis, such as natural language processing (NLP). The fully connected layer text analysis branch is the third branch in the text analysis channel, which can be a specific deep learning model, such as a multi-layer perceptron, a deep neural network, etc., used to integrate and analyze the features extracted from the text, so as to better understand the text information. Further comprising a step S220, performing text span analysis on the text information through the first text analysis branch to obtain P text span sets, wherein P is a positive integer. The first text analysis branch performs text span analysis and obtains P text span sets. Text span analysis is a text processing technique aimed at extracting relevant phrases or fragments from text, which can be defined according to specific semantic or structural standards. For example, in a question and answer system, the answer to a question can be defined as a text span. Through text span analysis, the text information is divided into multiple spans, each of which represents a text fragment related to the target. Further comprising a step S230, based on the second text analysis branch receiving the P text span sets returned by the first text analysis branch, performing time window configuration of text span on the task queue, and based on the task queue, calling P text span thickness sets according to the configured time window. The second text analysis branch receives the P text span sets returned by the first text analysis branch, performs time window configuration of text span, and schedules and processes the text span sets, wherein the time window configuration refers to obtaining many text information of the same kind in a historical time period. After configuring the time window, P text span thickness sets are called according to the defined time window and task queue, to ensure that the data in the text span set is processed within an appropriate time period. Further comprising a step S240, performing span information homogeneous analysis on the P text span thickness sets through the fully connected layer text analysis branch to generate homogeneous information and the heterogeneous information.The text span thickness set refers to a set containing P text spans, each span representing a segment or interval in the text, and the span information homogeneity analysis refers to analyzing the similarity or consistency between the spans, the homogeneous information refers to the text information with high similarity or consistency, and the heterogeneous information refers to the text information with difference or inconsistency.

[0032] In a possible implementation, step S240 performs span information homogeneity analysis on the P text span thickness sets through the full connection layer text analysis branch to generate homogeneous information and the heterogeneous information, and further includes step S241, obtaining a first text span thickness set from the P text span thickness sets. The first text span thickness set is a text span thickness set randomly selected from the P text span thickness sets. Step S242 further includes receiving the first text span thickness set through the full connection layer text analysis branch to perform thickness characteristic homogeneity judgment, and if the first text span thickness set satisfies the thickness characteristic homogeneity, storing the first text span thickness set to the homogeneous information. The thickness characteristic homogeneity judgment refers to analyzing the first text span thickness set to determine whether it meets the homogeneity characteristic, and evaluating the similarity and relevance between the text spans. If the first text span thickness set is judged to meet the homogeneity characteristic, it will be stored as homogeneous information. Step S243 further includes, if the first text span thickness set does not satisfy the thickness characteristic homogeneity, performing asynchronous thickness collection, and determining the heterogeneous information based on the asynchronous thickness. Specifically, asynchronous method is adopted to obtain more text spans, to ensure more extensive information, which may involve different data sources, different text processing techniques or feature extraction methods. When the text spans of asynchronous thickness are collected, they are analyzed to determine the heterogeneous information, including identifying the features different from the previous text spans, to ensure that the text data is more comprehensive and diversified, and to support more accurate decision-making.

[0033] In a possible implementation, the step S243 further includes a step S243-A of presetting a heterogeneity threshold. For example, when the preset heterogeneity threshold is 1, it means that each data is different. The step S243 further includes a step S243-B of comparing the asynchronous thickness with the heterogeneity threshold, and if the asynchronous thickness is less than or equal to the heterogeneity threshold, performing the homogeneity information analogy confirmation on the first text span thickness set. Specifically, the asynchronous thickness is compared with the heterogeneity threshold, and if the asynchronous thickness is less than or equal to the preset heterogeneity threshold, the homogeneity information analogy confirmation is performed on the first text span thickness set. In this way, the asynchronous data is compared with the first text span thickness set to confirm that the similar customer information is compared. The step S243 further includes a step S243-C of performing the heterogeneity information confirmation on the first text span thickness set if the asynchronous thickness is greater than the heterogeneity threshold. Specifically, if the asynchronous thickness is greater than the preset heterogeneity threshold, the heterogeneity information confirmation is performed on the first text span thickness set, that is, the asynchronous data is compared with the first text span thickness set to confirm the difference between them, so as to achieve more detailed classification and confirmation of the text information and better understand the text information.

[0034] In a possible implementation, the step S240 of training the full connection layer text parsing branch further includes a step S245 of constructing a network architecture of the full connection layer text parsing branch, and receiving output data of the first text parsing branch and the second text parsing branch connected in parallel. Specifically, the network structure of the full connection layer text parsing branch is constructed, the level and connection mode of each branch are determined, and after the output of the first and second text parsing branches, a full connection layer is added to further process and integrate the text features extracted by the parallel connection. The step S246 of obtaining sample training data is further included, and the sample training data includes a plurality of sample text span thickness sets, sample homogeneity information and sample heterogeneity information. The step S247 of completing training of the full connection layer text parsing branch based on the sample training data is further included. Specifically, according to the classification task, the difference between the model output and the real label is measured, the loss function is defined, the sample data is input to calculate the loss function and update the parameters of the full connection layer text parsing branch through back propagation, the performance of the full connection layer text parsing branch is gradually optimized, and finally the full connection layer text parsing branch is trained.

[0035] After the heterogeneous information is generated, step S300 is performed to determine the sending information by matching the sending information database based on the heterogeneous information. Specifically, the sending information is determined by matching the sending information database based on the obtained heterogeneous information of different sources and different types, that is, the specific information to be sent is determined by searching and matching the sending information database according to the heterogeneous information. The sending information database refers to a database for storing various types of data to be sent. By matching the sending information database based on the heterogeneous information, the sending information is determined, which realizes the use of diversified data information to quickly make accurate information sending selection and improves the efficiency of the customer service management system.

[0036] In a possible implementation, step S300 of determining the sending information by matching the sending information database based on the heterogeneous information further includes step S310 of retrieving the sending information database including the association between the heterogeneous information span and the sending information span. Specifically, the heterogeneous information span refers to the span between information and data of different sources, and the length of each text in the text information refers to the text length. The sending information span refers to the range or span of the information to be sent. The association between the heterogeneous information span and the sending information span can help determine the effectiveness of text feature extraction and the reliability of the matching result, ensure that the feature representation between the heterogeneous information span and the sending information span is consistent, and facilitate effective matching. Step S320 includes inputting the heterogeneous information into the sending information database to match the heterogeneous information span and determine the sending information span based on the matching result. Step S330 includes filling the sending information span based on the homogenous information analogy confirmation to determine the sending information. By homogenous information analogy confirmation, the heterogeneous information is compared and matched with known homogenous information. According to the result of homogenous information category confirmation, similar customer information is found, and the sending information is finally determined. This facilitates subsequent communication with multiple customers having the same problem and problem solving.

[0037] After the sending information is determined, step S400 is performed to determine the information sending type according to the sending information and call the sending interface. According to the determined sending information, the information sending type is determined, for example, the sending message can be a telephone, an email, a WeChat message, a QQ message, etc. The different sending interfaces are called according to the determined information sending type, for example, if the sending information is a short message, a short message sending interface is called; if the sending information is a WeChat message, a WeChat message sending interface is called, which realizes the quick and effective processing of the sending request of different types of multi-channel information.

[0038] After determining the information sending type and calling the sending interface, step S500 is performed to process the sending information through the collaborative system interfacing the sending interface. By interfacing the sending interface through the collaborative system, the processing and sending of the sending information can be achieved. Specifically, the collaborative system can be integrated with the sending interface, and the processed information can be sent to the target recipient through the sending interface. The collaborative system can perform various processing on the sending information, such as format conversion, data arrangement, association matching, etc. The association matching may involve the association between the heterogeneous information span and the sending information span. According to specific requirements, the collaborative system can establish the association between them by matching the features, tags or other shared attributes between the heterogeneous information span and the sending information span. For example, in an application scenario, the collaborative system can extract keywords or tags from the heterogeneous information span and match them with the corresponding parts in the sending information span, which can determine the association between the heterogeneous information span and the sending information span, thereby ensuring the accuracy and integrity of the sending information. By interfacing the sending interface, the collaborative system can send the processed sending information to the target recipient, completing the entire sending process. This integration can improve the efficiency of the customer service management system and ensure the association between the sent information and the source data.

[0039] In a possible implementation, step S500 processes the sending information through the collaborative system interfacing the sending interface, and further includes step S510 of interfacing the customer service system with the collaborative system to receive the sending information. By interfacing the customer service system with the collaborative system, the sending information can be received and processed and responded accordingly. Specifically, the customer service system can be integrated with the collaborative system, and the information sent from the collaborative system can be transmitted to the customer service system. The customer service system can obtain the sending information in real time through the interface with the collaborative system and display it to the customer service personnel for viewing and processing. After receiving the sending information, the customer service system can classify, distribute, process and reply to the information according to specific business processes and rules. The customer service personnel can interact with the customer, provide support and answer questions based on the content of the sent information and related information. Further, step S520 is included to solve the call based on the sending information and generate a solution processing scheme for the call recipient. Specifically, the sending information is used to solve the call and interface with the call recipient (customer service personnel) to generate a solution processing scheme. Further, step S530 is included to establish a customer communication interface according to the solution processing scheme and send the solution processing scheme. A communication interface is established with the customer who generated the abnormal order, and the corresponding solution processing scheme is sent to solve the problem.

[0040] In the foregoing, the multi-channel integrated-based customer service intelligent management method according to the embodiment of the present application is described in detail with reference to FIG. 1. Next, the multi-channel integrated-based customer service intelligent management system according to the embodiment of the present application will be described with reference to FIG. 2.

[0041] The multi-channel integrated-based customer service intelligent management system according to the embodiment of the present application is used to solve the technical problem that the existing customer service management cannot effectively integrate multi-channel data, the matching degree of abnormal data and communication scheme is poor, and the unified management cannot quickly generate a solution, thereby leading to low efficiency of customer service management, so as to achieve the technical effects of intelligently and effectively integrating multi-channel data, providing better data analysis, and improving the work efficiency of customer service management. The multi-channel integrated-based customer service intelligent management system comprises an order information extraction module 10, a text information homogeneity analysis module 20, a sending information database matching module 30, an information sending type judgment module 40, and a sending information processing module 50.

[0042] The order information extraction module 10 is used to receive order failure information of an order system, query order failure logs, extract order information, and send the order information to a task queue.

[0043] The text information homogeneity analysis module 20 is used to listen to the task queue based on an abnormal order system, perform text information capture, perform homogeneity analysis on the text information through a text analysis channel, and generate heterogeneous information.

[0044] The sending information database matching module 30 is used to match a sending information database based on the heterogeneous information and determine sending information.

[0045] The information sending type judgment module 40 is used to determine the information sending type according to the sending information and call a sending interface.

[0046] The sending information processing module 50 is used to process the sending information through a sending interface of a collaborative system.

[0047] In the following, the specific configuration of the text information homogeneity analysis module 20 will be described in detail. As described above, the text information is analyzed by the text parsing channel to generate the heterogeneous information, and the text information homogeneity analysis module 20 can further include: the text parsing channel includes a first text parsing branch, a second text parsing branch and a fully connected layer text parsing branch connected in parallel; the text span set P is obtained by executing the text span analysis on the text information through the first text parsing branch, wherein P is a positive integer; the P text span sets returned by the first text parsing branch are received based on the second text parsing branch, the time window configuration of the text span is executed on the task queue, and the P text span thickness sets are called based on the task queue according to the configured time window; the span information homogeneity analysis is performed based on the P text span thickness sets through the fully connected layer text parsing branch to generate the homogeneity information and the heterogeneous information.

[0048] In the following, the specific configuration of the text information homogeneity analysis module 20 will be described in detail. As described above, the text information is analyzed by the text parsing channel to generate the heterogeneous information, and the text information homogeneity analysis module 20 can further include: the text parsing channel includes a first text parsing branch, a second text parsing branch and a fully connected layer text parsing branch connected in parallel; the text span set P is obtained by executing the text span analysis on the text information through the first text parsing branch, wherein P is a positive integer; the P text span sets returned by the first text parsing branch are received based on the second text parsing branch, the time window configuration of the text span is executed on the task queue, and the P text span thickness sets are called based on the task queue according to the configured time window; the span information homogeneity analysis is performed based on the P text span thickness sets through the fully connected layer text parsing branch to generate the homogeneity information and the heterogeneous information.

[0049] In the following, the specific configuration of the text information homogeneity analysis module 20 will be described in detail. As described above, the text information is analyzed by the text parsing channel to generate the heterogeneous information, and the text information homogeneity analysis module 20 can further include: the text parsing channel includes a first text parsing branch, a second text parsing branch and a fully connected layer text parsing branch connected in parallel; the text span set P is obtained by executing the text span analysis on the text information through the first text parsing branch, wherein P is a positive integer; the P text span sets returned by the first text parsing branch are received based on the second text parsing branch, the time window configuration of the text span is executed on the task queue, and the P text span thickness sets are called based on the task queue according to the configured time window; the span information homogeneity analysis is performed based on the P text span thickness sets through the fully connected layer text parsing branch to generate the homogeneity information and the heterogeneous information.

[0050] In the following, the specific configuration of the text information homogeneity analysis module 20 will be described in detail. As described above, the text information is analyzed by the text parsing channel to generate the heterogeneous information, and the text information homogeneity analysis module 20 can further include: the text parsing channel includes a first text parsing branch, a second text parsing branch and a fully connected layer text parsing branch connected in parallel; the text span set P is obtained by executing the text span analysis on the text information through the first text parsing branch, wherein P is a positive integer; the P text span sets returned by the first text parsing branch are received based on the second text parsing branch, the time window configuration of the text span is executed on the task queue, and the P text span thickness sets are called based on the task queue according to the configured time window; the span information homogeneity analysis is performed based on the P text span thickness sets through the fully connected layer text parsing branch to generate the homogeneity information and the heterogeneous information.

[0051] The specific configuration of the sending information library matching module 30 will be described in detail below. As described above, the sending information library matching module 30 further comprises: calling the sending information library comprising the association relationship between the heterogeneous information span and the sending information span; determining the sending information span based on the input of the heterogeneous information into the sending information library matching the heterogeneous information span; and determining the sending information by filling the sending information span based on the analogy confirmation of the homogeneous information.

[0052] The specific configuration of the sending information processing module 50 will be described in detail below. As described above, the sending information processing module 50 can further comprise: receiving the sending information by interfacing the collaborative system through the customer service system; generating a solution processing scheme by interfacing the call receiver based on the sending information to solve the call; and establishing a customer communication interface according to the solution processing scheme to send the solution processing scheme.

[0053] The multi-channel integrated customer service intelligent management system provided by the embodiments of the present application can execute the multi-channel integrated customer service intelligent management method provided by any of the embodiments of the present application, and has the corresponding function modules and beneficial effects of the execution method.

[0054] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, and each unit and module included is only divided according to the functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific name of each functional unit is only for easy mutual differentiation, and does not limit the protection scope of the present application.

[0055] The above specific embodiments do not constitute a limitation on the protection scope of the present application. Those skilled in the art should understand that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent replacement and improvement made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A customer service intelligent management method based on multi-channel integration, characterized in that: The method is applied to a customer service intelligent management system based on multi-channel integration, wherein the system is in communication with an order system, an abnormal order system, a collaborative system, and a customer service system, and the method includes: Receive order failure information from the order system, query the order failure log, extract the order information and send it to the task queue; Based on the abnormal order system, the task queue is monitored and text information is captured, and the text information is subjected to homogeneous analysis through a text parsing channel to generate heterogeneous information; Determining the sending information based on matching the heterogeneous information to a sending information database; Determine the information sending type according to the sending information, and call the sending interface; The sending interface is connected to the collaborative system to process the sending information.

2. The customer service intelligent management method based on multi-channel integration according to claim 1, characterized in that: The text parsing channel performs homogeneous analysis on text information, including: The text parsing channel includes a first text parsing branch, a second text parsing branch and a fully connected layer text parsing branch connected in parallel; Performing text span parsing on the text information through the first text parsing branch to obtain P text span sets, where P is a positive integer; Based on the second text parsing branch receiving the P text span sets returned by the first text parsing branch, performing time window configuration of text spans on the task queue, and retrieving the P text span thickness sets based on the task queue according to the configured time window; Through the fully connected layer text parsing branch, span information homogeneity analysis is performed according to the P text span thickness sets to generate homogeneous information and the heterogeneous information.

3. The customer service intelligent management method based on multi-channel integration according to claim 2, characterized in that: By using the fully connected layer text parsing branch, span information homogeneity analysis is performed based on the P text span thickness sets to generate homogeneous information and the heterogeneous information, including: Obtaining a first text span thickness set according to the P text span thickness sets; Receiving the first text span thickness set through the fully connected layer text parsing branch to perform thickness characteristic homogeneity judgment, and if the first text span thickness set satisfies the thickness characteristic homogeneity, storing the first text span thickness set in the homogeneity information; If the first text span thickness set does not satisfy thickness characteristic homogeneity, asynchronous thickness collection is performed, and the heterogeneous information is determined based on the asynchronous thickness.

4. The customer service intelligent management method based on multi-channel integration according to claim 3, characterized in that: Determining the heterogeneous information based on the asynchronous thickness includes: Preset heterogeneity threshold; Performing heterogeneous threshold comparison on the asynchronous thickness, and if the asynchronous thickness is less than or equal to the heterogeneous threshold, performing homogeneous information analogy confirmation on the first text span thickness set; If the asynchronous thickness is greater than the heterogeneous threshold, the heterogeneous information is confirmed for the first text span thickness set.

5. The customer service intelligent management method based on multi-channel integration according to claim 4, characterized in that: The step of matching the heterogeneous information with a sending information database and determining the sending information includes: Retrieving the sending information database, including the association between the heterogeneous information span and the sending information span; Based on the heterogeneous information input into the transmission information database, the heterogeneous information span is matched to determine the transmission information span; The sending information span is filled based on the homogeneous information analogy confirmation to determine the sending information.

6. The customer service intelligent management method based on multi-channel integration according to claim 3, characterized in that: The method further comprises: Constructing a network architecture of the fully connected layer text parsing branch, and receiving output data of the first text parsing branch and the second text parsing branch connected in parallel; Obtain sample training data, including multiple sample text span thickness sets, sample homogeneity information, and sample heterogeneity information; Based on the sample training data, the training of the fully connected layer text parsing branch is completed.

7. The customer service intelligent management method based on multi-channel integration according to claim 1, characterized in that: Connecting the sending interface through the collaborative system to process the sending information includes: Connecting the collaborative system through the customer service system to receive the sent information; Based on the sent information, a solution is generated for the call recipient; A customer communication interface is established according to the solution and the solution is sent.

8. The customer service intelligent management system based on multi-channel integration is characterized by: The system is used to implement the customer service intelligent management method based on multi-channel integration according to any one of claims 1 to 7, and the system includes: An order information extraction module is used to receive order failure information from the order system, query the order failure log, extract order information, and send it to the task queue; A text information homogeneity analysis module, which is used to monitor the task queue based on the abnormal order system, execute text information capture, perform homogeneity analysis on the text information through a text parsing channel, and generate heterogeneous information; a sending information database matching module, the sending information database matching module being used to match the sending information database based on the heterogeneous information to determine the sending information; An information sending type judgment module, the information sending type judgment module is used to judge the information sending type according to the sending information and call the sending interface; The sending information processing module is used to process the sending information by docking with the sending interface through the collaborative system.

Images