Automatic customer complaint processing method and device and storage medium

Abstract

The invention discloses a customer complaint automatic processing method and device and a storage medium. The method comprises the following steps: generating a complaint work order in a preset format according to complaint information of a customer; under the condition that the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold value, determining a processing suggestion of the complaint work order according to a processing scheme of the historical complaint case; under the condition that the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach a preset threshold value, the complaint work order is input into a target model, processing suggestions of the complaint work order are output through the target model, and the target model is a model obtained through training of the historical complaint case; and processing the complaint work order according to the processing suggestion. Through the method and the device, the problem of low customer complaint processing efficiency caused by low automation level of a complaint processing flow in related technologies is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence, in particular, to a method and device for automatically processing customer complaints and a storage medium. BACKGROUND

[0002] In today's fast-paced financial service industry, customer satisfaction has become one of the important indicators to measure the quality of service. However, the effectiveness and customer experience of the traditional customer complaint management system are often limited by many factors, and the most prominent problem is that the complaint handling process relies heavily on manual intervention and has a low level of automation. When customers submit complaints through telephone, email or online platforms, this information is initially manually entered into the system by customer service representatives, and then classified according to the content. This process not only takes time, but also due to the limitations of manual operation, it is easy to cause classification errors, affecting the accuracy and efficiency of subsequent processing. The assignment and processing stage of the complaint work order also highly depends on manual judgment and action, and the process based on manual interpretation and decision not only increases the response time, but also may lead to the omission or misunderstanding of important information, especially when facing emotional customer complaints, manual judgment is difficult to fully consider the urgency of the complaint, thereby affecting the priority setting and response speed of the complaint handling. In addition, the existing complaint handling method has low utilization efficiency of historical complaint data, which can only be used as post-analysis materials.

[0003] In view of the low level of automation of the complaint handling process in the related art, which leads to low efficiency in handling customer complaints, there is currently no effective solution. SUMMARY

[0004] The main purpose of the present application is to provide a method and device for automatically processing customer complaints and a storage medium to solve the problem of low level of automation of the complaint handling process in the related art, which leads to low efficiency in handling customer complaints.

[0005] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a method for automatically processing customer complaints is provided. The method comprises: generating a complaint work order in a preset format according to the complaint information of a customer; in the case that the similarity of the complaint content of the complaint work order and the complaint content of a historical complaint case reaches a preset threshold, determining a handling suggestion of the complaint work order according to the handling scheme of the historical complaint case; in the case that the similarity of the complaint content of the complaint work order and the complaint content of a historical complaint case does not reach the preset threshold, inputting the complaint work order into a target model, splitting the complaint work order into a plurality of words through the target model, converting each word into an embedding vector and adding position coding to each embedding vector, and sequentially predicting the handling suggestion of the complaint work order according to each embedding vector to output the handling suggestion of the complaint work order, wherein the target model is a model trained by historical complaint cases; and processing the complaint work order according to the handling suggestion.

[0006] Optionally, generating the complaint work order in the preset format according to the complaint information of the customer comprises: identifying an emergency degree of the complaint information according to the complaint information; converting the complaint information into text information; determining field information corresponding to the text information according to a preset complaint standardization label; and generating the complaint work order in the preset format according to the field information, the emergency degree, and a preset template.

[0007] Optionally, processing the complaint work order according to the processing suggestion comprises: pushing the processing suggestion to a target object, wherein the target object is determined by the following steps: determining a responsible department of the complaint work order according to a business type in the complaint work order; and determining the target object processing the complaint work order in the responsible department according to a work load condition of each object in the responsible department.

[0008] Optionally, in a case where a similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, determining the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case comprises: extracting entities and relationships between the entities in the complaint content of the historical complaint case, and constructing a first knowledge graph according to the entities and the relationships between the entities in the complaint content of the historical complaint case; extracting entities and relationships between the entities in the complaint content of the complaint work order, and constructing a second knowledge graph according to the entities and the relationships between the entities in the complaint content of the complaint work order; calculating a similarity between the first knowledge graph and the second knowledge graph, and determining the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case in a case where the similarity reaches the preset threshold.

[0009] Optionally, the training steps of the target model comprise: extracting complaint questions and processing schemes from historical complaint cases, and grouping the complaint questions and the processing schemes as paired data; inputting the complaint questions into an initial model, and outputting a preset scheme through the initial model, wherein the preset scheme is obtained by splitting the complaint question into multiple words through the initial model, converting each word into an embedding vector, adding position encoding to each embedding vector, and sequentially predicting a scheme according to each embedding vector; calculating a loss value of the preset scheme and the processing scheme through a loss function; adjusting parameters of the initial model according to the loss value to obtain an adjusted model; repeatedly performing the steps of inputting the complaint question into the adjusted model, outputting a new preset scheme by the adjusted model, calculating a loss value of the new preset scheme and the processing scheme through the loss function, and adjusting the model parameters according to the loss value until a preset stopping condition is reached to obtain the target model.

[0010] Optionally, after processing the complaint work order according to the processing suggestion, the method further comprises: monitoring each node of a processing flow of the complaint work order, and storing information of each node to a blockchain, wherein the nodes of the processing flow comprise receiving the complaint work order, assigning the complaint work order, processing the complaint work order, and a processing result of the complaint work order.

[0011] Optionally, after storing each node to the blockchain, the method further comprises: identifying common complaint content of historical complaint cases stored in the blockchain; determining a business and a product corresponding to the common complaint content, and formulating an optimization strategy for the business and the product corresponding to the common complaint content.

[0012] To achieve the above object, according to another aspect of the present application, an automatic processing device for customer complaints is provided. The device comprises: a generating unit configured to generate a complaint work order in a preset format according to complaint information of a customer; a determining unit configured to determine a processing suggestion for the complaint work order according to a processing scheme of a historical complaint case if a similarity between complaint content of the complaint work order and complaint content of the historical complaint case reaches a preset threshold; a first input unit configured to input the complaint work order into a target model if the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold, split the complaint work order into a plurality of words by the target model, convert each word into an embedding vector and add position encoding to each embedding vector, and sequentially predict a processing suggestion for the complaint work order according to each embedding vector to output the processing suggestion for the complaint work order, wherein the target model is a model trained by the historical complaint cases; and a processing unit configured to process the complaint work order according to the processing suggestion.

[0013] Optionally, the generating unit comprises: an identifying module configured to identify an urgency of the complaint information according to the complaint information; a converting module configured to convert the complaint information into text information; a determining module configured to determine field information corresponding to the text information according to a preset complaint standardization label; and a generating module configured to generate the complaint work order in the preset format according to the field information, the urgency and a preset template.

[0014] Optionally, the processing unit comprises a pushing module configured to push the processing suggestion to a target object, wherein the target object is determined by the following steps: determining a responsible department of the complaint work order according to a business type in the complaint work order; and determining the target object in the responsible department that processes the complaint work order according to a work load of each object in the responsible department.

[0015] Optionally, the determining unit includes: a first extraction module, used to extract entities and relationships between entities from the complaint content of historical complaint cases, and construct a first knowledge graph based on the entities and relationships between entities from the complaint content of historical complaint cases; a second extraction module, used to extract entities and relationships between entities from the complaint content of complaint work orders, and construct a second knowledge graph based on the entities and relationships between entities from the complaint content of complaint work orders; and a calculation module, used to calculate the similarity between the first knowledge graph and the second knowledge graph, and, when the similarity reaches a preset threshold, determine the processing suggestion for the complaint work order based on the processing scheme of historical complaint cases.

[0016] Optionally, the device further includes: an extraction unit for extracting complaint issues and handling solutions from historical complaint cases and pairing complaint issues and handling solutions into data; a second input unit for inputting complaint issues into an initial model and outputting a preset solution through the initial model, wherein the preset solution is obtained by splitting the complaint issue into multiple words through the initial model, converting each word into an embedding vector and adding a positional code to each embedding vector, and predicting the solution according to each embedding vector in turn; a calculation unit for calculating the loss value of the preset solution and the handling solution through a loss function; an adjustment unit for adjusting the parameters of the initial model according to the loss value to obtain an adjusted model; and a third input unit for repeatedly executing the steps of inputting complaint issues into the adjusted model, the adjusted model outputting a new preset solution, calculating the loss value of the new preset solution and the handling solution through a loss function, and adjusting the model parameters according to the loss value, until a preset stopping condition is reached to obtain the target model.

[0017] Optionally, the device further includes a storage unit for monitoring each node of the complaint work order processing flow after processing the complaint work order according to the processing suggestions, and storing the information of each node to the blockchain, wherein the nodes of the processing flow include accepting the complaint work order, assigning the complaint work order, processing the complaint work order, and the processing result of the complaint work order.

[0018] Optionally, the device further includes: an identification unit, used to identify common complaint content of historical complaint cases stored in the blockchain after each node in the process of monitoring the complaint work order is stored in the blockchain; and a formulation unit, used to determine the business and products corresponding to the common complaint content, and formulate optimization strategies for the business and products corresponding to the common complaint content.

[0019] According to another aspect of this application, a computer-readable storage medium is provided, which includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform an automatic processing method for any customer complaint.

[0020] According to another aspect of the present application, an electronic device is provided, comprising one or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs comprise instructions for performing any one of the automatic processing methods of customer complaints.

[0021] According to another aspect of the present application, a computer program product is provided, comprising computer instructions for implementing the steps of any one of the automatic processing methods of customer complaints when executed by a processor.

[0022] In the embodiments of the present application, a complaint work order in a preset format is generated according to the complaint information of a customer; in a case where the similarity between the complaint content of the complaint work order and the complaint content of a historical complaint case reaches a preset threshold, a processing suggestion for the complaint work order is determined according to the processing scheme of the historical complaint case; in a case where the similarity between the complaint content of the complaint work order and the complaint content of a historical complaint case does not reach the preset threshold, the complaint work order is input into a target model, the complaint work order is split into a plurality of words by the target model, each word is converted into an embedding vector and each embedding vector is added with position encoding, and a processing suggestion for the complaint work order is predicted in sequence according to each embedding vector, and the processing suggestion for the complaint work order is output, wherein the target model is a model trained by historical complaint cases; and the complaint work order is processed according to the processing suggestion, thereby solving the technical problem of low automation level of complaint processing procedures, resulting in low efficiency in processing customer complaints.

[0023] In the present application, customer complaint information is automatically received and analyzed to be converted into a standardized complaint work order in a preset format, and then the similarity between the complaint content of the complaint work order and the complaint content of a historical complaint case is checked; if the similarity reaches a preset threshold, the processing scheme of the historical complaint case is automatically referenced to generate a processing suggestion for the current work order; for those complaint work orders with low similarity to historical complaint cases, a target model trained in advance by a large amount of historical data is used for processing to generate a processing suggestion, and the complaint work order is processed according to the processing suggestion, thereby achieving the technical effect of improving the efficiency of processing customer complaints. BRIEF DESCRIPTION OF DRAWINGS

[0024] The accompanying drawings, which form a part of the present application, are intended to provide further understanding of the present application and are incorporated herein for a purpose of explanations and are not intended as an improper limitation to the present application. In the drawings:

[0025] Figure 1 Fig. 1 shows a hardware structure block diagram of a computer terminal for implementing the automatic processing method of customer complaints;

[0026] Figure 2is a flow chart of an automatic processing method of customer complaints according to an embodiment of the application;

[0027] Figure 3 is a schematic diagram of an automatic processing system of customer complaints according to an embodiment of the application;

[0028] Figure 4 is a schematic diagram of an automatic processing method of customer complaints according to an embodiment of the application;

[0029] Figure 5 is a schematic diagram of an automatic processing device of customer complaints according to an embodiment of the application;

[0030] Figure 6 is a structural block diagram of an electronic device according to an embodiment of the application. DETAILED DESCRIPTION

[0031] In order to enable persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by persons skilled in the art without creative labor should fall within the scope of protection of the present application.

[0032] It should be noted that the terms "first", "second", and the like in the specification and claims of the present application and the above-described drawings are used to distinguish similar objects, and do not necessarily indicate a specific order or a chronological sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units that are not clearly listed or inherent to the process, method, product or device.

[0033] It should be noted that the relevant information (including but not limited to historical case information, user personal information, etc.) and data (including but not limited to data for processing, stored data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties. For example, an interface is provided between the system and the relevant user or institution. Before obtaining the relevant information, the interface needs to send a request to the aforementioned user or institution, and after receiving the consent information feedback from the aforementioned user or institution, the relevant information is obtained.

[0034] Embodiment 1

[0035] According to the embodiment of the present application, a method for automatically processing customer complaints is also provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown.

[0036] The method provided by the embodiment of the present application can be executed in a mobile terminal, a computer terminal or similar computing device. Figure 1 A hardware structure block diagram of a computer terminal (or mobile device) for implementing the method for automatically processing customer complaints is shown. As shown in Figure 1 , the computer terminal 10 (or mobile device) can include one or more processors 102 (the processor 102 can include but not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it can also include a display, an input / output interface (Input / Output interface, I / O interface), a universal serial bus (Universal Serial Bus, USB) port (which can be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. Those skilled in the art can understand that Figure 1 The structure shown is only schematic, which does not limit the structure of the above-mentioned electronic device. For example, the computer terminal 10 can include more or less components than those shown in Figure 1 , or have a different configuration than that shown in Figure 1 .

[0037] It should be noted that the one or more processors 102 and / or other data processing circuits described above can be referred to herein as "data processing circuits" in general. The data processing circuit can be embodied in whole or in part as software, hardware, firmware or any other combination. In addition, the data processing circuit can be a single independent processing module, or any one of the other elements combined into the computer terminal 10 (or mobile device) in whole or in part. As referred to in the embodiments of the present application, the data processing circuit as a kind of processor control (for example, the selection of the variable resistance terminal path connected with the interface).

[0038] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the automatic customer complaint handling method in this embodiment of the application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned automatic customer complaint handling method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor 102, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0039] The transmission device 106 is used to receive or send data via a network. Specific examples of the network described above may include a wireless network provided by the communication provider of the computer terminal 10. In one example, the transmission device 106 includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 106 may be a Radio Frequency (RF) module, used for wireless communication with the Internet.

[0040] The display can be, for example, a touchscreen liquid crystal display (LCD), which allows the user to interact with the user interface of the computer terminal 10 (or mobile device).

[0041] Under the aforementioned operating environment, this application provides the following: Figure 2 The method for automatically handling customer complaints is shown. Figure 2 This is a flowchart of an automatic customer complaint processing method according to Embodiment 1 of this application.

[0042] Step S201: Generate a complaint work order in a preset format based on the customer's complaint information.

[0043] Optionally, customers can submit complaint information via telephone, email, online platforms, etc. Complaint information can be in the form of voice, text, or images. A pre-formatted complaint form refers to converting various forms of complaint information into a standardized, structured document. Key fields such as the complainant, complaint date, complaint recipient, and complaint content can be extracted from the complaint information and filled into a template according to a predetermined format to generate a pre-formatted complaint form.

[0044] Step S202, in the case that the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, determining a processing suggestion of the complaint work order according to a processing scheme of the historical complaint case.

[0045] Optionally, the historical complaint case includes past customer complaint content and its processing record, by calculating the similarity between the complaint content of the newly generated complaint work order and the complaint content of the historical complaint case, it is determined whether the similarity exceeds a preset threshold, if the preset threshold is exceeded, it can be indicated that the complaint content of the newly generated complaint work order is similar to the past processed complaint case, the successful processing scheme of the historical complaint case can be used for reference, and the corresponding processing suggestion is directly generated.

[0046] Step S203, in the case that the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold, inputting the complaint work order into a target model, splitting the complaint work order into multiple words by the target model, converting each word into an embedding vector and adding position coding to each embedding vector, and sequentially predicting the processing suggestion of the complaint work order according to each embedding vector to output the processing suggestion of the complaint work order, wherein the target model is a model trained by the historical complaint case.

[0047] Optionally, when the similarity between the complaint content of the newly generated complaint work order and the complaint content of the historical complaint case is lower than the preset threshold, the complaint work order can be input into the target model for processing. The target model first splits the work order text into words, generates an embedding vector for each word, and then adds position coding to the embedding vector. The target model predicts the next possible word according to each embedding vector with position information, gradually constructs a string of words, and finally forms a processing suggestion.

[0048] Step S204, processing the complaint work order according to the processing suggestion.

[0049] Optionally, after the processing suggestion is generated, the relevant department or personnel responsible for processing the complaint work order will take specific actions according to the processing suggestion to solve the problems in the customer complaint.

[0050] The method for automatically processing customer complaints provided by the embodiments of the present application generates a complaint work order in a preset format according to the complaint information of the customer; in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, the processing proposal of the complaint work order is determined according to the processing scheme of the historical complaint case; in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold, the complaint work order is input into a target model, the complaint work order is split into multiple words by the target model, each word is converted into an embedding vector and each embedding vector is added with position coding, and the processing proposal of the complaint work order is predicted according to each embedding vector in sequence, and the processing proposal of the complaint work order is output, wherein the target model is a model trained by historical complaint cases; and the complaint work order is processed according to the processing proposal, which solves the technical problem of low automation level of the complaint processing procedure and low efficiency in processing customer complaints. In the present application, customer complaint information is automatically received and analyzed to be converted into a standardized complaint work order in a preset format, and then the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case is checked; if the similarity reaches a preset threshold, the processing proposal of the current work order is automatically generated by referring to the processing scheme of the historical complaint case; for those complaint work orders with low similarity to the historical complaint cases, a target model trained by a large amount of historical data is used for processing, and the complaint work order is processed according to the processing proposal after the processing proposal is generated, thereby achieving the technical effect of improving the efficiency of processing customer complaints.

[0051] To generate more accurate and standardized complaint work orders, optionally, in the method for automatically processing customer complaints provided by the embodiments of the present application, generating a complaint work order in a preset format according to the complaint information of the customer comprises:

[0052] First, identifying the urgency of the complaint information according to the complaint information.

[0053] Optionally, the complaint information submitted by the customer can be analyzed by using techniques such as sentiment analysis and keyword detection to identify and quantify the urgency of the complaint information. For example, by analyzing the emotional characteristics such as the tone and words of the customer, it is determined whether the complaint belongs to an emergency state.

[0054] Second, converting the complaint information into text information.

[0055] Optionally, if the complaint information of the customer is submitted in the form of voice or image, etc., voice recognition, optical character recognition, etc. can be used to convert these non-text information into text information that can be analyzed and processed.

[0056] Third, determining the field information corresponding to the text information according to the preset complaint standardization label.

[0057] Optionally, the preset complaint standardization label is a structured field set for the complaint information, such as complaint area, complaint type, business type, complaint content, responsible department, urgency, and required resolution time limit. For each label, field information is extracted from the text information.

[0058] In the fourth step, a preset format complaint work order is generated according to the field information, the urgency, and a preset template.

[0059] Optionally, the preset template is a designed complaint work order format for standardizing the recording and display of complaint information. The preset format complaint work order is obtained by filling the field information and the urgency into the preset template.

[0060] In summary, through the above steps, the emergency of the complaint content is automatically identified, the complaint information is converted into analyzable text, the structured key information is obtained, and finally the standardized complaint work order is generated. The entire process does not require manual intervention. This design not only significantly improves the automation level of complaint processing, but also realizes the structuring and consistency of information through standardized labels and templates, thereby improving the processing efficiency.

[0061] In order to improve the efficiency of complaint processing, in the automatic processing method of customer complaints provided in the embodiments of the present application, the processing of the complaint work order according to the processing suggestion includes: pushing the processing suggestion to the target object, wherein the target object is determined by the following steps: determining the responsible department of the complaint work order according to the business type in the complaint work order; and determining the target object in the responsible department for processing the complaint work order according to the work load of each object in the responsible department.

[0062] Optionally, the business type refers to a specific bank service category involved in the customer complaint, such as loan service, credit card service, account management, etc. According to the business type information identified in the complaint work order, the department responsible for processing the work order can be automatically determined through intelligent matching or preset rules. For example, if the complaint is about credit card service, the complaint work order is sent to the credit card service center. After determining the responsible department, the work load of each employee in the department can be analyzed to find the employee with the least workload who can respond the fastest as the target processing object.

[0063] In summary, through the above steps, the most suitable responsible department is automatically identified and matched based on the business type of the complaint content, and then the complaint work order is assigned to the employee who can quickly handle the problem by analyzing the work load of the employees in the responsible department, thereby improving the processing efficiency.

[0064] To improve the efficiency and accuracy of complaint processing, optionally, in the automatic processing method of customer complaints provided in the embodiments of the present application, in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, the processing suggestion of the complaint work order is determined according to the processing scheme of the historical complaint case, which includes:

[0065] First, extract the entities in the complaint content of the historical complaint case and the relationship between the entities, and construct a first knowledge graph according to the entities in the complaint content of the historical complaint case and the relationship between the entities.

[0066] Optionally, the entity is a specific object mentioned in the complaint content, such as a product, a service link, customer personal information, etc., and the relationship between entities is a link describing the interaction between entities, such as "the customer is associated with the account" and "the product causes revenue problems". Using natural language processing technology to extract entities and relationships between entities from the text of historical complaint cases, a first knowledge graph can be constructed to reflect the details of historical complaint cases.

[0067] Second, extract the entities in the complaint content of the complaint work order and the relationship between the entities, and construct a second knowledge graph according to the entities in the complaint content of the complaint work order and the relationship between the entities.

[0068] Optionally, natural language processing technology can be used to extract entities and relationships between entities from the text content of the current complaint work order, and a second knowledge graph can be constructed based on this, which can reflect the main features and details of the current complaint.

[0069] Third, calculate the similarity of the first knowledge graph and the second knowledge graph, and in the case where the similarity reaches a preset threshold, determine the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case.

[0070] Optionally, by calculating the similarity between the first knowledge graph and the second knowledge graph, if the similarity reaches a preset threshold, it is considered that the current complaint work order is highly related to a historical case, and the processing scheme of the historical case can be referred to to form and output the processing suggestion for the current complaint work order.

[0071] In summary, through the above steps, the first knowledge graph and the second knowledge graph are constructed, and the similarity of the two knowledge graphs is calculated to determine whether the current complaint work order has similarities with historical complaint cases. When the similarity reaches a preset threshold, the processing scheme of the historical case is used to quickly generate a processing suggestion suitable for the current complaint work order, improving the efficiency and accuracy of complaint processing.

[0072] To improve the accuracy of the complaint suggestions generated by the target model, optionally, in the automatic processing method of customer complaints provided in the embodiments of the present application, the training step of the target model includes:

[0073] First, extract complaint problems and handling solutions from historical complaint cases, and form paired data of complaint problems and handling solutions.

[0074] Optionally, complaint problems and handling solutions can be identified and extracted from the historical complaint case database to form multiple sets of paired data as the basis for training the target model.

[0075] Second, input the complaint problem into the initial model, and output the preset solution through the initial model, wherein the preset solution is obtained by splitting the complaint problem into multiple words, converting each word into an embedding vector, and adding position encoding to each embedding vector, and then sequentially predicting the solution based on each embedding vector.

[0076] Optionally, the initial model accepts text input of the complaint problem, splits it into words, then converts each word into an embedding vector and adds position encoding information. The initial model gradually predicts the next word based on these embedding vectors and position encodings until a complete preset solution is generated.

[0077] Third, calculate the loss value of the preset solution and the handling solution through the loss function.

[0078] Optionally, the loss function can be used to compare the preset solution generated by the model with the real handling solution in the historical complaint case, calculate the difference between the two, and obtain the loss value.

[0079] Fourth, adjust the parameters of the initial model according to the loss value to obtain the adjusted model.

[0080] Optionally, according to the loss value, the parameters in the initial model can be adjusted through the backpropagation algorithm, expecting to reduce the prediction error in the next iteration, and obtain a model with better performance.

[0081] Fifth, repeat the steps of inputting the complaint problem into the adjusted model, the adjusted model outputting a new preset solution, calculating the loss value of the new preset solution and the handling solution through the loss function, and adjusting the model parameters according to the loss value until the preset stopping condition is reached, and obtain the target model.

[0082] Optionally, repeat steps two to four, each time using the adjusted model to make a new round of prediction, then calculate the loss value again and adjust the parameters until the performance of the model is stable or the preset stopping condition is reached. The preset stopping condition can be reaching a preset number of iterations or the loss value reaching a preset value, etc.

[0083] In summary, through the above steps, the target model is trained through deep learning techniques and iterative optimization strategies, enabling the target model to learn and master the processing logic and methods of historical complaint cases, so that when facing new customer complaints, it can quickly and accurately generate reasonable processing suggestions.

[0084] To improve the traceability of the complaint process, optionally, in the automatic processing method for customer complaints provided in the embodiments of the present application, after processing the complaint work order according to the processing suggestion, the method further comprises: monitoring each node of the processing flow of the complaint work order, and storing the information of each node to the blockchain, wherein the nodes of the processing flow include receiving the complaint work order, assigning the complaint work order, processing the complaint work order, and the processing result of the complaint work order.

[0085] Optionally, after completing the processing of the complaint work order and taking action according to the processing suggestion, each key node in the entire processing flow can be monitored in real time. Whenever the process advances to a new node (such as the completion of complaint acceptance, the assignment of the work order to a specific department, etc.), the detailed information of the node, such as the timestamp, the processing status, the responsible department and personnel, etc., can be automatically encrypted and stored to the blockchain, forming a complete and tamper-proof complaint processing flow record. By chaining the complaint processing information, the authenticity and traceability of the information are improved, and the customer's trust is enhanced.

[0086] To improve the quality of subsequent services and customer experience, optionally, in the automatic processing method for customer complaints provided in the embodiments of the present application, after monitoring the nodes of the processing flow of the complaint work order and storing each node to the blockchain, the method further comprises:

[0087] First, identify the common complaint content of the historical complaint cases stored in the blockchain.

[0088] Optionally, historical complaint case data can be read from the blockchain, and data mining and text clustering algorithms can be used to identify common content in the complaint cases.

[0089] Second, determine the business and product corresponding to the common complaint content, and develop optimization strategies for the business and product corresponding to the common complaint content.

[0090] Optionally, after identifying the common complaint content, the specific business and product involved in these complaint contents can be further analyzed, and corresponding optimization strategies can be developed. For example, if credit card services frequently trigger complaints, measures such as optimizing the credit card approval process, increasing customer support, and adjusting the charging policy can be considered.

[0091] In summary, by identifying common problems in historical complaint cases and customizing optimization strategies, the use of blockchain technology improves the authenticity and traceability of data, which helps to improve service quality and prevent potential problems, thereby improving customer experience.

[0092] For example, Figure 3 A schematic diagram of an automatic processing system for customer complaints according to the present embodiment includes seven modules, namely a basic framework module, an extraction and identification module, an intelligent processing engine, a blockchain storage module, a feedback module, a closed-loop tracking module, and an analysis module. The basic framework module includes a standardized label framework. After receiving multi-modal customer complaint information, the standardized complaint work order is generated by inputting the extraction and identification engine. The complaint work order is input into the intelligent processing engine. The responsible department of the complaint work order is determined by the intelligent dispatch subsystem. The intelligent auxiliary decision-making subsystem proposes a processing suggestion. The responsible part handles the complaint work order according to the processing suggestion. The complaint process is input into the blockchain storage module to realize double-chain collaborative storage. The results of complaint handling are input into the feedback module and delivered to the client. The analysis module analyzes the common problems of historical complaint cases and generates optimization suggestions. In summary, the automatic processing system for customer complaints according to the present embodiment can improve the efficiency of complaint handling and improve customer satisfaction.

[0093] For example, Figure 4 A schematic diagram of an automatic processing method for customer complaints according to the present embodiment is shown. First, access the complaint telephone or picture signal, convert the voice or picture form of the complaint content into text content, extract the keywords in the text content, identify the customer's core complaint point, match the business type, responsible department, and emergency level to generate a complaint work order. After receiving the complaint signal, the matching results are stored and dispatched to the relevant responsible department. The responsible part receives the complaint work order and gives rectification measures. The rectification measures are pushed to the customer service manager and the customer through APP, SMS, telephone, etc. to realize automatic processing of customer complaints.

[0094] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown.

[0095] Embodiment 2

[0096] The present application also provides an automatic processing device for customer complaints. It should be noted that the automatic processing device for customer complaints according to the present application can be used to execute the automatic processing method for customer complaints provided by the present application. The automatic processing device for customer complaints provided by the present application is introduced as follows.

[0097] According to the embodiment of the application, the device for implementing the automatic processing method of the customer complaint is also provided, as shown in the figure, the device comprises a generating unit 501, a determining unit 502, a first input unit 503 and a processing unit 504. Figure 5

[0098] Specifically, the generating unit 501 is configured to generate a complaint work order in a preset format according to the complaint information of the customer;

[0099] The determining unit 502 is configured to determine a processing suggestion of the complaint work order according to a processing scheme of a historical complaint case in a case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold value;

[0100] The first input unit 503 is configured to input the complaint work order into a target model in a case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold value, split the complaint work order into a plurality of words through the target model, convert each word into an embedding vector and add position coding to each embedding vector, sequentially predict the processing suggestion of the complaint work order according to each embedding vector, and output the processing suggestion of the complaint work order, wherein the target model is a model trained through the historical complaint cases;

[0101] The processing unit 504 is configured to process the complaint work order according to the processing suggestion.

[0102] The device for automatically processing the customer complaint provided by the embodiment of the application generates a complaint work order in a preset format according to the complaint information of the customer through the generating unit 501, determines a processing suggestion of the complaint work order according to a processing scheme of a historical complaint case in a case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold value through the determining unit 502, inputs the complaint work order into a target model in a case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold value through the first input unit 503, splits the complaint work order into a plurality of words through the target model, converts each word into an embedding vector and adds position coding to each embedding vector, sequentially predicts the processing suggestion of the complaint work order according to each embedding vector, and outputs the processing suggestion of the complaint work order, wherein the target model is a model trained through the historical complaint cases; and the processing unit 504 processes the complaint work order according to the processing suggestion, thereby solving the problem of low automation level of the complaint processing flow in the related art, leading to low efficiency in processing the customer complaint, and achieving the effect of improving the efficiency of processing the customer complaint.

[0103] ​Optionally, in the device for automatically processing customer complaints provided in the embodiments of the present application, the generating unit 501 comprises: an identifying module configured to identify the urgency of the complaint information according to the complaint information; a converting module configured to convert the complaint information into text information; a determining module configured to determine field information corresponding to the text information according to a preset complaint standardization label; and a generating module configured to generate a complaint work order in a preset format according to the field information, the urgency, and a preset template.

[0104] Optionally, in the device for automatically processing customer complaints provided in the embodiments of the present application, the processing unit 504 comprises a pushing module configured to push the processing suggestion to a target object, wherein the target object is determined by the following steps: determining a responsible department of the complaint work order according to the business type in the complaint work order; and determining the target object in the responsible department that processes the complaint work order according to the work load of each object in the responsible department.

[0105] Optionally, in the device for automatically processing customer complaints provided in the embodiments of the present application, the determining unit 502 comprises: a first extracting module configured to extract entities and relationships between the entities in the complaint content of the historical complaint case, and construct a first knowledge graph according to the entities and the relationships between the entities in the complaint content of the historical complaint case; a second extracting module configured to extract entities and relationships between the entities in the complaint content of the complaint work order, and construct a second knowledge graph according to the entities and the relationships between the entities in the complaint content of the complaint work order; and a calculating module configured to calculate the similarity of the first knowledge graph and the second knowledge graph, and determine the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case in the case where the similarity reaches a preset threshold.

[0106] Optionally, in the device for automatically processing customer complaints provided in the embodiments of the present application, the device further comprises: an extracting unit configured to extract complaint problems and processing schemes from historical complaint cases, and group the complaint problems and the processing schemes into paired data; a second input unit configured to input the complaint problems into an initial model, and output a preset scheme through the initial model, wherein the preset scheme is obtained by splitting the complaint problems into a plurality of words through the initial model, converting each word into an embedding vector and adding position encoding to each embedding vector, and sequentially predicting a scheme according to each embedding vector; a calculating unit configured to calculate a loss value of the preset scheme and the processing scheme through a loss function; an adjusting unit configured to adjust parameters of the initial model according to the loss value to obtain an adjusted model; and a third input unit configured to repeatedly perform the steps of inputting the complaint problems into the adjusted model, outputting a new preset scheme through the adjusted model, calculating a loss value of the new preset scheme and the processing scheme through the loss function, and adjusting the model parameters according to the loss value until a preset stopping condition is reached to obtain a target model.

[0107] Optionally, in the device for automatically processing customer complaints provided in the embodiment of the present application, the device further comprises a storage unit configured to, after processing the complaint work order according to the processing suggestion, monitor each node of a processing flow of the complaint work order, and store information of each node to a block chain, wherein the nodes of the processing flow include receiving the complaint work order, assigning the complaint work order, processing the complaint work order, and a processing result of the complaint work order.

[0108] Optionally, in the device for automatically processing customer complaints provided in the embodiment of the present application, the device further comprises an identification unit configured to, after monitoring each node of the processing flow of the complaint work order and storing each node to the block chain, identify common complaint content of historical complaint cases stored in the block chain; and a formulation unit configured to determine a business and a product corresponding to the common complaint content, and formulate an optimization strategy for the business and the product corresponding to the common complaint content.

[0109] It should be noted that the generation unit 501, the determination unit 502, the first input unit 503, and the processing unit 504 correspond to steps S201 to S204 in Embodiment 1, and the four units have the same instances and application scenarios as the corresponding steps, but are not limited to the content disclosed in Embodiment 1. It should be noted that the above-mentioned modules or units can be hardware components or software components stored in a memory (for example, the memory 104) and processed by one or more processors (for example, the processors 102a, 102b, …, 102n), and the above-mentioned modules can also be a part of the device and can run in the computer terminal 10 provided in Embodiment 1.

[0110] Embodiment 3

[0111] The embodiment of the present application can provide a computer terminal, which can be any one of the computer terminal devices in the computer terminal group. Optionally, in the embodiment, the computer terminal can be replaced by a mobile terminal or an electronic device or other terminal device.

[0112] Optionally, in the embodiment, the computer terminal can be located in at least one of the network devices in the computer network.

[0113] In the embodiment, the computer terminal can execute program codes of the following steps in the automatic processing method of customer complaints: generating a complaint work order in a preset format according to the complaint information of the customer; in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, determining the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case; in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold, inputting the complaint work order into a target model, splitting the complaint work order into multiple words through the target model, converting each word into an embedding vector and adding position coding to each embedding vector, sequentially predicting the processing suggestion of the complaint work order according to each embedding vector, and outputting the processing suggestion of the complaint work order, wherein the target model is a model trained through historical complaint cases; and processing the complaint work order according to the processing suggestion.

[0114] Optionally, the computer terminal can execute program codes of the following steps in the automatic processing method of customer complaints: generating a complaint work order in a preset format according to the complaint information of the customer includes: identifying the urgency of the complaint information according to the complaint information; converting the complaint information into text information; determining the field information corresponding to the text information according to the preset complaint standardization label; and generating the complaint work order in the preset format according to the field information, the urgency and the preset template.

[0115] Optionally, the computer terminal can execute program codes of the following steps in the automatic processing method of customer complaints: processing the complaint work order according to the processing suggestion includes: pushing the processing suggestion to a target object, wherein the target object is determined by the following steps: determining the responsible department of the complaint work order according to the business type in the complaint work order; and determining the target object for processing the complaint work order in the responsible department according to the work load of each object in the responsible department.

[0116] Optionally, the computer terminal can execute program codes of the following steps in the automatic processing method of customer complaints: in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, determining the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case includes: extracting entities and relationships between entities in the complaint content of the historical complaint case, constructing a first knowledge graph according to the entities and relationships between entities in the complaint content of the historical complaint case; extracting entities and relationships between entities in the complaint content of the complaint work order, constructing a second knowledge graph according to the entities and relationships between entities in the complaint content of the complaint work order; calculating the similarity of the first knowledge graph and the second knowledge graph, and in the case where the similarity reaches a preset threshold, determining the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case.

[0117] Optionally, the computer terminal can execute program codes of the following steps in the automatic processing method of customer complaints: the training step of the target model comprises: extracting complaint problems and processing solutions from historical complaint cases, and grouping the complaint problems and processing solutions as paired data; inputting the complaint problems into an initial model, and outputting a preset solution through the initial model, wherein the preset solution is obtained by splitting the complaint problems into multiple words through the initial model, converting each word into an embedding vector and adding position coding to each embedding vector, and sequentially predicting a solution according to each embedding vector; calculating the loss value of the preset solution and the processing solution through a loss function; adjusting the parameters of the initial model according to the loss value to obtain an adjusted model; repeatedly executing the steps of inputting the complaint problems into the adjusted model, outputting a new preset solution by the adjusted model, calculating the loss value of the new preset solution and the processing solution through the loss function, and adjusting the model parameters according to the loss value until a preset stopping condition is reached to obtain the target model.

[0118] Optionally, the computer terminal can execute program codes of the following steps in the automatic processing method of customer complaints: after processing the complaint work order according to the processing suggestion, the method further comprises: monitoring each node of the processing flow of the complaint work order, and storing the information of each node to a blockchain, wherein the nodes of the processing flow include receiving the complaint work order, assigning the complaint work order, processing the complaint work order, and the processing result of the complaint work order.

[0119] Optionally, the computer terminal can execute program codes of the following steps in the automatic processing method of customer complaints: after monitoring the nodes of the processing flow of the complaint work order and storing each node to the blockchain, the method further comprises: identifying the common complaint content of the historical complaint cases stored in the blockchain; determining the business and product corresponding to the common complaint content, and formulating an optimization strategy for the business and product corresponding to the common complaint content.

[0120] Optionally, Figure 6 is a structural block diagram of an electronic device according to an embodiment of the present application. As shown in Figure 6 the electronic device can include one or more (only one is shown in the figure) processors 602, memories 604, storage controllers, and peripheral interfaces, wherein the peripheral interfaces are connected with radio frequency modules, audio modules, and displays. Figure 6

[0121] ​The memory can be configured to store software programs and modules, such as program instructions / modules corresponding to the automatic processing method and device for customer complaints in the embodiments of the present application. The processor executes various functions and data processing by running the software programs and modules stored in the memory, that is, implements the automatic processing method for customer complaints. The memory can include a high-speed random access memory, and can further include a non-volatile memory, such as one or more magnetic storage devices, flash memories, or other non-volatile solid-state memories. In some examples, the memory can further include a memory remotely arranged with respect to the processor, which can be connected to the terminal through 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 a combination thereof.

[0122] The processor can call information and application programs stored in the memory through the transmission device to execute the above-mentioned steps in the automatic processing method for customer complaints.

[0123] By generating a complaint work order in a preset format according to the customer complaint information; in the case where the similarity of the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, determining a processing suggestion of the complaint work order according to the processing scheme of the historical complaint case; in the case where the similarity of the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold, inputting the complaint work order into a target model, splitting the complaint work order into a plurality of words through the target model, converting each word into an embedding vector and adding position coding to each embedding vector, and sequentially predicting a processing suggestion of the complaint work order according to each embedding vector to output the processing suggestion of the complaint work order, wherein the target model is a model trained through historical complaint cases; and processing the complaint work order according to the processing suggestion, the technical problem of low automation level of the complaint processing flow and low efficiency of processing customer complaints is solved, and the technical effect of improving the efficiency of processing customer complaints is achieved.

[0124] Those skilled in the art can understand that Figure 6 The structure shown is only schematic, and the electronic device can also be a terminal device such as a smart phone, a tablet computer, a palm computer, a mobile Internet device (MID), a PAD, and the like. Figure 6 It does not limit the structure of the above-mentioned electronic device. For example, the electronic device can further include more or less components (such as a network interface, a display device, and the like) than those shown in the figure, or have a different configuration from that shown in the figure. Figure 6 Figure 6

[0125] ​​Those skilled in the art can understand that all or part of the steps in the above-mentioned embodiments can be completed by instructing the terminal device related hardware through a program, and the program can be stored in a computer readable storage medium, which can include a flash disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0126] Embodiment 4

[0127] The embodiments of the present application also provide a storage medium. Optionally, in the embodiments, the storage medium can be used to store the program code executed by the automatic processing method of customer complaints provided in the embodiment 1.

[0128] Optionally, in the embodiments, the storage medium can be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group.

[0129] Optionally, in the embodiments, the storage medium is configured to store program code for performing the following steps: generating a complaint work order in a preset format according to the complaint information of the customer; determining a processing suggestion of the complaint work order according to the processing scheme of the historical complaint case in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold; inputting the complaint work order into a target model to split the complaint work order into a plurality of words, convert each word into an embedding vector and add position coding to each embedding vector, and sequentially predict the processing suggestion of the complaint work order according to each embedding vector, and output the processing suggestion of the complaint work order in the case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case does not reach the preset threshold, wherein the target model is a model trained by the historical complaint cases; and processing the complaint work order according to the processing suggestion.

[0130] Optionally, the storage medium is further configured to store program code for performing the following steps: generating a complaint work order in a preset format according to the complaint information of the customer includes: identifying the urgency of the complaint information according to the complaint information; converting the complaint information into text information; determining the field information corresponding to the text information according to the preset complaint standardization label; and generating the complaint work order in the preset format according to the field information, the urgency and the preset template.

[0131] Optionally, the storage medium is further configured to store program code for performing the following steps: processing the complaint work order according to the processing suggestion comprises: pushing the processing suggestion to the target object, wherein the target object is determined by the following steps: determining the responsible department of the complaint work order according to the business type in the complaint work order; and determining the target object in the responsible department processing the complaint work order according to the work load of each object in the responsible department.

[0132] Optionally, the storage medium is further configured to store program code for performing the following steps: in a case where the similarity between the complaint content of the complaint work order and the complaint content of the historical complaint case reaches a preset threshold, determining the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case comprises: extracting entities and relationships between entities in the complaint content of the historical complaint case, and constructing a first knowledge graph according to the entities and relationships between entities in the complaint content of the historical complaint case; extracting entities and relationships between entities in the complaint content of the complaint work order, and constructing a second knowledge graph according to the entities and relationships between entities in the complaint content of the complaint work order; calculating the similarity of the first knowledge graph and the second knowledge graph, and in a case where the similarity reaches the preset threshold, determining the processing suggestion of the complaint work order according to the processing scheme of the historical complaint case.

[0133] Optionally, the storage medium is further configured to store program code for performing the following steps: the training step of the target model comprises: extracting complaint problems and processing schemes from historical complaint cases, and grouping the complaint problems and processing schemes as paired data; inputting the complaint problem into the initial model, and outputting a preset scheme through the initial model, wherein the preset scheme is obtained by splitting the complaint problem into multiple words through the initial model, converting each word into an embedding vector and adding position encoding to each embedding vector, and sequentially predicting a scheme according to each embedding vector; calculating the loss value of the preset scheme and the processing scheme through a loss function; adjusting the parameters of the initial model according to the loss value to obtain an adjusted model; repeating the steps of inputting the complaint problem into the adjusted model, the adjusted model outputting a new preset scheme, calculating the loss value of the new preset scheme and the processing scheme through a loss function, and adjusting the model parameters according to the loss value until a preset stopping condition is reached to obtain the target model.

[0134] Optionally, the storage medium is further configured to store program code for performing the following steps: after processing the complaint work order according to the processing suggestion, the method further comprises: monitoring each node of the processing flow of the complaint work order, and storing the information of each node to a blockchain, wherein the nodes of the processing flow include receiving the complaint work order, assigning the complaint work order, processing the complaint work order, and the processing result of the complaint work order.

[0135] Optionally, the storage medium is further configured to store program code for performing the following steps: monitoring the nodes of the processing flow of the complaint work order, after storing each node to the blockchain, the method further comprises: identifying the common complaint content of the historical complaint cases stored in the blockchain; determining the business and product corresponding to the common complaint content, and formulating an optimization strategy for the business and product corresponding to the common complaint content.

[0136] The application further provides a computer program product, which is suitable for executing the steps of the automatic processing method of customer complaints when executed on a data processing device.

[0137] The above-mentioned sequence numbers of the embodiments of the application are only for description, and do not represent advantages or disadvantages of the embodiments.

[0138] In the above-mentioned embodiments of the application, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0139] In the several embodiments provided by the present application, it should be understood that the disclosed technology can be implemented in other ways. Of course, the unit embodiment described above is only schematic. For example, the division of the units is only a logical function division. There can be another division manner for actual implementation, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed coupling or direct coupling or communication connection between the units can be indirect coupling or communication connection through some interface, or electrical or other form.

[0140] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment.

[0141] In addition, each functional unit in each embodiment of the application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit. The above integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0142] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or say the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the embodiments of the present application. The aforementioned storage medium includes: a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0143] The above is only the preferred embodiment of the present application, and it should be pointed out that for those skilled in the art, without departing from the principles of the present application, a number of improvements and refinements can be made, which should be considered as the protection scope of the present application.

Claims

1. An automated method for handling customer complaints, characterized in that, include: Generate a pre-formatted complaint form based on the customer's complaint information; If the similarity between the complaint content of the complaint ticket and the complaint content of historical complaint cases reaches a preset threshold, the processing suggestion for the complaint ticket is determined based on the processing solution of the historical complaint cases. If the similarity between the complaint content of the complaint ticket and the complaint content of historical complaint cases does not reach a preset threshold, the complaint ticket is input into the target model. The target model splits the complaint ticket into multiple words, converts each word into an embedding vector, adds positional encoding to each embedding vector, and predicts the processing suggestions for the complaint ticket based on each embedding vector in turn. The processing suggestions for the complaint ticket are then output. The target model is a model trained using the historical complaint cases. The complaint ticket will be processed according to the processing suggestions.

2. The method according to claim 1, characterized in that, The system generates pre-formatted complaint forms based on customer complaint information, including: The urgency of the complaint information is identified based on the complaint information; Convert the complaint information into text information; The field information corresponding to the text information is determined based on the preset standardized complaint tags; The complaint form in the preset format is generated based on the field information, the urgency level, and the preset template.

3. The method according to claim 1, characterized in that, Processing the complaint ticket according to the aforementioned processing suggestions includes: The processing suggestion is pushed to the target object, wherein the target object is determined by the following steps: determining the responsible department of the complaint work order based on the business type in the complaint work order; determining the target object in the responsible department to process the complaint work order based on the workload of each object in the responsible department.

4. The method according to claim 1, characterized in that, If the similarity between the complaint content of the complaint ticket and the complaint content of historical complaint cases reaches a preset threshold, the processing suggestions for the complaint ticket are determined based on the processing solutions of the historical complaint cases, including: Extract the entities and relationships between entities from the complaint content of the historical complaint cases, and construct a first knowledge graph based on the entities and relationships between entities in the complaint content of the historical complaint cases; Extract the entities and relationships between entities from the complaint content of the complaint ticket, and construct a second knowledge graph based on the entities and relationships between entities in the complaint content of the complaint ticket; Calculate the similarity between the first knowledge graph and the second knowledge graph. If the similarity reaches the preset threshold, determine the processing suggestion for the complaint ticket based on the processing scheme of the historical complaint cases.

5. The method according to claim 1, characterized in that, The training steps for the target model include: Extract the complaint issues and solutions from the historical complaint cases, and combine the complaint issues and solutions into paired data. The complaint is input into an initial model, and the initial model outputs a preset solution. The preset solution is obtained by splitting the complaint into multiple words through the initial model, converting each word into an embedding vector, adding a position code to each embedding vector, and predicting the solution according to each embedding vector in turn. The loss values ​​of the preset scheme and the processing scheme are calculated using a loss function; The parameters of the initial model are adjusted based on the loss value to obtain the adjusted model; The process of inputting the complaint into the adjusted model, the adjusted model outputting a new preset solution, calculating the loss value of the new preset solution and the processing solution using a loss function, and adjusting the model parameters based on the loss value is repeated until a preset stopping condition is reached, thus obtaining the target model.

6. The method according to claim 1, characterized in that, After processing the complaint ticket according to the processing suggestions, the method further includes: Monitor each node of the complaint work order processing flow and store the information of each node in the blockchain. The nodes of the processing flow include accepting the complaint work order, assigning the complaint work order, processing the complaint work order, and the processing result of the complaint work order.

7. The method according to claim 6, characterized in that, After storing each node monitoring the processing flow of the complaint ticket on the blockchain, the method further includes: Identify common complaint content among historical complaint cases stored in the blockchain; Identify the business and products corresponding to the common complaints, and formulate optimization strategies for the business and products corresponding to the common complaints.

8. An automatic customer complaint processing device, characterized in that, include: The generation unit is used to generate a complaint work order in a preset format based on the customer's complaint information; The determining unit is used to determine a processing suggestion for the complaint work order based on the processing scheme of the historical complaint cases when the similarity between the complaint content of the complaint work order and the complaint content of historical complaint cases reaches a preset threshold. The first input unit is used to input the complaint form into a target model when the similarity between the complaint content of the complaint form and the complaint content of historical complaint cases does not reach a preset threshold. The target model then splits the complaint form into multiple words, converts each word into an embedding vector, adds positional encoding to each embedding vector, predicts processing suggestions for the complaint form based on each embedding vector, and outputs the processing suggestions for the complaint form. The target model is a model trained using the historical complaint cases. The processing unit is used to process the complaint work order according to the processing suggestions.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored executable program, wherein, when the executable program is executed, it controls the device on which the computer-readable storage medium is located to perform the automatic customer complaint processing method according to any one of claims 1 to 7.

10. An electronic device, characterized in that, include: Memory, which stores executable programs; A processor for running the program, wherein the program, when running, performs the automatic processing method for customer complaints as described in any one of claims 1 to 7.

11. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the automatic customer complaint processing method according to any one of claims 1 to 7.

Images