User complaint processing method, apparatus, device, and storage medium

Abstract

The application discloses a user complaint processing method and device, equipment and a storage medium, and belongs to the technical field of data processing. The method comprises the following steps: obtaining a structured complaint template, a manually filled complaint work order of a target customer service account and user complaint voice; wherein the manually filled complaint work order and the user complaint voice correspond to each other; converting the user complaint voice into user complaint text information; filling in the structured complaint template based on the user complaint text information and the manually filled complaint work order, and obtaining a standard complaint work order. The application fills in the structured complaint template by using the user complaint text information and the manually filled complaint work order, can avoid the situation that key information is missed or random errors exist when a customer service personnel manually records complaint content, and makes the generated complaint work order match the actual complaint demand of a user.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a user complaint handling method, apparatus, device, and storage medium. Background Technology

[0002] With increasing consumer awareness of rights protection, the number of complaints from bank users regarding their services is also rising. The current process for handling user complaints at banks is as follows: users call a complaint hotline, and an operator manually records the work order; the work order and the transcript of the complaint are then forwarded to the complex issues team for manual analysis and confirmation of the complaint content (service, business, product, etc.), before being forwarded to the responsible department for follow-up processing.

[0003] However, in the current complaint handling process, customer service personnel are prone to omitting key information or making random errors when manually filling out complaint forms. For example, if the customer service staff makes a mistake in inputting the information, the generated complaint form may not match the user's actual complaint needs. Summary of the Invention

[0004] The main purpose of this application is to provide a user complaint handling method, apparatus, device, and storage medium, which aims to solve the technical problem of mismatch between generated complaint work orders and actual user complaint needs.

[0005] To achieve the above objectives, this application provides a user complaint handling method, comprising:

[0006] Obtain a structured complaint template, a manually filled complaint form from the target customer service account, and a user complaint voice message; wherein the manually filled complaint form and the user complaint voice message correspond to each other;

[0007] Convert the user's voice complaint into a text complaint message;

[0008] Based on the user complaint text information and the manually filled complaint form, a structured complaint template is filled out to obtain a standard complaint form.

[0009] Optionally, the step of filling out a structured complaint template based on the user complaint text information and the manually filled complaint form to obtain a standard complaint form includes:

[0010] The user complaint text information is extracted according to a preset complaint vocabulary to obtain the first key information, and the manually filled complaint work order is extracted to obtain the second key information.

[0011] Identify distinguishing keywords from the first key information that differ from the second key information, as well as common keywords that are the same as the second key information from the first key information;

[0012] If the total number of items of the first keyword in the first key information is the same as the total number of items of the second keyword in the second key information, then the structured complaint template is filled in based on the distinguishing keyword and the common keyword to obtain the standard complaint work order;

[0013] If the total number of items of the first keyword in the first key information is inconsistent with the total number of items of the second keyword in the second key information, then the structured complaint template is filled in based on the distinguishing keyword and the second key information to obtain the standard complaint work order.

[0014] Optionally, identifying distinguishing keywords from the first key information that differ from the second key information, and common keywords that are the same as the second key information, from the first key information includes:

[0015] Calculate the feature distance between each of the first keyword terms and each of the second keyword terms;

[0016] Based on the feature distance, distinguishing keywords that are different from the second key information and common keywords that are the same as the second key information are identified from the first key information.

[0017] Optionally, after converting the user complaint voice into user complaint text information, the method further includes:

[0018] Extract sentiment keywords from the user complaint text information;

[0019] Based on the aforementioned sentiment keywords, the standard complaint work orders are sorted to obtain a sorting order, so that staff can process the standard complaint work orders according to the sorting order.

[0020] Optionally, after obtaining a standard complaint form by filling in the structured complaint template based on the user complaint text information and the manually filled complaint form, the method further includes:

[0021] The complaint work order judgment model is used to determine whether the standard complaint work order is a complaint work order. The complaint work order judgment model is an NLP language model.

[0022] If the standard complaint work order is a complaint work order, then the standard complaint work order is input into the complaint work order classification model for classification, and the classification result output by the complaint work order classification model is obtained.

[0023] Optionally, after determining whether the standard complaint work order is a complaint work order, the method further includes:

[0024] If the standard complaint form is not a complaint form, it will be sent to a human for manual confirmation.

[0025] If a manually entered complaint work order confirmation is received, the process of inputting the standard complaint work order into the complaint work order classification model for classification is executed, and the classification result output by the complaint work order classification model is obtained.

[0026] Optionally, after obtaining a standard complaint form by filling in the structured complaint template based on the user complaint text information and the manually filled complaint form, the method further includes:

[0027] Based on the type of the standard complaint work order, the processing time of the standard complaint work order, and the complaint indicators of the standard complaint work order, a complaint system report is generated.

[0028] Secondly, this application provides a user complaint processing device, comprising:

[0029] The acquisition module is used to acquire structured complaint templates, manually filled complaint forms from the target customer service account, and user complaint voice recordings; wherein the manually filled complaint forms and user complaint voice recordings correspond to each other.

[0030] The conversion module is used to convert the user's voice complaint into text information.

[0031] The filling module is used to fill in the structured complaint template based on the user complaint text information and the manually filled complaint work order to obtain a standard complaint work order.

[0032] Thirdly, this application provides a user complaint processing device, including a memory, a processor, and a user complaint processing program stored on the processor and executable on the processor, the user complaint processing program being configured to implement the steps of the user complaint processing method described above.

[0033] Fourthly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when processed, executes the user complaint processing method of any embodiment of this application.

[0034] The user complaint handling method proposed in this application, compared with the prior art where customer service personnel manually input key information from user complaint calls, converts user complaint voice into user complaint text information, and then combines the user complaint text information with manually filled-out complaint forms to fill in a structured complaint template to obtain a standard complaint form. In other words, it uses user complaint text information to improve the information omitted by customer service personnel in manually filled-out complaint forms, avoiding situations where key information is omitted or random errors occur when customer service personnel manually record complaint content, resulting in a higher degree of matching between the generated complaint form and the user's actual complaint needs. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the hardware structure of a user complaint processing device according to an embodiment of this application;

[0036] Figure 2 This is a flowchart illustrating the first embodiment of the user complaint handling method of this application;

[0037] Figure 3 This is a flowchart illustrating the second embodiment of the user complaint handling method of this application;

[0038] Figure 4 This is a flowchart illustrating the third embodiment of the user complaint handling method of this application;

[0039] Figure 5 This is a flowchart illustrating the fourth embodiment of the user complaint handling method of this application;

[0040] Figure 6 This is a schematic diagram of the functional modules of the user complaint handling and identification device in this application.

[0041] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0042] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0043] Due to the limitations of current technology, customer service personnel may omit key information or make random errors when manually filling out complaint forms in the current complaint handling process. For example, if customer service staff make input errors, the generated complaint form may not match the user's actual complaint needs.

[0044] To address this, this application provides a solution that, compared to existing technologies where customer service personnel manually input key information from user complaint calls, converts user complaint voice into user complaint text. This text is then combined with manually filled-out complaint forms to fill out a structured complaint template, resulting in a standard complaint form. This approach utilizes the user complaint text to supplement information omitted by customer service personnel in manually filled-out complaint forms, preventing situations where key information is missed or random errors occur when customer service personnel manually record complaint content. This results in a higher degree of matching between the generated complaint form and the user's actual complaint needs.

[0045] Reference Figure 1 , Figure 1 This is a schematic diagram of the structure of a user complaint processing device in the hardware operating environment involved in the embodiments of this application.

[0046] like Figure 1 As shown, the user complaint processing device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0047] Those skilled in the art will understand that Figure 1 The structure shown does not constitute a limitation on the user complaint handling device and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0048] like Figure 1 As shown, the memory 1005, which serves as a storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and a user complaint handling program.

[0049] exist Figure 1In the user complaint processing device shown, the network interface 1004 is mainly used for data communication with the network server; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the user complaint processing device of this application can be set in the user complaint processing device. The user complaint processing device calls the user complaint processing program stored in the memory 1005 through the processor 1001 and executes the user complaint processing method provided in the embodiment of this application.

[0050] Based on, but not limited to, the hardware structure of the user complaint processing device described above, this application provides a first embodiment of a user complaint processing method. (Refer to...) Figure 2 , Figure 2 A flowchart illustrating the first embodiment of the user complaint handling method is shown.

[0051] It should be noted that although the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0052] In this embodiment, the user complaint handling method includes:

[0053] Step S10: Obtain the structured complaint template, the manually filled complaint form from the target customer service account, and the user's complaint voice recording.

[0054] Among them, there is a correspondence between manually filled-out complaint forms and user complaint voice messages.

[0055] It should be noted that the user complaint handling method is implemented by the user complaint handling device, which is equipped with a user complaint handling program. When the user complaint handling device executes the user complaint handling program, it implements the user complaint handling method of this application.

[0056] It's understandable that users may encounter unpleasant situations when using services provided by service providers, leading to user complaints. Generally, users can establish a connection with the service provider's customer service by calling a complaint hotline. The user then describes the situation (the problem) and the issues and / or compensation they require from the service provider (the user's actual complaint needs) via voice. During this call, the customer service representative can ask the user if they agree to have their complaint recorded. With the user's consent, the call can be recorded, resulting in the user's complaint voice recording. After the call ends, the user's complaint voice recording can be saved locally or in the cloud.

[0057] Then, customer service personnel can log in to their own customer service account (the target customer service account) on the service provider's internal system, call up the corresponding complaint form template, fill in and submit the complaint form template, and thus obtain a manually filled complaint form.

[0058] Structured complaint templates are reusable documents that can be pre-configured by practitioners, which helps save time on filling out user complaint text information and manual complaint forms, and standardizes the format of complaint forms.

[0059] A structured complaint template can consist of five fields: user information, complaint text, departmental handling opinion, handling result, and follow-up result. Of course, a structured complaint template can include more fields, such as complaint type information and complaint form information. For example, the complaint type can be service-related or business-related, and the complaint form can be a telephone complaint or an in-person complaint.

[0060] In this embodiment, the structured complaint template can be pre-saved locally or in the cloud in the user complaint processing device. When step S10 is executed, the user complaint processing device may be triggered to directly call the structured complaint template and user complaint voice from the local device or the cloud when it receives a manually filled complaint form submitted by the target customer service account.

[0061] Alternatively, when the user complaint processing device is scheduled or other triggering conditions configured by the administrator are met, it can call the manually filled complaint form submitted by the target customer service account and call the structured complaint template and user complaint voice from local or cloud.

[0062] Step S20: Convert the user's voice complaint into a text message.

[0063] In this embodiment, ASR (Automatic Speech Recognition) technology can be used to convert user complaint speech into user complaint text information.

[0064] Specifically, when using ASR technology to convert user complaint speech into user complaint text information, the STT (Speech-to-Text) algorithm is mainly employed. This involves utilizing CNN (Convolutional Neural Networks) and RNN (Recurrent Neural Network) models, with CTC (Connectionist Temporal Classification) as the loss function to construct a speech-to-text network for segmenting the word structure of the speech. The CNN model processes the speech input spectrogram and outputs a speech feature map. The RNN model is a recurrent network composed of multiple bidirectional LSTM (Long Short-Term Memory) networks, further processing the speech feature map into character sequences at different times or frames. The CTC loss function is used to calculate character probabilities, deriving the correct character sequence. Finally, a Softmax normalization layer is used to generate the probability of producing a character at each step and align it with the character sequence, outputting the final translation result of the speech-to-text network, i.e., the user complaint text information.

[0065] Step S30: Fill in the structured complaint template based on the user complaint text information and the manually filled complaint work order to obtain a standard complaint work order.

[0066] After obtaining user complaint text information and manually filled complaint forms, the original information recorded in the user complaint text information and the information processed by customer service recorded in the manually filled complaint forms can be combined to fill in the blank content in the structured complaint template, so as to obtain a standard complaint form that does not omit complaint information and can fully reflect the actual complaint needs of users.

[0067] Furthermore, as an optional implementation, step S30 specifically includes:

[0068] Step S301: Extract key information from user complaint text information according to the preset complaint vocabulary list to obtain the first key information; extract key information from manually filled complaint work orders to obtain the second key information.

[0069] In this embodiment, the preset complaint vocabulary includes multiple preset complaint terms. These terms reflect the problems encountered by users or the compensation requests they seek. The preset complaint vocabulary can be pre-configured by professionals based on their experience. For example, after establishing a call connection with a customer, bank customer service personnel can identify multiple keywords based on the customer's voice description, such as "money not credited in real time" or "bank account abnormality." Alternatively, the preset complaint vocabulary can be generated by extracting key information from historical user complaint text information manually or through user complaint devices.

[0070] Understandably, after converting user complaints from voice to text, the text may contain not only the primary key information related to the user's request, but also irrelevant information such as complaints or grumbles. Therefore, keywords can be extracted from the user complaint text based on a pre-defined complaint vocabulary to quickly and accurately identify the primary keyword information.

[0071] It should be noted that there may be multiple first keywords in the first key information extracted from the user complaint text information, but the extracted first keywords may not be closely related to the user's complaint. Therefore, in order to avoid the influence of other keywords on the first key information and improve the relevance between the first key information and the user complaint text, as an optional implementation method, keywords can be extracted from the user complaint text information according to the TF-IDF (Term Frequency-inverse Document Frequency) algorithm and a preset complaint vocabulary.

[0072] In this embodiment, customer service personnel can process user complaint voice messages based on a preset complaint vocabulary list. That is, when filling out the complaint form template, customer service personnel can simply enter keywords, allowing for direct extraction of key information from manually filled complaint forms. Alternatively, the steps for extracting key information from user complaint text can be referenced, which will not be elaborated upon in this embodiment.

[0073] Step S302: Identify the distinguishing keywords that are different from the second key information and the common keywords that are the same as the second key information from the first key information.

[0074] In this embodiment, distinguishing keywords and common keywords can be identified by comparing each first keyword in the first key information with each second keyword in the second key information.

[0075] Of course, since customer service representatives naturally process the user's voice complaint when filling out the complaint form, the first and second keywords for the same complaint content may not be completely consistent. To avoid misclassifying such first and second keywords as different keywords during comparison, as an optional implementation, step S302 specifically includes:

[0076] Step A10: Calculate the feature distance between the first keyword and the second keyword of each item.

[0077] Step A20: Based on the feature distance, identify the distinguishing keywords that are different from the second key information and the common keywords that are the same as the second key information from the first key information.

[0078] In this embodiment, the feature distance can be used to represent the similarity between the first keyword and the second keyword.

[0079] Specifically, the first and second keywords can be input into the DSSM (Deep Structured Semantic Models) model. The DSSM model calculates the feature distance between the first and second keywords. First and second keywords with feature distances less than a first preset value are considered to reflect the same complaint content, i.e., they are identical shared keywords. Of course, it is also possible to identify first keywords whose feature distances to multiple second keywords are all greater than the second preset value, meaning that there are no second keywords reflecting the same complaint content for that first keyword, thus obtaining distinguishing keywords.

[0080] Alternatively, other methods can be used to calculate the feature distance between each first keyword and each second keyword; this embodiment does not limit this method.

[0081] Step S303: If the total number of items of the first keyword in the first key information is the same as the total number of items of the second keyword in the second key information, then fill in the structured complaint template based on the distinguishing keywords and common keywords to obtain a standard complaint work order.

[0082] Step S304: If the total number of items of the first keyword in the first key information is inconsistent with the total number of items of the second keyword in the second key information, then fill in the structured complaint template based on the distinguishing keywords and the second key information to obtain a standard complaint work order.

[0083] It should be noted that the total number of items for the first keyword in the first key information and the total number of items for the second keyword in the second key information may or may not be the same.

[0084] Specifically, when the total number of items in the first key information is the same as the total number of items in the second key information, it can be determined that the customer service personnel did not omit any key information when manually filling out the complaint form. However, in order to avoid the situation where the key information of the complaint is incorrect due to the input error of the customer service staff, the structured complaint template can be filled out according to the distinguishing keywords and common keywords to obtain a standard complaint form.

[0085] When the total number of items for the first keyword in the first key information does not match the total number of items for the second keyword in the second key information, it can be determined that customer service personnel omitted key information when manually filling out the complaint form. To avoid omitting key complaint information, the structured complaint template should be filled out according to the distinguishing keywords and the second key information to comprehensively cover the user's complaint requests.

[0086] Furthermore, as an optional implementation, after obtaining the user complaint text information, the FastText algorithm can be used in the NLP algorithm model to classify the user complaint text and obtain text classification labels. Specifically, the received word vectors are transmitted to a CNN network structure, and word vector labels, i.e., text classification labels, are output. It should be noted that since the NLP algorithm model cannot directly process unstructured text, it is necessary to convert the unstructured text into structured text. That is, the first key information is extracted from the user complaint text information using a preset complaint vocabulary, a dedicated keyword lexicon is generated based on the first key information, word vectors are generated by combining the keyword lexicon, and the word vectors are input into the NLP algorithm model to obtain text classification labels.

[0087] In this embodiment, after obtaining text classification tags based on user complaint text information, standard complaint work orders can be assigned to the corresponding processing departments according to the text classification tags, so as to resolve user problems more quickly and accurately and improve user experience.

[0088] Furthermore, as an optional implementation, in addition to extracting the first key information from the user complaint text information using a preset complaint vocabulary to generate structured text, it is also necessary to perform data cleaning and text correction on the first key information. Finally, the text classification results, work order association attributes, and complaint information from other channels are combined and summarized into the information input and collection module to obtain formatted and unified structured complaint data, i.e., standard complaint work orders. Based on the type of standard complaint work order, the standard complaint work order is assigned to the corresponding processing department.

[0089] In this embodiment, after converting user complaints from voice to text, a structured complaint template is filled out based on the user complaint text and manually filled-out complaint forms to obtain a standard complaint form. This process uses the user complaint text to supplement information omitted by customer service personnel in manually filled-out complaint forms, preventing situations where key information is missed or random errors occur when customer service personnel manually record complaint content. This ensures that the generated complaint form matches the user's actual complaint needs. Simultaneously, the user complaint text is categorized, and the standard complaint form is assigned to the appropriate processing department based on the categorization results. This allows for faster and more accurate resolution of user issues, improving the user experience.

[0090] Furthermore, a second embodiment of the user complaint handling method of this application is proposed.

[0091] Reference Figure 3 In this embodiment, after step S30, the method further includes:

[0092] Step S31: Extract sentiment keywords from user complaint text information.

[0093] Step S32: Based on sentiment keywords, sort the standard complaint work orders to obtain the sorting order of the standard complaint work orders, so that staff can process the standard complaint work orders based on the sorting order.

[0094] Understandably, when a user makes a complaint, their current emotion can be determined by their choice of words. For example, words expressing emotion such as "I'm very angry" or "I'm very disappointed" can be extracted from the user's complaint text. Alternatively, as an optional implementation, to determine the user's emotional level at the time, such as angry or extremely angry, words indicating degree can be extracted from the user's complaint text. For instance, if the user's complaint text is "I'm extremely angry," the extracted emotional keywords could be "extremely" and "angry" to indicate the user's level of anger.

[0095] Specifically, after extracting sentiment keywords from user complaint texts, standard complaint tickets can be sorted according to these keywords to obtain the order of quasi-complaint tickets. That is, the urgency of standard complaint tickets is determined based on sentiment keywords. For example, when the sentiment keywords are "angry" or "very angry," the standard complaint tickets corresponding to the sentiment keyword "very angry" are placed first, and the standard complaint tickets corresponding to the sentiment keyword "angry" are placed second. Based on the urgency of the standard complaint tickets, a reasonable order for processing them is determined. That is, standard complaint tickets with very urgent urgency are processed first, while standard complaint tickets with moderate urgency are processed according to their order of priority.

[0096] Alternatively, as an optional implementation, after acquiring the user's complaint voice, emotion spectrum analysis can be performed on the voice information to obtain the user's emotion analysis results. Specifically, after acquiring the user's complaint voice, the user's complaint voice is divided into multiple voice segments according to a preset unit time. The number of syllables in each voice segment is detected, and the maximum number of syllables is determined from the multiple syllable counts. Then, based on the number of syllables in the multiple voice segments and multiple preset unit time periods, the average number of syllables is obtained. The user's emotion is determined based on the difference between the average number of syllables and the maximum number of syllables. At the same time, it is also necessary to determine the user's volume level in each preset unit time period, and determine the maximum volume and the average volume from the multiple volumes. The user's emotion is determined based on the difference between the average volume and the maximum volume. For example, if the user speaks quickly and loudly during the interaction, the difference between the average number of syllables and the maximum number of syllables is close to zero, and the difference between the average volume and the maximum volume is also close to zero. In this case, it can be determined that the user's emotion is agitated when making the complaint, indicating that the event is important to the user and needs to be dealt with quickly.

[0097] Furthermore, as an optional implementation method, after obtaining a standard complaint form, the user's complaint background and reasons can be further divided by combining the user profile and the user's transaction information in the bank, so as to more accurately determine the user's complaint demands.

[0098] In this embodiment, the user's emotional state at the time of complaint is determined based on sentiment keywords, and standard complaint tickets are sorted according to these keywords. Complaint tickets with higher ranking can be assigned to departments for expedited processing to quickly resolve user issues. Simultaneously, this application integrates artificial intelligence technology, combining complaint data, customer information, and other data sources to intelligently recommend strategies for processing user complaint information, and provides 24 / 7 real-time monitoring of the entire process, enabling rapid and high-quality responses to user complaints and improving service quality.

[0099] Furthermore, a third embodiment of the user complaint handling method of this application is proposed, referring to... Figure 4 .

[0100] In this embodiment, step S30 is followed by:

[0101] Step S33: Use the complaint work order judgment model to determine whether a standard complaint work order is a complaint work order. The complaint work order judgment model is an NLP language model.

[0102] Step S34: If the standard complaint work order is a complaint work order, then input the standard complaint work order into the complaint work order classification model for classification, and obtain the classification result output by the complaint work order classification model.

[0103] It is understandable that there may be some errors in the standard complaint work orders output from the complaint information input, collection, and processing process; that is, there may be work orders that are incorrectly classified as complaints. To avoid the occurrence of incorrectly classified complaint work orders, it is also necessary to use a complaint work order judgment model to determine whether a standard complaint work order is indeed a complaint work order.

[0104] Specifically, when the complaint ticket judgment model is an NLP language model, after inputting the standard complaint ticket into the NLP language model, the third key information in the standard complaint ticket is extracted according to the preset complaint vocabulary. The third key information is then verified using the second key information. When each second key word in the second key word matches each third key word in the third key information, that is, when the similarity between each second key word and each third key word is high, the standard complaint ticket is then input into the complaint ticket classification model to classify the standard complaint ticket, obtain the classification result output by the complaint ticket classification model, and assign the standard complaint ticket to the corresponding processing department according to the classification result.

[0105] In this embodiment, judging standard complaint work orders can avoid subsequently assigning non-complaint work orders to processing departments, thereby reducing the workload of staff. Simultaneously, when a standard complaint work order is classified as a complaint work order, it is input into a complaint work order classification model for classification. Based on the classification results, the standard complaint work order is assigned to the corresponding processing department. This not only allows for quick and accurate identification of the processing department and shortens the processing cycle for standard complaint work orders, but also significantly reduces the overall efficiency. Specifically, before adopting this method, the number of complaints processed annually was over 80,000, with an average processing time of 5 days per complaint and an average classification time of 210 seconds per standard complaint work order. After adopting this method, the number of complaints processed annually is over 500,000, with an average processing time of 3.5 days per complaint and an average classification time of 1.5 seconds per standard complaint work order. Furthermore, it reduces human intervention errors, enabling efficient and stable analysis and processing of standard complaint work orders, further improving workflow efficiency.

[0106] Furthermore, a fourth embodiment of the user complaint handling method of this application is proposed, referring to... Figure 5 .

[0107] In this embodiment, after step S33, the following steps are also included:

[0108] Step S35: If the standard complaint work order is not a complaint work order, then output it to a human for manual confirmation of the complaint work order.

[0109] Step S36: If a manually input complaint work order confirmation message is received, then the standard complaint work order is input into the complaint work order classification model for classification, and the classification result output by the complaint work order classification model is obtained.

[0110] In this embodiment, to avoid misjudgment leading to untimely handling of user complaints, when a standard complaint work order is not a complaint work order, it is sent to the staff. After receiving the standard complaint work order, the staff can determine whether the standard complaint work order is a complaint work order based on the preset complaint glossary and the content recorded in the standard complaint work order. Based on the determination result, the staff can operate the corresponding controls to generate complaint work order confirmation information, that is, the standard complaint work order is not a complaint work order, or the standard complaint work order is a complaint work order.

[0111] Understandably, when a standard complaint ticket is not a complaint ticket, it is closed. When a standard complaint ticket is a complaint ticket, it can be sent to the appropriate department for processing, or it can be input into a complaint ticket classification model for classification. The classification results are then used to assign the standard complaint ticket to the corresponding department. Simultaneously, the standard complaint ticket is input again into the complaint ticket judgment model for retraining, further optimizing the model and improving its accuracy.

[0112] Furthermore, a sixth embodiment of the user complaint handling method of this application is proposed.

[0113] In this embodiment, step S30 is followed by:

[0114] Step S37: Generate a complaint system report based on the type of standard complaint work order, the processing time of the standard complaint work order, and the complaint indicators of the standard complaint work order.

[0115] In this embodiment, the type of a standard complaint form can be determined based on its content. For example, if the content of a standard complaint form involves RMB savings, the type is "Amount Savings"; if it involves credit cards, the type is "Credit Card". The processing time for a standard complaint form can be determined based on its generation time and closure time. Complaint metrics can include completion rate, feedback rate, satisfaction rate, response rate, and transfer rate.

[0116] In this embodiment, statistical analysis is performed based on parameters such as the type of standard complaint work order, processing time, and complaint indicators, and a complaint system report is generated. The complaint process can be monitored and managed from different dimensions, and shortcomings can be addressed in a timely manner. For example, if the processing time is too long, corresponding measures are taken to shorten the processing time.

[0117] Based on the same application concept, this application proposes a user complaint handling device, referring to... Figure 6 , Figure 6This is a schematic diagram of the modules of the first embodiment of the user complaint handling device of this application.

[0118] The acquisition module is used to acquire structured complaint templates, manually filled complaint forms from the target customer service account, and user complaint voice recordings; among them, the manually filled complaint forms and user complaint voice recordings correspond to each other.

[0119] The conversion module is used to convert user complaint voice into user complaint text information;

[0120] The form-filling module is used to fill in structured complaint templates based on user complaint text information and manually filled complaint forms to obtain standard complaint forms.

[0121] It should be noted that, in this embodiment, the various implementation methods of the user complaint handling device and the technical effects they achieve can be referred to the various implementation methods of the user complaint handling method in the foregoing embodiments, and will not be repeated here.

[0122] The technical solution of this embodiment, through the cooperation of various functional modules, obtains a structured complaint template, a manually filled complaint form from the target customer service account, and user complaint voice recordings. The manually filled complaint form and the user complaint voice recordings correspond to each other. The user complaint voice recordings are converted into user complaint text information. Based on the user complaint text information and the manually filled complaint form, the structured complaint template is filled out to obtain a standard complaint form. Compared to the prior art, where customer service personnel manually input key information from user complaint calls, this application converts user complaint voice recordings into user complaint text information and combines this text information with the manually filled complaint form to fill out the structured complaint template, obtaining a standard complaint form. This means that the user complaint text information is used to complete information omitted by customer service personnel in manually filled complaint forms, avoiding situations where key information is omitted or random errors occur when customer service personnel manually record complaint content. This results in a higher degree of matching between the generated complaint form and the user's actual complaint needs.

[0123] Furthermore, embodiments of this application also propose a computer storage medium storing a user complaint handling program. When executed by a processor, the user complaint handling program implements the steps of the user complaint handling method described above. Therefore, it will not be repeated here. Additionally, the beneficial effects of using the same method will not be repeated here either. For technical details not disclosed in the computer-readable storage medium embodiments of this application, please refer to the description of the method embodiments of this application. As an example, program instructions can be deployed to execute on a single computing device, or on multiple computing devices located at one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0124] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0125] It should also be noted that the device embodiments described above are merely illustrative. 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 may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided in this application, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0126] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware, or it can be implemented by special-purpose hardware including application-specific integrated circuits, special-purpose CPUs, special-purpose memory, special-purpose components, etc. Generally, any function performed by a computer program can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be diverse, such as analog circuits, digital circuits, or special-purpose circuits. However, for this application, software program implementation is more often the preferred implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a readable storage medium, such as a computer floppy disk, USB flash drive, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of this application.

[0127] The above are merely preferred embodiments of this application and do not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

Claims

1. A method for handling user complaints, characterized in that, The method includes: Obtain a structured complaint template, a manually filled complaint form from the target customer service account, and a user complaint voice message; wherein the manually filled complaint form and the user complaint voice message correspond to each other; Convert the user's voice complaint into a text complaint message; Based on the user complaint text information and the manually filled complaint form, the structured complaint template is filled out to obtain a standard complaint form; The step of filling out the structured complaint template based on the user complaint text information and the manually filled complaint form to obtain a standard complaint form includes: The user complaint text information is extracted according to a preset complaint vocabulary to obtain the first key information, and the manually filled complaint work order is extracted to obtain the second key information. Calculate the feature distance between each first keyword in the first key information and each second keyword in the second key information; Based on the feature distance, distinguishing keywords that are different from the second key information and common keywords that are the same as the second key information are identified from the first key information. If the total number of items of the first keyword in the first key information is the same as the total number of items of the second keyword in the second key information, then the structured complaint template is filled in based on the distinguishing keyword and the common keyword to obtain the standard complaint work order; If the total number of items of the first keyword in the first key information is inconsistent with the total number of items of the second keyword in the second key information, then the structured complaint template is filled in based on the distinguishing keyword and the second key information to obtain the standard complaint work order.

2. The user complaint handling method according to claim 1, characterized in that, After converting the user complaint voice into user complaint text information, the method further includes: Extract sentiment keywords from the user complaint text information; Based on the aforementioned sentiment keywords, the standard complaint work orders are sorted to obtain a sorting order, so that staff can process the standard complaint work orders according to the sorting order.

3. The user complaint handling method according to claim 1, characterized in that, After obtaining a standard complaint form by filling in the structured complaint template based on the user complaint text information and the manually filled complaint form, the method further includes: The complaint work order judgment model is used to determine whether the standard complaint work order is a complaint work order. The complaint work order judgment model is an NLP language model. If the standard complaint work order is a complaint work order, then the standard complaint work order is input into the complaint work order classification model for classification, and the classification result output by the complaint work order classification model is obtained.

4. The user complaint handling method according to claim 3, characterized in that, After determining whether the standard complaint work order is a complaint work order, the method further includes: If the standard complaint form is not a complaint form, it will be sent to a human for manual confirmation. If a manually entered complaint work order confirmation is received, the process of inputting the standard complaint work order into the complaint work order classification model for classification is executed, and the classification result output by the complaint work order classification model is obtained.

5. The user complaint handling method according to claim 1, characterized in that, After obtaining a standard complaint form by filling in the structured complaint template based on the user complaint text information and the manually filled complaint form, the method further includes: Based on the type of the standard complaint work order, the processing time of the standard complaint work order, and the complaint indicators of the standard complaint work order, a complaint system report is generated.

6. A user complaint handling device, characterized in that, The user complaint handling device includes: The acquisition module is used to acquire structured complaint templates, manually filled complaint forms from the target customer service account, and user complaint voice recordings; wherein the manually filled complaint forms and the user complaint voice recordings correspond to each other. The conversion module is used to convert the user's voice complaint into text information. The filling module is used to fill in the structured complaint template based on the user complaint text information and the manually filled complaint work order to obtain a standard complaint work order; The filling module is also used for: The user complaint text information is extracted according to a preset complaint vocabulary to obtain the first key information, and the manually filled complaint work order is extracted to obtain the second key information. Calculate the feature distance between each first keyword in the first key information and each second keyword in the second key information; Based on the feature distance, distinguishing keywords that are different from the second key information and common keywords that are the same as the second key information are identified from the first key information. If the total number of items of the first keyword in the first key information is the same as the total number of items of the second keyword in the second key information, then the structured complaint template is filled in based on the distinguishing keyword and the common keyword to obtain the standard complaint work order; If the total number of items of the first keyword in the first key information is inconsistent with the total number of items of the second keyword in the second key information, then the structured complaint template is filled in based on the distinguishing keyword and the second key information to obtain the standard complaint work order.

7. A user complaint handling device, characterized in that, include: A processor, a memory, and a user complaint handling program stored in the memory, wherein the user complaint handling program is executed by the processor to implement the steps of the user complaint handling method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a user complaint handling program, which, when executed by a processor, implements the user complaint handling method as described in any one of claims 1 to 5.

Images