Customer service information upgrade early warning method and system

Abstract

The invention provides a customer service information upgrade early warning method and system, and the method comprises the following steps: carrying out the preprocessing of information data and an upgrade complaint condition, and obtaining final cleaning information data; performing statistical analysis on the correlation factors to obtain final correlation factors; training according to the final cleaning data and the final correlation factor to obtain an early warning model; grading the upgrade risk information to obtain an upgrade risk grade; and according to the upgrade risk level, carrying out upgrade inquiry early warning by using the early warning model. According to the technical scheme, the inquiry risk early warning model is established, the inquiry work order is divided according to different grades, and differential tracking processing is performed, so that cases of consultation upgrading complaints and complaint upgrading supervision are reduced, and the complaint processing quality and efficiency are improved; working pain points of the mechanism are solved, customer experience is improved, complaints of the mechanism are reduced, and dominant pressure is relieved.

Description

Background Art

[0002] Conventional warnings are linear formulas generated by combining universal factors with artificial factor scoring. Under the current severe regulatory environment, companies are in urgent need of quantifiable scenarios to improve consulting services and user satisfaction. However, customer service agents can only understand user satisfaction and intention to upgrade complaints through subjective feelings during user communication. The factors are too single and the accuracy is low. Summary of the invention

[0003] This application provides a customer service consultation and complaint upgrade early warning method and system to reduce the number of consultation upgrade complaints and complaint upgrade supervision cases, improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve owner pressure.

[0004] In a first aspect, a customer service complaint escalation early warning method is provided, comprising the following steps:

[0005] Pre-process the consultation and complaint data and upgraded complaint information to obtain the final cleaned consultation and complaint data;

[0006] Conduct statistical analysis on the correlation factors to obtain the final correlation factors;

[0007] An early warning model is obtained by training the final cleaned data and the final correlation factors;

[0008] Classify the escalation risk of consultation and complaint to obtain the escalation risk level;

[0009] According to the escalation risk level, the early warning model is used to carry out consultation and complaint escalation early warning.

[0010] In the above technical scheme, the final cleaned consultation and complaint data are obtained by pre-processing the consultation and complaint data and the upgraded complaint situation; the relevant factors are statistically analyzed to obtain the final relevant factors; the early warning model is trained based on the final cleaned data and the final relevant factors; the consultation and complaint upgrade risks are graded to obtain the upgrade risk level; according to the upgrade risk level, the early warning model is used to issue consultation and complaint upgrade warnings; by building a consultation and complaint risk early warning model, the consultation and complaint work orders are divided into different levels, and differentiated tracking and processing are carried out to reduce consultation upgrade complaints and complaint upgrade supervision cases, and improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve the pressure on owners.

[0011] In a specific implementation plan, the steps of pre-processing the consultation and complaint data and the upgraded complaint information to obtain the final cleaned consultation and complaint data are as follows:

[0012] Summarize the consultation and complaint data and upgraded complaint status to obtain summary consultation and complaint data;

[0013] The summarized consultation data is preprocessed to obtain final cleaned consultation data.

[0014] In a specific implementation scheme, the step of performing statistical analysis on the correlation factors to obtain the final correlation factors specifically includes:

[0015] Perform statistics on the correlation factors to obtain the initial correlation factors;

[0016] Screening the initial correlation factors to obtain remaining correlation factors;

[0017] Based on the remaining correlation factor, the final correlation factor is obtained using a prediction model algorithm.

[0018] In a specific implementation scheme, the relevant factors include complaint dimension, customer dimension and policy dimension.

[0019] In a specific implementation scheme, the step of screening the initial correlation factors to obtain the remaining correlation factors is specifically:

[0020] Based on the prediction model algorithm, the correlation between all the escalation situations of the consultation and the initial correlation factors is analyzed, and the initial correlation factors with low correlation are excluded to obtain the remaining correlation factors.

[0021] In a specific implementation scheme, based on the remaining correlation factor, the step of obtaining the final correlation factor using a prediction model algorithm is specifically as follows:

[0022] Based on the prediction model algorithm and the remaining correlation factors, the missing data supplementation scenarios of each factor in the final cleaning consultation data and the extreme factor peak shaving algorithm are planned to obtain the final correlation factors.

[0023] In a specific implementation plan, the steps of grading the escalation risk of the consultation and complaint to obtain the escalation risk level are as follows:

[0024] According to the actual distribution of consultation and complaint warning scores and the actual escalation situation, the high, medium and low escalation risk segments are confirmed to obtain the escalation risk level.

[0025] In a specific implementation scheme, according to the escalation risk level, the steps of using the early warning model to perform a consultation and complaint escalation early warning are specifically as follows:

[0026] According to the upgrade risk level, the early warning model is used to collect historical fluctuation complaint data within a specific time interval, and the upgrade score is calculated after data cleaning and missing factors are supplemented. The final warning score and risk segmentation are generated and notified to the customer service agent to complete the complaint upgrade warning.

[0027] In a specific embodiment, the specific time interval includes every four hours or every six hours.

[0028] Second, a customer service complaint upgrade early warning system is provided, including:

[0029] The consultation and complaint data module is used to pre-process the consultation and complaint data and upgrade the complaint status to obtain the final cleaned consultation and complaint data;

[0030] The correlation factor module is used to perform statistical analysis on the correlation factors to obtain the final correlation factors;

[0031] An early warning model module, used for training an early warning model according to the final cleaned data and the final correlation factors;

[0032] The risk level module is used to classify the escalation risk of consultation and complaint and obtain the escalation risk level;

[0033] The consultation and complaint warning module uses the warning model to issue consultation and complaint upgrade warnings according to the upgrade risk level.

[0034] In the above technical scheme, the final cleaned consultation and complaint data are obtained by pre-processing the consultation and complaint data and the upgraded complaint situation; the relevant factors are statistically analyzed to obtain the final relevant factors; the early warning model is trained based on the final cleaned data and the final relevant factors; the consultation and complaint upgrade risks are graded to obtain the upgrade risk level; according to the upgrade risk level, the early warning model is used to issue consultation and complaint upgrade warnings; by building a consultation and complaint risk early warning model, the consultation and complaint work orders are divided into different levels, and differentiated tracking and processing are carried out to reduce consultation upgrade complaints and complaint upgrade supervision cases, and improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve the pressure on owners. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A flowchart of a customer service complaint upgrade warning method provided in an embodiment of the present application;

[0036] Figure 2 A structural block diagram of the customer service complaint escalation warning system provided in an embodiment of the present application. DETAILED DESCRIPTION

[0037] The present application is further described in detail below through the accompanying drawings and embodiments. Through these descriptions, the characteristics and advantages of the present application will become clearer and more specific.

[0038] The word "exemplary" is used exclusively herein to mean "serving as an example, embodiment, or illustration." Any embodiment described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise noted.

[0039] In addition, the technical features involved in different embodiments of the present application described below can be combined with each other as long as they do not conflict with each other.

[0040] To facilitate understanding of the customer service consultation and complaint upgrade warning method and system provided in the embodiment of the present application, its application scenario is first explained. The customer service consultation and complaint upgrade warning method and system provided in the embodiment of the present application are used to reduce consultation and complaint upgrades, cases of complaint upgrade supervision, and improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve the pressure of the owner. Conventional warning is a linear formula generated by manually assigning points to factors in combination with general factors. Under the current severe regulatory background, companies are in urgent need of quantifiable scenarios to improve consultation and complaint services and improve user satisfaction, but customer service agents can only understand user satisfaction and intention to upgrade complaints through subjective feelings during user communication. The factors are too single and the accuracy is low. For this reason, the embodiment of the present application provides a customer service consultation and complaint upgrade warning method and system to reduce consultation and complaint upgrades, cases of complaint upgrade supervision, and improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve the pressure of the owner. The following is a detailed description of the embodiment in conjunction with specific drawings.

[0041] refer to Figure 1 and Figure 2 , Figure 1 A flowchart of a customer service complaint upgrade warning method provided in an embodiment of the present application; Figure 2 A structural block diagram of the customer service complaint escalation warning system provided in an embodiment of the present application.

[0042] exist Figure 1 In the embodiment of the present application, a customer service complaint upgrade warning method is provided, comprising the following steps:

[0043] Pre-process the consultation and complaint data and upgraded complaint information to obtain the final cleaned consultation and complaint data;

[0044] Conduct statistical analysis on the correlation factors to obtain the final correlation factors;

[0045] An early warning model is obtained by training the final cleaned data and the final correlation factors;

[0046] Classify the escalation risk of consultation and complaint to obtain the escalation risk level;

[0047] According to the escalation risk level, the early warning model is used to carry out consultation and complaint escalation early warning.

[0048] In the above technical scheme, the final cleaned consultation and complaint data are obtained by pre-processing the consultation and complaint data and the upgraded complaint situation; the relevant factors are statistically analyzed to obtain the final relevant factors; the early warning model is trained based on the final cleaned data and the final relevant factors; the consultation and complaint upgrade risks are graded to obtain the upgrade risk level; according to the upgrade risk level, the early warning model is used to issue consultation and complaint upgrade warnings; by building a consultation and complaint risk early warning model, the consultation and complaint work orders are divided into different levels, and differentiated tracking and processing are carried out to reduce consultation upgrade complaints and complaint upgrade supervision cases, and improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve the pressure on owners.

[0049] In a specific implementation plan, the steps of pre-processing the consultation and complaint data and the upgraded complaint information to obtain the final cleaned consultation and complaint data are as follows:

[0050] Summarize the consultation and complaint data and upgraded complaint status to obtain summary consultation and complaint data;

[0051] The summarized consultation data is preprocessed to obtain final cleaned consultation data.

[0052] In a specific implementation scheme, the step of performing statistical analysis on the correlation factors to obtain the final correlation factors specifically includes:

[0053] Perform statistics on the correlation factors to obtain the initial correlation factors;

[0054] Screening the initial correlation factors to obtain remaining correlation factors;

[0055] Based on the remaining correlation factor, the final correlation factor is obtained using a prediction model algorithm.

[0056] In a specific implementation scheme, the relevant factors include complaint dimension, customer dimension and policy dimension.

[0057] In a specific implementation scheme, the step of screening the initial correlation factors to obtain the remaining correlation factors is specifically:

[0058] Based on the prediction model algorithm, the correlation between all the escalation situations of the consultation and the initial correlation factors is analyzed, and the initial correlation factors with low correlation are excluded to obtain the remaining correlation factors.

[0059] In a specific implementation scheme, based on the remaining correlation factor, the step of obtaining the final correlation factor using a prediction model algorithm is specifically as follows:

[0060] Based on the prediction model algorithm and the remaining correlation factors, the missing data supplementation scenarios of each factor in the final cleaning consultation data and the extreme factor peak shaving algorithm are planned to obtain the final correlation factors.

[0061] It should be noted that: the prediction model algorithm includes:

[0062] 1. Regression analysis model: Research and analyze historical data to determine the correlation between the various factors that affect the prediction target and the target value, and use a statistical analysis method to estimate or predict by establishing a relevant mathematical model.

[0063] 2. Time series model: Data is a sequence of time and value, used to describe the past, analyze laws, and predict the future. Common algorithms:

[0064] Seasonal decomposition: Based on the changes in the data over the cycle, determine whether the series fluctuates greatly over the cycle, and then use multiplication and superposition models to perform seasonal decomposition.

[0065] Exponential smoothing: including simple smoothing, Holt linear trend model, damped trend model, simple seasonality and Winter model, etc. The choice of smoothing parameter α depends on the volatility of the time series.

[0066] ARIMA model: The full name is autoregressive difference moving average model, which consists of three parts: autoregression (AR), difference (I) and moving average (MA). It is suitable for stationary or non-stationary time series data that can be converted into stationary data, and can capture long-term trends and seasonal changes.

[0067] 3. Expert system model

[0068] The prediction model built based on expert knowledge and experience makes predictions by simulating the expert's decision-making process.

[0069] 4. Artificial Neural Network Model

[0070] The predictive model is constructed based on the structure and function of biological neural networks and has powerful learning and adaptability.

[0071] BP neural network: Back propagation neural network is a multi-layer feedforward neural network. It consists of an input layer, a hidden layer, and an output layer. The weights and biases are adjusted through the back propagation algorithm to minimize the prediction error. It can handle nonlinear relationships, but it has the disadvantages of high model complexity and easy to fall into local optimal solutions.

[0072] 5. Fuzzy prediction model

[0073] The prediction model based on fuzzy theory and fuzzy logic is suitable for processing uncertain and ambiguous data.

[0074] 6. Support Vector Machine Model

[0075] The prediction model built based on statistical learning theory and support vector machine algorithm is suitable for processing high-dimensional data and complex nonlinear relationships.

[0076] Support Vector Machine Regression (SVR): The resulting model depends only on a subset of the training data, as the cost function ignores samples whose predictions are close to their target. Support Vector Regression has three different implementations: SVR, NuSVR, and LinearSVR.

[0077] 7. Wavelet analysis model

[0078] The prediction model built based on wavelet transform and signal processing theory is suitable for processing data of non-stationary and sudden signal.

[0079] With the development of big data and computing power, predictive model algorithms are also constantly developing and innovating, providing more powerful tools for prediction and decision-making in various fields.

[0080] In a specific implementation plan, the steps of grading the escalation risk of the consultation and complaint to obtain the escalation risk level are as follows:

[0081] According to the actual distribution of consultation and complaint warning scores and the actual escalation situation, the high, medium and low escalation risk segments are confirmed to obtain the escalation risk level.

[0082] In a specific implementation scheme, according to the escalation risk level, the steps of using the early warning model to perform a consultation and complaint escalation early warning are specifically as follows:

[0083] According to the upgrade risk level, the early warning model is used to collect historical fluctuation complaint data within a specific time interval, and the upgrade score is calculated after data cleaning and missing factors are supplemented. The final warning score and risk segmentation are generated and notified to the customer service agent to complete the complaint upgrade warning.

[0084] In a specific embodiment, the specific time interval includes every four hours or every six hours.

[0085] Specifically, the specific process of the customer service complaint upgrade warning method is as follows:

[0086] Summarize the consultation and complaint data of the past year and upgrade the complaint status;

[0087] Statistical information on more than 50 factors including complaint latitude, customer latitude, policy latitude, etc.

[0088] Based on the prediction model algorithm, analyze the correlation between all escalation situations and factors, and exclude factors with low correlation;

[0089] Based on the prediction model algorithm and the remaining correlation factors, the missing data supplement scenarios of each factor in the planning data and the extreme factor peak shaving algorithm;

[0090] Perform early warning model training based on the final cleaned data and factors to generate an early warning model;

[0091] Determine the high, medium and low escalation risk segments based on the actual distribution of consultation and warning scores and actual escalation situations;

[0092] After the model is launched, historical fluctuation complaint data is collected every four hours, and the upgraded score is calculated after data cleaning and missing factors are supplemented. The final warning score and risk segmentation are generated and notified to the customer service agent.

[0093] In this embodiment, conventional warnings are linear formulas generated by manually assigning points to common factors, while the warning model of the embodiment of this application is combined with a big data algorithm, and each factor influences each other to generate a final warning model, and the warning score is more detailed and accurate. By building an intelligent classification and classification "Sky Eye" risk warning model for consultation and complaint, the consultation and complaint work orders are divided into different levels, and differentiated tracking and processing are carried out to reduce the number of consultation upgrades and complaints escalated to supervision, and improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve the pressure on owners.

[0094] It should be noted that:

[0095] 1. The application of big data algorithms in early warning models mainly includes the following aspects:

[0096] Data mining algorithms: Data mining algorithms can extract useful information and patterns from massive amounts of data to provide data support for early warning models. For example, through association rule mining algorithms, the correlation and dependency between data can be discovered, thereby predicting future trends and changes.

[0097] Machine learning algorithms: Machine learning algorithms can automatically extract features and patterns from data, improving the accuracy and intelligence of early warning models. Common machine learning algorithms include neural networks, support vector machines, decision trees, etc. These algorithms can perform operations such as classification, regression, and clustering on data to achieve accurate prediction and early warning of risks.

[0098] Statistical forecasting algorithm: Statistical forecasting algorithm predicts future development trends and change patterns through statistical analysis of historical data. For example, time series analysis algorithm can analyze the change trend of data over time and predict future data values. This algorithm is widely used in financial market forecasting, weather forecasting and other fields.

[0099] 2. The advantages of combining early warning models with big data algorithms are:

[0100] Real-time monitoring: The early warning model can monitor data changes in real time. Once an abnormality or potential risk is found, the early warning mechanism is immediately triggered to provide timely decision support for decision makers;

[0101] Accurate prediction: Big data algorithms can deeply explore the laws and patterns in the data and improve the prediction accuracy of the early warning model. By analyzing and learning from historical data, the model can predict future risk events and provide enterprises with targeted response measures;

[0102] High level of intelligence: The application of algorithms such as machine learning makes the early warning model more intelligent. The model can automatically extract features and patterns in the data, and perform self-optimization and updates to adapt to the ever-changing market environment and business needs.

[0103] exist Figure 2 In the embodiment of the present application, a customer service complaint upgrade warning system is provided, including:

[0104] The consultation and complaint data module is used to pre-process the consultation and complaint data and upgrade the complaint status to obtain the final cleaned consultation and complaint data;

[0105] The correlation factor module is used to perform statistical analysis on the correlation factors to obtain the final correlation factors;

[0106] An early warning model module, used for training an early warning model according to the final cleaned data and the final correlation factors;

[0107] The risk level module is used to classify the escalation risk of consultation and complaint and obtain the escalation risk level;

[0108] The consultation and complaint warning module uses the warning model to issue consultation and complaint upgrade warnings according to the upgrade risk level.

[0109] In the above technical scheme, the final cleaned consultation and complaint data are obtained by pre-processing the consultation and complaint data and the upgraded complaint situation; the relevant factors are statistically analyzed to obtain the final relevant factors; the early warning model is trained based on the final cleaned data and the final relevant factors; the consultation and complaint upgrade risks are graded to obtain the upgrade risk level; according to the upgrade risk level, the early warning model is used to issue consultation and complaint upgrade warnings; by building a consultation and complaint risk early warning model, the consultation and complaint work orders are divided into different levels, and differentiated tracking and processing are carried out to reduce consultation upgrade complaints and complaint upgrade supervision cases, and improve the quality and efficiency of complaint handling; solve the pain points of institutional work, improve customer experience, reduce the number of institutional complaints, and relieve the pressure on owners.

[0110] Those skilled in the art will appreciate that the present application may be implemented as a system, method or computer program product.

[0111] Therefore, the present disclosure may be specifically implemented in the following forms, namely: it may be completely hardware, it may be completely software (including firmware, resident software, microcode, etc.), or it may be a combination of hardware and software, generally referred to herein as a "circuit", "module" or "system". In addition, in some embodiments, the present application may also be implemented in the form of a computer program product in one or more computer-readable media, and the computer-readable medium may contain computer-readable program code.

[0112] Any combination of one or more computer-readable media can be used. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or devices, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: electrical connections with one or more wires, portable computer disks, hard disks, random access memories (RAM), read-only memories (ROM), erasable programmable read-only memories (EPROM or flash memory), optical fibers, portable compact disk read-only memories (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this document, computer-readable storage media can be any tangible medium containing or storing a program, which can be used by an instruction execution system, device or device or used in combination with it.

[0113] Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art can change, modify, replace and modify the above embodiments within the scope of the present application. On this basis, a variety of replacements and improvements can be made to the present application, all of which fall within the scope of protection of the present application.

Claims

1. A customer service complaint upgrade early warning method, characterized in that: The following steps are involved: Pre-process the consultation and complaint data and upgraded complaint information to obtain the final cleaned consultation and complaint data; Conduct statistical analysis on the correlation factors to obtain the final correlation factors; An early warning model is obtained by training the final cleaned data and the final correlation factors; Classify the escalation risk of consultation and complaint to obtain the escalation risk level; According to the escalation risk level, the early warning model is used to carry out consultation and complaint escalation early warning.

2. The customer service complaint escalation early warning method according to claim 1, characterized in that: Pre-process the consultation and complaint data and escalated complaint information to obtain the final cleansing consultation and complaint data, specifically: Summarize the consultation and complaint data and upgraded complaint status to obtain summary consultation and complaint data; The summarized consultation data is preprocessed to obtain final cleaned consultation data.

3. The customer service complaint escalation early warning method according to claim 2, characterized in that: The steps of performing statistical analysis on the correlation factors to obtain the final correlation factors specifically include: Perform statistics on the correlation factors to obtain the initial correlation factors; Screening the initial correlation factors to obtain remaining correlation factors; Based on the remaining correlation factor, the final correlation factor is obtained using a prediction model algorithm.

4. The customer service complaint escalation early warning method according to claim 3, characterized in that: The relevant factors include complaint dimension, customer dimension and policy dimension.

5. The customer service complaint escalation early warning method according to claim 4, characterized in that: The step of screening the initial correlation factors to obtain the remaining correlation factors is specifically: Based on the prediction model algorithm, the correlation between all the escalation situations of the consultation and the initial correlation factors is analyzed, and the initial correlation factors with low correlation are excluded to obtain the remaining correlation factors.

6. The customer service complaint escalation early warning method according to claim 5, characterized in that: Based on the remaining correlation factor, the step of obtaining the final correlation factor using a prediction model algorithm is specifically as follows: Based on the prediction model algorithm and the remaining correlation factors, the missing data supplementation scenarios of each factor in the final cleaning consultation data and the extreme factor peak shaving algorithm are planned to obtain the final correlation factors.

7. The customer service complaint escalation early warning method according to claim 6, characterized in that: The steps to categorize the escalation risk of consultation and complaint and obtain the escalation risk level are as follows: According to the actual distribution of consultation and complaint warning scores and the actual escalation situation, the high, medium and low escalation risk segments are confirmed to obtain the escalation risk level.

8. The customer service complaint escalation early warning method according to claim 7, characterized in that: According to the escalation risk level, the steps of using the early warning model to conduct consultation and complaint escalation early warning are specifically as follows: According to the upgrade risk level, the early warning model is used to collect historical fluctuation complaint data within a specific time interval, and the upgrade score is calculated after data cleaning and missing factors are supplemented. The final warning score and risk segmentation are generated and notified to the customer service agent to complete the complaint upgrade warning.

9. The customer service complaint escalation early warning method according to claim 8, characterized in that: The specific time interval includes every four hours and every six hours.

10. A customer service complaint upgrade warning system, characterized in that: include: The consultation and complaint data module is used to pre-process the consultation and complaint data and upgrade the complaint status to obtain the final cleaned consultation and complaint data; The correlation factor module is used to perform statistical analysis on the correlation factors to obtain the final correlation factors; An early warning model module, used for training an early warning model according to the final cleaned data and the final correlation factors; The risk level module is used to classify the escalation risk of consultation and complaint and obtain the escalation risk level; The consultation and complaint warning module uses the warning model to issue consultation and complaint upgrade warnings according to the upgrade risk level.

Images