Outbound call forwarding method and system for realizing sectional manual forwarding

Abstract

The invention discloses an outbound call forwarding method and system for realizing segmented manual forwarding, and relates to the technical field of call centers. The method comprises the following steps: acquiring service processing information of each manual transfer service in a historical time period and current appeal information of a target user, wherein the service processing information comprises at least one of call duration, question type, question content, docking customer service and satisfaction score; based on the service processing information of each manual transfer service in the historical time period, determining the capability expression degree of each customer service; based on the current appeal information and each ability expression degree, respectively determining a first adaptation degree between each customer service and the target user; and based on the first adaptation degree between each customer service and the target user, performing customer service allocation on the target user. According to the invention, the distribution accuracy of the manual customer service can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of call centers, and particularly to an outbound transfer method and system for realizing segmented manual transfer. Background Art

[0002] In current application scenarios such as call centers, customer service systems, and sales outbound calls, the outbound transfer method of manual transfer has been widely adopted. The core of this method is that it can accurately transfer the customer's phone call or relevant information to the corresponding service link or professional service personnel according to the actual situation, so as to provide a more refined service experience that meets the customer's personalized needs.

[0003] At present, with the help of technologies such as intelligent voice assistants and automated queries, the front-end service links in user requirements have been automated. Only those more complex and special problems will be transferred to the manual customer service. This approach not only greatly reduces the labor cost but also effectively improves the customer satisfaction.

[0004] However, due to the complexity, diversity, and particularity of customer problems, as well as the uneven professional capabilities of manual customer service in different types of problems, in actual operation, the accuracy of manual customer service allocation is poor. Summary of the Invention

[0005] The embodiments of the present invention provide an outbound transfer method and system for realizing segmented manual transfer, which can improve the accuracy of manual customer service allocation.

[0006] In the first aspect of the embodiments of the present invention, an outbound transfer method for realizing segmented manual transfer is provided, including: Obtain the service processing information of each manual transfer service and the current demand information of the target user within the historical time period, where the service processing information includes at least one of call duration, problem type, problem content, connected customer service, and satisfaction score; Based on the service processing information of each manual transfer service within the historical time period, determine the ability performance degree of each customer service respectively; Based on the current demand information and each ability performance degree, determine the first matching degree between each customer service and the target user respectively; Based on the first matching degree between each customer service and the target user, perform customer service allocation for the target user.

[0007] In some possible implementation manners, based on the service processing information of each manual transfer service within the historical time period, determining the ability performance degree of each customer service respectively may specifically include: Based on the service processing information of each manual transfer service within the historical time period, determine the comprehensive service processing degree of each customer service respectively; Based on the problem content and satisfaction rating of each manual transfer service within a historical time period, determine the comprehensive language comprehension of each customer service representative respectively; Using each comprehensive service processing degree and each comprehensive language comprehension degree, determine the ability performance of each customer service representative respectively.

[0008] In some possible implementation manners, based on the service processing information of each manual transfer service within a historical time period, determine the comprehensive service processing degree of each customer service representative respectively. Specifically, it may include: For each customer service representative, perform the following steps respectively: Based on the number of manual transfers of each problem type, each satisfaction rating, and each call duration, determine the complexity of each problem type respectively; Based on the call duration of each call for the target customer service representative to handle each problem type, determine the local service processing degree of the target customer service representative for each problem type respectively, and based on each satisfaction rating of the target customer service representative to handle each problem type, determine the user satisfaction performance of the target customer service representative for each problem type respectively. The target customer service representative is any customer service representative; Based on each complexity, each local service processing degree, and each user satisfaction performance, determine the proficiency performance of the target customer service representative for each problem type respectively; Perform a mean processing on each proficiency performance to obtain the comprehensive service processing degree of the target customer service representative.

[0009] In some possible implementation manners, based on the number of manual transfers of each problem type, each satisfaction rating, and each call duration, determine the complexity of each problem type respectively. Specifically, it may include: Using the number of manual transfer services of the target problem type and the total number of manual transfer services, determine the manual transfer frequency of the target problem type. The target problem type is any problem type; For the target problem type, use each satisfaction rating and the corresponding call duration to determine the first problem difficulty of each manual transfer service respectively; Perform a mean processing on each first problem difficulty to obtain the second problem difficulty of the target problem type; Using the manual transfer frequency and the second problem difficulty of the target problem type, determine the complexity of the target problem type.

[0010] In some possible implementation manners, based on the call duration of each call for the target customer service representative to handle each problem type, determine the local service processing degree of the target customer service representative for each problem type respectively. Specifically, it may include: Perform a mean processing on the call duration of each call for the target customer service representative to handle the target problem type to obtain the average processing duration of the target customer service representative for the target problem type. The target problem type is any problem type; Average the call duration of each call for each customer service to handle the target problem type, and obtain the standard processing duration of the target problem type; Use the average processing duration and the standard processing duration to determine the local business processing degree of the target customer service to handle the target problem type.

[0011] In some possible implementation manners, based on each complexity, each local business processing degree, and each user satisfaction performance degree, respectively determine the proficiency performance degree of the target customer service to handle each problem type, which may specifically include: For each problem type, respectively perform the following steps: Use the complexity of the target problem type and the user satisfaction performance degree to determine the first evaluation value of the target customer service to handle the target problem type, where the target problem type is any problem type; Use the complexity of the target problem type and the local business processing degree to determine the second evaluation value of the target customer service to handle the target problem type; Use the first evaluation value and the second evaluation value to determine the proficiency performance degree of the target customer service to handle the target problem type.

[0012] In some possible implementation manners, based on the problem content and satisfaction score of each manual transfer service within the historical time period, respectively determine the comprehensive language comprehension degree of each customer service, including: For each customer service, respectively perform the following steps: Perform semantic analysis on the problem content of each manual transfer service handled by the target customer service to obtain the clear expression degree of the requirements of each manual transfer service, where the target customer service is any customer service; Use the clear expression degree of the requirements of each manual transfer service and the satisfaction score to determine the local language comprehension degree of the target customer service for each manual transfer service; Average the local language comprehension degrees to obtain the comprehensive language comprehension degree of the target customer service.

[0013] In some possible implementation manners, based on the current request information and each ability performance degree, respectively determine the first fitness degree between each customer service and the target user, including: Based on the current request information of the target user, determine at least one current problem type corresponding to the target user; Based on the proficiency performance degree of each customer service to handle each current problem type, respectively determine the second fitness degree of each customer service for the current request information; Use each second fitness degree and each ability performance degree to respectively determine the first fitness degree between each customer service and the target user.

[0014] In some possible implementation manners, based on the first fitness degree between each customer service and the target user, perform customer service allocation for the target user, which may specifically include: Select multiple candidate customer service representatives from each customer service representative, where the first fitness degree of each candidate customer service representative is greater than a preset fitness degree threshold; Sort each candidate customer service representative in descending order of the first fitness degree to obtain the arrangement order of each candidate customer service representative; Traverse each candidate customer service representative according to the arrangement order, and determine the target customer service representative as the candidate customer service representative whose first working status belongs to the idle state; Assign the target customer service representative to the target user.

[0015] In the second aspect of the embodiments of the present invention, an outbound transfer system for implementing segmented manual transfer is provided, including: An information acquisition module, configured to acquire service processing information of each manual transfer service and current demand information of a target user within a historical time period, where the service processing information includes at least one of call duration, problem type, problem content, docking customer service representative, and satisfaction score; An ability determination module, configured to respectively determine the ability performance degree of each customer service representative based on the service processing information of each manual transfer service within a historical time period; A fitness degree determination module, configured to respectively determine the first fitness degree between each customer service representative and the target user based on the current demand information and each ability performance degree; A customer service representative assignment module, configured to assign the target user to a customer service representative based on the first fitness degree between each customer service representative and the target user.

[0016] The present invention has the following beneficial effects: In the outbound transfer method for implementing segmented manual transfer provided by the embodiments of the present invention, by acquiring the service processing information of each manual transfer service within a historical time period, including call duration, problem type, problem content, docking customer service representative, and satisfaction score, etc., it is possible to comprehensively and deeply understand the processing ability and performance of each customer service representative in different types of problems. Based on these service processing information, the ability performance degree of each customer service representative is evaluated, and then according to the current demand information of the target user and the ability performance degree of each customer service representative, the first fitness degree between each customer service representative and the target user is calculated. In this way, the present invention can more accurately assign the target user to the customer service representative most suitable for handling its current demand by accurately identifying the current demand information of the target user and quantifying the working ability of each customer service representative, thereby improving the assignment accuracy of manual customer service representatives. Description of the Drawings

[0017] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 A schematic flowchart of an outbound transfer method for implementing segmented manual transfer provided by an embodiment of the present invention; Figure 2 A schematic flowchart of S200 provided by an embodiment of the present invention; Figure 3 A schematic flowchart of S210 provided by an embodiment of the present invention; Figure 4 A schematic structural diagram of an outbound transfer system for implementing segmented manual transfer provided by an embodiment of the present invention. Detailed implementation manners

[0019] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, will detail the specific implementation manners, structures, features, and effects of an outbound transfer method and system for implementing segmented manual transfer proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present invention belongs.

[0021] It should be noted that in the technical solutions of the present invention, the acquisition, storage, use, processing, etc. of data all comply with the relevant regulations of laws and regulations.

[0022] It should be noted that in the embodiments of the present invention, some industry-existing solutions such as certain software, components, models, etc. may be mentioned. They should be regarded as exemplary. The purpose is only to illustrate the feasibility in the implementation of the technical solutions of the present invention, but it does not mean that the applicant has already or necessarily used this solution.

[0023] In many application scenarios such as today's call centers, customer service systems, and outbound sales calls, the outbound transfer mode of manual transfer has been widely used. The key advantage of this mode is that it can accurately transfer the customer's phone call or relevant information to the corresponding service process or professional customer service personnel according to the specific situation, thus bringing a more meticulous and personalized service experience to the customer.

[0024] Currently, with the help of advanced technologies such as intelligent voice assistants and automated queries, it has been possible to automate the front-end service links in user requirements. Only relatively complex and special problems will be transferred to the hands of human customer service. This approach, on the one hand, significantly reduces labor costs, and on the other hand, also significantly improves customer satisfaction. However, due to the complexity, diversity, and particularity of customer problems, combined with the differences in professional capabilities of human customer service in handling different types of problems, in actual operation, the accuracy of the allocation of human customer service is often not satisfactory.

[0025] In the face of the above problems, the present invention first analyzes the key bottlenecks existing in the traditional manual customer service allocation mechanism. The prior art relies on static rules to evaluate the capabilities of customer service, resulting in the inability to accurately match the customer demands with complex semantics to the customer service with corresponding processing capabilities. The present invention discovers that dynamically evaluating the multi-dimensional ability performance demonstrated by customer service in historical services and calculating the adaptability in combination with the characteristics of real-time user demands is the core breakthrough point for improving the allocation accuracy.

[0026] Regarding this, as Figure 1 shown, the present invention provides a schematic flow diagram of an outbound transfer method for realizing segmented manual transfer. The outbound transfer method for realizing segmented manual transfer can be applied to the server side, and the outbound transfer method for realizing segmented manual transfer may include the following S100 to S400.

[0027] S100, obtain the service processing information of each manual transfer service within a historical time period and the current demands information of the target user. The service processing information includes at least one of call duration, problem type, problem content, corresponding customer service, and satisfaction score.

[0028] In this embodiment, the service processing information refers to multi-dimensional data indicators related to the manual transfer service in the historical call record. Specifically, the call duration can be used to reflect the processing efficiency, the problem type and content can be used to reflect the business field, the corresponding customer service can be used to identify the service personnel, and the satisfaction score can be used to measure the service quality to achieve this. These indicators together constitute the basic data for evaluating the capabilities of customer service.

[0029] As an example, the server extracts the records of each manual transfer service within a historical time period from the database of the call center or customer service system. These records should contain key information such as call duration, problem type, problem content, assigned customer service representative, and satisfaction rating. Specifically, these data can be obtained through methods such as Structured Query Language (SQL) query, data export tools, or Application Programming Interface (API) interfaces.

[0030] Meanwhile, when the target user initiates a call or submits a request, the server promptly captures the current request information. The request information can include the user's problem description, expected solution, urgency level, etc. Specifically, these information can be collected through methods such as speech recognition technology, online form submission, chatbot interaction, etc.

[0031] S200, Based on the service processing information of each manual transfer service within the historical time period, determine the performance level of each customer service representative respectively.

[0032] In this embodiment, the performance level refers to the quantitative evaluation result of the comprehensive ability demonstrated by the customer service representative in historical services. Specifically, it can be calculated by analyzing dimensions such as the processing efficiency, user satisfaction, and semantic understanding ability of the customer service representative under different problem types, and is used to objectively reflect the business expertise and skill level of the customer service representative.

[0033] As an example, the server calculates the performance level of each customer service representative for each problem type based on indicators such as call duration and satisfaction rating, and then calculates the average value of the performance level for each problem type to obtain the overall performance level of each customer service representative.

[0034] For example, a comprehensive scoring formula can be set to sum the weighted values of each indicator to obtain the performance level score of each customer service representative for each problem type. Among them, the longer the call duration, the lower the corresponding processing efficiency of the customer service representative can be understood; the higher the satisfaction rating, the higher the corresponding skill level of the customer service representative can be understood. Alternatively, machine learning algorithms such as decision trees and random forests can also be used to model the service processing information to predict the performance level of each customer service representative for each problem type.

[0035] S300, Based on the current request information and each performance level, determine the first fitness degree between each customer service representative and the target user respectively.

[0036] In this embodiment, the first fitness refers to the quantitative value of the matching degree between the current demands of the target user and the capabilities of each customer service. Specifically, it can be achieved by parsing the user demands into problem types and associating and calculating them with the processing capabilities of the corresponding types of customer service. This indicator directly determines the priority of customer service allocation.

[0037] As an example, the server calculates the fitness between them based on the performance degree of each customer service and the current demand information of the target user. Among them, the fitness can be obtained through methods such as similarity calculation and matching degree scoring.

[0038] For example, if the problem type of the target user matches the problem type that a certain customer service is good at handling, and the overall performance degree of this customer service is relatively high, then the fitness between them is relatively high.

[0039] S400. Based on the first fitness between each customer service and the target user, allocate customer service for the target user.

[0040] In this embodiment, customer service allocation refers to dynamically selecting the optimal service personnel based on the fitness results calculated in real time. Specifically, a sorting and screening mechanism can be adopted to preferentially allocate the customer service with the highest fitness and in an idle state to ensure the accurate matching of user needs and customer service expertise.

[0041] As an example, the server sorts the customer services according to the calculated first fitness between each customer service and the target user. Select the customer service with the highest first fitness as the docking customer service for the target user.

[0042] Then, allocate the call or demand of the target user to the selected docking customer service. Specifically, it can be allocated through the automatic allocation of the server, manual intervention, or a combination of both.

[0043] In the outbound transfer method for realizing segmented manual transfer provided in this embodiment, by obtaining the service processing information of each manual transfer service within the historical time period, including call duration, problem type, problem content, docking customer service, and satisfaction score, etc., it is possible to comprehensively and deeply understand the processing capabilities and performances of each customer service in different types of problems. Evaluate the performance degree of each customer service based on these service processing information, and then calculate the first fitness between each customer service and the target user according to the current demand information of the target user and the performance degree of each customer service. In this way, the present invention can more accurately allocate the target user to the most suitable customer service for handling their current demands by accurately identifying the current demand information of the target user and quantifying the working capabilities of each customer service, thereby improving the allocation accuracy of manual customer service.

[0044] In some of the above solutions of this application, a method for determining the performance of customer service capabilities based on historical data is proposed. However, in actual applications, relying solely on business processing indicators in a single dimension may not comprehensively reflect the comprehensive service capabilities of customer service, resulting in insufficient accuracy in subsequent adaptability calculations, thereby affecting the final allocation effect.

[0045] In response to this, as Figure 2 shown, the present invention further proposes that S200 may specifically include the following S210 to S230: S210, based on the service processing information of each manual transfer service within a historical time period, respectively determine the comprehensive business processing degree of each customer service; S220, based on the problem content and satisfaction score of each manual transfer service within a historical time period, respectively determine the comprehensive language comprehension degree of each customer service; S230, using each comprehensive business processing degree and each comprehensive language comprehension degree, respectively determine the performance degree of each customer service.

[0046] In this embodiment, the comprehensive business processing degree reflects the overall ability and efficiency of customer service in handling various business problems. It can be comprehensively calculated based on multiple factors such as the number of problems handled by the customer service, the handling time, and the user satisfaction, and is used to measure the comprehensive performance of the customer service in business processing.

[0047] The comprehensive language comprehension degree reflects the ability of the customer service to understand the language of the user's problems. It can be based on the correlation analysis of the problem content and the satisfaction score to evaluate the customer service's ability to understand the user's problems and provide accurate answers. The satisfaction score can be an important indicator of user feedback and reflect the effect of the customer service's language comprehension.

[0048] As an example, the server first calculates multiple evaluation indicators associated with the comprehensive business processing degree according to the service processing information of each manual transfer service within a historical time period. For example, the evaluation indicators may include processing efficiency (determined based on the average call duration) and skill level (determined based on the average satisfaction score). Then, according to each evaluation indicator, weighted summation is used to obtain the comprehensive business processing degree of each customer service.

[0049] Then, the server extracts the problem content and the corresponding satisfaction score of each manual transfer service within a historical time period. Perform text analysis on the problem content to extract information such as keywords and topics to understand the types of user problems. Then analyze the correlation between the problem content and the satisfaction score to evaluate the performance of the customer service in understanding the user's problems. For example, the satisfaction scores obtained by the customer service in handling a certain type of problem can be counted. Finally, according to the results of the correlation analysis, a target algorithm (such as regression analysis, clustering analysis, etc.) is used to calculate the comprehensive language comprehension degree of each customer service.

[0050] Finally, integrate the comprehensive business processing degree and comprehensive language comprehension degree of each customer service representative to form a complete ability evaluation data set. Construct an ability performance evaluation model, which can use weighted summation, machine learning algorithms (such as decision trees, neural networks, etc.) or other forms of combination methods, take the comprehensive business processing degree and comprehensive language comprehension degree as inputs, and output the ability performance degree of each customer service representative. Alternatively, the comprehensive business processing degree and comprehensive language comprehension degree of each customer service representative can also be directly multiplied to serve as the ability performance degree corresponding to each customer service representative.

[0051] Through this embodiment, comprehensively evaluate the business processing ability and language comprehension ability of customer service representatives, so as to more accurately measure the comprehensive ability of customer service representatives. Thereby, in the subsequent customer service assignment process, more suitable customer service representatives can be more accurately matched, improving service quality and customer satisfaction.

[0052] In some of the above solutions of this application, a method for determining the comprehensive business processing degree of customer service representatives based on historical service processing information is proposed. However, due to the differences in the processing difficulties of different problem types and the varying processing efficiencies and user satisfaction performances of customer service representatives for various problems, simply relying on mean processing or a single indicator is difficult to accurately reflect the actual ability of customer service representatives to handle specific problem types, resulting in a deviation between the evaluation result of the comprehensive business processing degree and the actual business requirements.

[0053] In response to this, as Figure 3 shown, the present invention further proposes that S210 may specifically include: For each customer service representative, respectively perform the following S211 to S214: S211, based on the number of manual transfers, each satisfaction score, and each call duration for each problem type, respectively determine the complexity of each problem type; S212, based on the call duration of each time the target customer service representative handles each problem type, respectively determine the local business processing degree of the target customer service representative for each problem type, and based on each satisfaction score of the target customer service representative for handling each problem type, respectively determine the user satisfaction performance degree of the target customer service representative for each problem type, where the target customer service representative is any customer service representative; S213, based on each complexity, each local business processing degree, and each user satisfaction performance degree, respectively determine the proficiency performance degree of the target customer service representative for each problem type; S214, perform mean processing on each proficiency performance degree to obtain the comprehensive business processing degree of the target customer service representative.

[0054] In this embodiment, the complexity reflects the difficulty level of handling a certain problem type. It comprehensively considers the number of manual transfers, each satisfaction score, and call duration, and is used to measure the knowledge, skills, and effort required to solve this type of problem.

[0055] The local business processing degree reflects the business processing ability and efficiency of the target customer service when dealing with specific problem types. It can be calculated based on the call duration of each call when the target customer service handles this problem type. The shorter the call duration, the higher the processing efficiency.

[0056] The user satisfaction performance reflects the satisfaction degree of users with the service of the target customer service when dealing with specific problem types. It can be determined based on the satisfaction score of each call when the target customer service handles this problem type. The higher the satisfaction score, the higher the user satisfaction performance.

[0057] The proficiency performance combines the complexity of the problem type, the local business processing degree of the target customer service, and the user satisfaction performance to evaluate the proficiency of the target customer service when dealing with specific problem types.

[0058] As an example, the server can use weighted summation or a preset algorithm to combine the number of manual transfers, satisfaction scores, and call durations to calculate the complexity of each problem type. For example, complexity = (number of manual transfers * weight 1) + (1 - average satisfaction score * weight 2) + (average call duration * weight 3), where weight 1, weight 2, and weight 3 are coefficients determined according to the actual situation.

[0059] Then, for each problem type, calculate the average call duration for the target customer service to handle this problem. Among them, the local business processing degree can be inversely proportional to the average call duration, that is, the shorter the average call duration, the higher the local business processing degree. For each problem type, calculate the average value of the satisfaction scores for the target customer service to handle this problem. Among them, the user satisfaction performance can directly use the average value of the satisfaction scores, or perform a certain transformation (such as a linear transformation) to conform to a specific score range.

[0060] Then, a weighted summation or a preset algorithm can be used to combine the complexity of the problem type, the local business processing degree of the target customer service, and the user satisfaction performance to calculate the proficiency performance. For example, proficiency performance = (local business processing degree * weight 4) + (user satisfaction performance * weight 5) - (complexity * weight 6), where weight 4, weight 5, and weight 6 are coefficients determined according to the actual situation.

[0061] Finally, take the average value of the proficiency performance of the customer service for all problem types to obtain the comprehensive business processing degree.

[0062] Through this embodiment, it is possible to accurately evaluate the processing ability of each customer service for different problem types, thereby providing reliable data support for subsequent customer service allocation. In this way, this method considers multiple factors such as the complexity of the problem type, the processing efficiency of the customer service, and user satisfaction, making the evaluation of customer service capabilities more comprehensive and objective.

[0063] In some of the above solutions of the present application, when determining the complexity of the problem type, the occurrence frequency and the solution difficulty of the problem type cannot be accurately measured, resulting in a deviation in the evaluation of the customer service ability.

[0064] In response to this, the present invention further proposes that S211 may specifically include: Using the number of manual transfer services of the target problem type and the total number of manual transfer services, determine the manual transfer frequency of the target problem type, where the target problem type is any one of the problem types; For the target problem type, use each satisfaction score and the corresponding call duration to respectively determine the first problem difficulty of each manual transfer service; Perform a mean processing on each first problem difficulty to obtain the second problem difficulty of the target problem type; Using the manual transfer frequency of the target problem type and the second problem difficulty, determine the complexity of the target problem type.

[0065] In this embodiment, the manual transfer frequency is calculated by the ratio of the number of transfers of the target problem type to the total number of services; since the higher the user satisfaction score and the shorter the call duration, the lower the complexity of this type of problem, the ratio of the single call duration to the satisfaction score can be used as the first problem difficulty; the second problem difficulty is obtained by calculating the arithmetic mean or weighted mean of all the first problem difficulties; the complexity is obtained by multiplying the manual transfer frequency by the second problem difficulty or by weighted summation. For example, using the product form highlights the impact of high-frequency and high-difficulty problems.

[0066] Specifically, the manual transfer frequency reflects the occurrence probability of the target problem type, and the higher the value, the more common the problem. The first problem difficulty quantifies the problem-solving difficulty through the processing duration of a single service and the user feedback. The longer the call duration and the lower the satisfaction, the higher the difficulty. The second problem difficulty integrates historical data to eliminate the contingency of a single service and reflects the average solution difficulty of this problem. The final complexity combines the frequency and the average difficulty, taking into account both the frequency of the problem occurrence and the solution difficulty, providing more comprehensive basic data for the subsequent evaluation of the customer service ability. For example, if the manual transfer frequency of a certain problem type is 30% and the second problem difficulty is 0.8, the calculated result of the complexity is 0.24, which can accurately represent the comprehensive processing challenge degree of this problem type.

[0067] As an example, the server first uses the number of manual transfer services of the target problem type and the total number of manual transfer services to determine the manual transfer frequency of the target problem type. For example, assume that the target problem type is "bill query", the number of manual transfer services of this type is 100 times, and the total number of manual transfer services is 1000 times, then the manual transfer frequency of "bill query" is 0.1.

[0068] Then, for the target problem type, using each satisfaction score and the corresponding call duration, determine the first problem difficulty of each manual transfer service respectively. Specifically, the ratio of the single call duration to the satisfaction score can be used as the first problem difficulty. For example, if the satisfaction score of a certain "bill inquiry" service is 10 points (out of 10 full marks) and the call duration is 8 minutes, then the first problem difficulty of this service can be calculated as 0.8.

[0069] Then, perform mean processing on each first problem difficulty to obtain the second problem difficulty of the target problem type. Continuing with the "bill inquiry" as an example, assuming there are 100 services of this type, and the average value of the 100 calculated first problem difficulties is 0.7, then the second problem difficulty of "bill inquiry" is 0.7.

[0070] Finally, using the manual transfer frequency of the target problem type and the second problem difficulty, determine the complexity of the target problem type. The weighted average of the manual transfer frequency and the second problem difficulty can be used as the complexity. For example, if the manual transfer frequency of "bill inquiry" is 0.1 and the second problem difficulty is 0.7, then its complexity can be calculated as (0.1 * 0.3 + 0.7 * 0.7) = 0.52.

[0071] Through this embodiment, the present application can accurately evaluate the complexity of different problem types. By comprehensively considering the manual transfer frequency and the problem difficulty, the complexity of the problem type can be more comprehensively reflected. This method not only considers the frequency of problem occurrence but also the difficulty of problem solving, thus providing a more reliable basis for subsequent customer service allocation. Thereby, the professional ability of the customer service can be better matched with the user's problem type, improving service efficiency and user satisfaction.

[0072] In some of the above solutions of the present application, there is a lack of an objective and unified measurement standard when determining the local business processing degree. Relying solely on the processing duration of a single customer service is difficult to accurately reflect the difference in its business processing ability relative to the overall service level, resulting in a deviation in the ability assessment.

[0073] In response to this, the present invention further proposes that S212 may specifically include: Perform mean processing on the call duration of each call of the target customer service for the target problem type to obtain the average processing duration of the target customer service for the target problem type, where the target problem type is any one of the problem types; Perform mean processing on the call duration of each call of each customer service for the target problem type to obtain the standard processing duration of the target problem type; Using the average processing duration and the standard processing duration, determine the local business processing degree of the target customer service for the target problem type.

[0074] In this embodiment, the average processing time is calculated by counting the historical call duration data of the target customer service for handling specific types of problems, reflecting the average processing efficiency of the customer service for this type of problem. The standard processing time is calculated by summarizing the call duration data of all customer service for handling the same type of problem and taking the average value to establish the benchmark processing efficiency for this type of problem. The local business processing degree is generated by comparing the standard processing time with the average processing time. The shorter the average processing time is than the standard processing time, the better the processing efficiency of the target customer service is than the overall level.

[0075] Specifically, for the target problem type, the server automatically extracts the historical call records of the target customer service, and calculates the average processing time after eliminating abnormal data. At the same time, the call time of all customer service staff handling this type of problem is summarized, and the standard processing time is generated after data cleaning. The two are input into the preset calculation model, for example, the standard processing time is divided by the average processing time to get the ratio. When the ratio is greater than 1, it indicates that the processing efficiency of the target customer service is better than the average level, thereby generating the local business processing degree. For example, if the standard processing time is 300 seconds and the average processing time of the target customer service is 250 seconds, the local business processing degree is 1.2, reflecting its efficient processing ability for this type of problem. This method eliminates subjective bias in individual evaluation through objective data comparison to ensure the accuracy and fairness of capability evaluation.

[0076] As an example, the server performs average processing on the duration of each call of the target customer service to handle the target problem type, and obtains the average processing time of the target customer service for the target problem type. For example, assuming that the target customer service handles 5 calls of a certain target problem type for 10 minutes, 12 minutes, 8 minutes, 15 minutes and 11 minutes respectively, the average processing time is calculated to be 11.2 minutes.

[0077] Then, the average duration of each call of each customer service handling the target problem type is processed to obtain the standard processing time of the target problem type. Specifically, the call duration of all customer service staff handling the target problem type can be counted and the average value can be calculated. Assuming that there are 10 customer service staff handling the target problem type, their average processing time is 11.2 minutes, 10.5 minutes, 12.3 minutes, 9.8 minutes, 11.7 minutes, 10.9 minutes, 11.5 minutes, 12.1 minutes, 10.2 minutes and 11.8 minutes respectively, then the standard processing time is calculated to be 11.2 minutes.

[0078] Finally, divide the standard processing time by the average processing time to determine the local business processing degree of the target customer service handling the target problem type. Therefore, assuming that the average processing time of the target customer service is 11.2 minutes and the standard processing time is 11.2 minutes, the local business processing degree is calculated to be 1.

[0079] Furthermore, for the user satisfaction performance of the target customer service in handling each problem type, the same method as above is used for determination. That is, first, the average value of each satisfaction score of the target customer service in handling the target problem type is calculated to obtain the average satisfaction score of the target customer service for the target problem type; then, the average value of each satisfaction score of each customer service in handling the target problem type is calculated to obtain the standard satisfaction score of the target problem type; finally, the average satisfaction score is divided by the standard satisfaction score to obtain the user satisfaction performance of the target customer service in handling the target problem type.

[0080] Through this embodiment, the efficiency of the target customer service in handling specific problem types can be accurately evaluated. By comparing the average processing time of the target customer service with the standard processing time, the business ability of the target customer service in handling this problem type can be objectively reflected. This evaluation method takes into account the complexity of different problem types and avoids the one-sidedness of simply using the call duration as the evaluation criterion. At the same time, by calculating the local business processing degree, the relative performance of the target customer service in handling specific problem types can be intuitively displayed, providing reliable data support for subsequent customer service allocation.

[0081] In some of the above solutions of this application, it is proposed to determine the proficiency performance of the customer service in handling problem types through complexity, local business processing degree, and user satisfaction performance. However, in the specific implementation process, if the proficiency performance is calculated only through a single dimension, it may lead to one-sided evaluation results and cannot accurately reflect the comprehensive ability of the customer service in handling specific problem types, thus affecting the accuracy of subsequent adaptability calculation.

[0082] In response to this, the present invention further proposes that S213 may specifically include: For each problem type, the following steps are respectively executed: Using the complexity and user satisfaction performance of the target problem type, determine the first evaluation value of the target customer service in handling the target problem type, where the target problem type is any one of the problem types; Using the complexity and local business processing degree of the target problem type, determine the second evaluation value of the target customer service in handling the target problem type; Using the first evaluation value and the second evaluation value, determine the proficiency performance of the target customer service in handling the target problem type.

[0083] In this embodiment, the proficiency performance of the target customer service in handling the target problem type can be specifically determined by the following formula 1: Formula 1 In formula 1, is used to represent the proficiency performance of the m-th customer service in handling the j-th problem type, For characterizing the complexity of the j-th type of problem, For characterizing the user satisfaction performance of the m-th customer service representative when handling the j-th type of problem, For characterizing the local business processing degree of the m-th customer service representative when handling the j-th type of problem.

[0084] Among them, when the difficulty of manual question answering is higher (i.e., the complexity of the problem type is higher), more attention is paid to the user satisfaction performance of the manual customer service when handling this type of problem (measured by customer satisfaction to highlight personal ability); while when the difficulty of manual question answering is lower (i.e., the complexity of the problem type is lower), more attention is paid to the ability performance of the manual customer service when handling this type of problem.

[0085] Through this embodiment, it is possible to comprehensively consider the complexity of the customer service handling problems, the user satisfaction performance, and the local business processing degree, so as to more accurately evaluate the proficiency of the customer service in handling different types of problems. Thus, the accuracy of customer service allocation can be improved, enabling the customer's problems to be assigned to the most suitable customer service representative, thereby enhancing the quality and satisfaction of customer service.

[0086] In some of the above solutions of the present application, a comprehensive service processing degree and a comprehensive language comprehension degree are proposed to determine the customer service ability performance. However, in this process, only structured data such as call duration and problem type are relied on to evaluate the customer service ability, ignoring the accuracy of the customer service's understanding of the user's demand content, resulting in a deviation in the ability evaluation and affecting the accuracy of subsequent customer service allocation.

[0087] In response to this, the present application further proposes that S220 may specifically include: For each customer service representative, the following steps are respectively executed: Perform semantic analysis on the problem content of each manual transfer service handled by the target customer service representative to obtain the clear expression degree of the demand for each manual transfer service, where the target customer service representative is any customer service representative; Use the clear expression degree of the demand for each manual transfer service and the satisfaction score to determine the local language comprehension degree of the target customer service representative for each manual transfer service; Perform a mean processing on each local language comprehension degree to obtain the comprehensive language comprehension degree of the target customer service representative.

[0088] In this embodiment, semantic analysis extracts keywords and intentions in the problem content through a natural language processing model, and calculates the clear expression degree of the demand in combination with the context relevance; the ratio of the satisfaction score to the clear expression degree of the demand is the local language comprehension degree. Among them, when the clear expression degree of the demand is lower, but the satisfaction score is higher, the local language comprehension degree of the customer service representative is higher; the mean processing uses the arithmetic average method to integrate multiple local language comprehension degrees to eliminate the accidental error of single data and ensure the stability of the comprehensive language comprehension degree.

[0089] Specifically, for each piece of problem content processed by the target customer service, word segmentation, entity recognition, and intent classification are performed through a pre-trained language model to generate a multi-dimensional feature vector containing keyword density, semantic coherence, and intent matching degree. After normalization, the clarity of demand expression is obtained. The satisfaction score is compared with the clarity of demand expression and used as the local language understanding degree. Finally, the average value of the local language understanding degrees corresponding to all problems processed by the target customer service is taken to obtain the comprehensive language understanding degree. This process quantifies the correlation between the accuracy of the customer service's understanding of user needs and user feedback, compensates for the deficiency of simply relying on the evaluation ability of structured data, makes the calculation of the ability performance more comprehensive, and thus improves the matching accuracy of customer service allocation.

[0090] As an example, the server first performs semantic analysis on each piece of problem content of the artificial transfer service processed by the target customer service to obtain the clarity of demand expression for each artificial transfer service. Semantic analysis can be achieved through natural language processing techniques. For example, a word vector model and deep learning algorithms are used to analyze the problem content, extract keywords and semantic features, and calculate the clarity and integrity of the problem description, so as to obtain a quantitative index of the clarity of demand expression.

[0091] Then, the local language understanding degree of the target customer service for each artificial transfer service is determined by using the clarity of demand expression and the satisfaction score of each artificial transfer service. Specifically, the satisfaction score can be compared with the clarity of demand expression to calculate the local language understanding degree.

[0092] Finally, the mean value processing is performed on each local language understanding degree to obtain the comprehensive language understanding degree of the target customer service. Methods such as arithmetic mean or geometric mean can be used to calculate all local language understanding degrees, and finally a comprehensive scoring index is obtained.

[0093] Through this embodiment, the language understanding ability of the customer service can be objectively evaluated, providing a basis for subsequent customer service allocation. This method comprehensively considers the clarity of customer demand expression and customer satisfaction, and can more comprehensively reflect the language understanding level of the customer service. By statistically analyzing the data of multiple services, the performance of the customer service's language understanding ability in long-term work can be obtained, which helps to identify customer services with excellent language understanding ability performance.

[0094] In some of the above solutions of the present application, during the artificial customer service allocation process, the adaptation degree is calculated only based on the comprehensive ability performance degree, without combining the specific problem type corresponding to the current user's appeal for targeted matching, resulting in a deviation between the customer service allocation result and the actual demand.

[0095] In response to this, the present application further proposes that S300 specifically may include: Based on the current demand information of the target user, determine at least one current problem type corresponding to the target user; Based on the proficiency performance of each customer service in handling each current problem type, respectively determine the second adaptation degree of each customer service to the current demand information; Utilize each second adaptation degree and each ability performance degree to respectively determine the first adaptation degree between each customer service and the target user.

[0096] In this embodiment, the determination of the current problem type is achieved by extracting keywords and classifying the demand text through natural language processing technology. For example, a support vector machine model is used to map the user description to preset classification labels such as account management, technical failure, after-sales service, etc. The calculation of the second adaptation degree is realized by weighted summation. Specifically, the proficiency performance obtained by the customer service in the historical tasks of handling each current problem type is multiplied by the weight coefficient corresponding to the problem type and then accumulated. The weight coefficient is dynamically adjusted according to the occurrence frequency of the keywords related to each problem type in the current demand. The generation of the first adaptation degree adopts a linear combination method, and the second adaptation degree and the ability performance degree are fused according to a preset ratio, where the proportion of the ability performance degree in the combination is 40%-60%.

[0097] Specifically, when the target user submits a demand containing keywords such as "account login failed" and "password reset" through voice or text, it is classified as an account management problem. At the same time, when the current demand information of the target user may also contain demands with keywords such as "return goods" and "exchange goods", it can also be classified as an after-sales service problem. Subsequently, screen the proficiency performance of all customer services in handling account management problems and after-sales service problems. Then, the second adaptation degree of each customer service for the current demand information is obtained by weighted calculation after standardizing the proficiency performance of account management problems and after-sales service problems. The weight coefficient is dynamically adjusted according to the occurrence frequency of keywords such as "login", "password", "return goods", and "exchange goods" in the user demand. Finally, the second adaptation degree and the comprehensive ability performance degree of the customer service are fused in a 6:4 ratio to generate the first adaptation degree, ensuring that both the comprehensive service ability of the customer service and its processing advantage in specific problem areas are considered during the allocation process. For example, if the second adaptation degree of a certain customer service for the current demand information is 92 points and its ability performance degree is 85 points, then the final first adaptation degree is calculated as 92×0.6 + 85×0.4 = 89.2 points. This dual-dimensional evaluation mechanism effectively avoids the misallocation of customer services with excellent comprehensive abilities but lacking experience in specific fields, and improves the service matching accuracy in complex problem scenarios.

[0098] As an example, the server extracts keywords from the speech request text of the target user through natural language processing algorithms, and identifies two current problem types: account anomaly and tariff dispute. The server retrieves the proficiency performance of all customer service representatives in handling account anomaly problems and tariff dispute problems from the historical database. The weighted average of the proficiency performance of 0.87 for handling account anomaly problems and 0.76 for handling tariff dispute problems by the customer service representatives is calculated, where the weights are dynamically allocated according to the occurrence frequency of the problem types in the request information, and the second fitness index of 0.82 is obtained. Further, this second fitness index is multiplied by the comprehensive ability performance of 0.91 of Customer Service A to generate the final first fitness value of 0.75.

[0099] Through this embodiment, an accurate customer service matching mechanism based on multi-dimensional ability portraits is realized, effectively solving the problem of misaligned type matching in the manual allocation process. By dynamically associating the user request types with the customer service special ability data, customer service representatives in non-expert fields are actively avoided in the allocation process, enabling user problems to reach customer service representatives with corresponding processing experience directly, significantly reducing the transfer times and improving the first problem resolution rate.

[0100] In some of the above solutions of this application, although the target user is allocated a customer service representative based on the first fitness between each customer service representative and the target user, it does not consider whether each customer service representative is currently available. As a result, after the target user is allocated a customer service representative, the target user needs to wait for the customer service representative, resulting in a poor user experience.

[0101] In response to this, this application further proposes that S400 may specifically include: Select multiple candidate customer service representatives from all customer service representatives whose first fitness is greater than a preset fitness threshold; Sort the candidate customer service representatives in descending order of the first fitness to obtain the ranking order of each candidate customer service representative; Traverse each candidate customer service representative in the ranking order, and determine the candidate customer service representative whose first working status belongs to the idle state as the target customer service representative; Allocate the target customer service representative to the target user.

[0102] In this embodiment, the preset fitness threshold is a preset numerical standard for screening customer service representatives. Only customer service representatives whose first fitness is greater than this threshold will be recognized as candidate customer service representatives meeting the basic fitness requirements and enter the subsequent screening and allocation process. This threshold is determined according to factors such as actual business requirements, system resources, and expectations for customer service quality.

[0103] Candidate customer service representatives are multiple representatives selected from all customer service representatives based on the condition that the first fitness degree is greater than a preset fitness degree threshold. These customer service representatives have the basic conditions to match with the target user and have the potential to be assigned to serve the target user.

[0104] The arrangement order is the position order of each candidate customer service representative in the sorting result after sorting the selected candidate customer service representatives in descending order of the first fitness degree. The higher the fitness degree of a customer service representative, the more forward the position in the arrangement order.

[0105] The working status is used to describe whether the customer service representative can currently accept new user requests and provide services. It usually includes "idle status" and "busy status", etc. "Idle status" means that the customer service representative is not currently dealing with other users' problems and can immediately respond to new users; "Busy status" means that the customer service representative is dealing with other users' problems and temporarily cannot accept new user requests.

[0106] The target customer service representative is the first candidate customer service representative in the idle status during the process of traversing each candidate customer service representative in the arrangement order, that is, the customer service representative determined to be finally assigned to serve the target user.

[0107] As an example, the server compares the first fitness degree of each customer service representative calculated with the preset fitness degree threshold, and selects the customer service representatives whose first fitness degree is greater than the threshold. These customer service representatives constitute the candidate customer service representative set.

[0108] Then, a common sorting algorithm (such as bubble sort, quick sort, merge sort, etc.) is used to sort the candidate customer service representative set. Using the first fitness degree as the sorting key, the candidate customer service representatives are arranged in descending order. For example, in the quick sort algorithm, the first fitness degree of a candidate customer service representative is selected as the benchmark value, and the first fitness degrees of other candidate customer service representatives are compared with the benchmark value. Those greater than the benchmark value are placed on the left, and those less than the benchmark value are placed on the right. Then, the left and right parts are sorted recursively. Finally, a list of candidate customer service representatives arranged in descending order of the first fitness degree is obtained, and the position of each customer service representative in the list is the arrangement order.

[0109] Then, starting from the candidate customer service representative with the most forward arrangement order, the working status of each candidate customer service representative is checked in turn. The current working status of the customer service representative can be determined by querying the customer service status database. During the traversal process, once it is found that the working status of a certain candidate customer service representative is the idle status, the traversal is immediately stopped, and this customer service representative is determined as the target customer service representative, and the subsequent candidate customer service representatives are not checked anymore.

[0110] Finally, associate the request information of the target user with the target customer service, and send a notification (such as popping up a prompt box, sound reminder, etc.) to the terminal device where the target customer service is located, informing the target customer service that there is a new user request to be processed. At the same time, feedback the information that the customer service has been successfully assigned to the target user, so that the target user can establish a connection with the target customer service and communicate.

[0111] Through this embodiment, the candidate customer services are sorted according to the first fitness degree, and the customer service that is idle and has a high fitness degree is preferentially selected as the target customer service, avoiding the waste of resources caused by blindly assigning customer services. It not only makes full use of the capabilities of customer services with high fitness degrees, but also reasonably utilizes the resources of idle customer services, improving the utilization efficiency of customer service resources and reducing the service cost.

[0112] An outbound transfer method for implementing segmented manual transfer. Correspondingly, the present invention also provides a specific embodiment of an outbound transfer system for implementing segmented manual transfer.

[0113] As Figure 4 shown, a structural schematic diagram of an outbound transfer system for implementing segmented manual transfer is provided. The outbound transfer system 400 for implementing segmented manual transfer may include an information acquisition module 410, a capability determination module 420, a fitness degree determination module 430, and a customer service assignment module 440.

[0114] The information acquisition module 410 is configured to acquire the service processing information of each manual transfer service and the current demand information of the target user within a historical time period, and the service processing information includes at least one of call duration, problem type, problem content, docking customer service, and satisfaction score; The capability determination module 420 is configured to respectively determine the capability performance degree of each customer service based on the service processing information of each manual transfer service within a historical time period; The fitness degree determination module 430 is configured to respectively determine the first fitness degree between each customer service and the target user based on the current demand information and each capability performance degree; The customer service assignment module 440 is configured to assign customer services to the target user based on the first fitness degree between each customer service and the target user.

[0115] In the outbound transfer system for implementing segmented manual transfer provided in this embodiment, by obtaining the service processing information of each manual transfer service within a historical time period, including call duration, problem type, problem content, assigned customer service representative, and satisfaction score, etc., it is possible to comprehensively and deeply understand the processing capabilities and performance of each customer service representative in different types of problems. Based on this service processing information, the performance degree of each customer service representative is evaluated, and then according to the current demand information of the target user and the performance degree of each customer service representative, the first fitness between each customer service representative and the target user is calculated. In this way, by accurately identifying the current demand information of the target user and quantifying the work ability of each customer service representative, the present invention can more precisely assign the target user to the most suitable customer service representative for handling their current demands, improving the accuracy of manual customer service assignment.

[0116] It should be clear that the present invention is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, the detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present invention is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order between steps after understanding the spirit of the present invention.

[0117] It should also be noted that the exemplary embodiments mentioned in the present invention describe some methods or systems based on a series of steps or devices. However, the present invention is not limited to the order of the above steps, that is, the steps can be executed in the order mentioned in the embodiments, or different from the order in the embodiments, or several steps can be executed simultaneously.

[0118] As mentioned above, the above is only the specific implementation manner of the present invention. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, modules, and units can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated here. It should be understood that the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of various equivalent modifications or substitutions, and these modifications or substitutions should all be covered within the protection scope of the present invention.

Claims

1. An outbound transfer method for implementing segmented manual transfer, characterized in that, The method includes: Obtaining the service processing information of each manual transfer service and the current demand information of the target user within a historical time period, where the service processing information includes at least one of call duration, problem type, problem content, assigned customer service, and satisfaction score; Based on the service processing information of each manual transfer service within the historical time period, respectively determining the ability performance of each customer service; Based on the current demand information and each ability performance, respectively determining the first fitness between each customer service and the target user; Based on the first fitness between each customer service and the target user, performing customer service assignment for the target user; The step of respectively determining the ability performance of each customer service based on the service processing information of each manual transfer service within the historical time period includes: Based on the service processing information of each manual transfer service within the historical time period, respectively determining the comprehensive service processing degree of each customer service; Based on the problem content and the satisfaction score of each manual transfer service within the historical time period, respectively determining the comprehensive language comprehension degree of each customer service; Using each comprehensive service processing degree and each comprehensive language comprehension degree, respectively determining the ability performance of each customer service; The step of respectively determining the first fitness between each customer service and the target user based on the current demand information and each ability performance includes: Based on the current demand information of the target user, determining at least one current problem type corresponding to the target user; Based on the proficient performance of each customer service in handling each current problem type, respectively determining the second fitness of each customer service for the current demand information; Using each second fitness and each ability performance, respectively determining the first fitness between each customer service and the target user.

2. The outbound transfer method for realizing segmented manual transfer according to claim 1, characterized in that The step of respectively determining the comprehensive service processing degree of each customer service based on the service processing information of each manual transfer service within the historical time period includes: For each customer service, respectively performing the following steps: Based on the number of manual transfers of each problem type, each satisfaction score, and each call duration, respectively determining the complexity of each problem type; Based on the call duration of each time the target customer service handles each problem type, respectively determining the local service processing degree of the target customer service for each problem type, and based on the satisfaction score of each time the target customer service handles each problem type, respectively determining the user satisfaction performance of the target customer service for each problem type, where the target customer service is any one of the customer services; Based on each complexity, each local service processing degree, and each user satisfaction performance, respectively determining the proficient performance of the target customer service in handling each problem type; Performing an average processing on each proficient performance to obtain the comprehensive service processing degree of the target customer service.

3. The outbound transfer method for implementing segmented manual transfer according to claim 2, wherein Based on the number of manual transfers for each of the said problem types, each of the said satisfaction scores, and each of the said call durations, respectively determine the complexity of each of the said problem types, including: Using the number of manual transfer services for the target problem type and the total number of the said manual transfer services, determine the manual transfer frequency of the target problem type, where the target problem type is any one of the said problem types; For the target problem type, respectively determine the first problem difficulty degree of each of the said manual transfer services by using each of the said satisfaction scores and the corresponding said call duration; Perform a mean processing on each of the said first problem difficulty degrees to obtain the second problem difficulty degree of the target problem type; Using the said manual transfer frequency and the said second problem difficulty degree of the target problem type, determine the said complexity of the target problem type.

4. The outbound transfer method for implementing segmented manual transfer according to claim 2, characterized in that Based on each of the said call durations for the target customer service to handle each of the said problem types, respectively determine the local business handling degree of the target customer service to handle each of the said problem types, including: Perform a mean processing on each of the said call durations for the target customer service to handle the target problem type to obtain the average processing duration of the target customer service for the target problem type, where the target problem type is any one of the said problem types; Perform a mean processing on each of the said call durations for each customer service to handle the target problem type to obtain the standard processing duration of the target problem type; Using the said average processing duration and the said standard processing duration, determine the said local business handling degree of the target customer service to handle the target problem type.

5. The outbound transfer method for implementing segmented manual transfer according to claim 2, characterized in that, Based on each of the said complexities, each of the said local business handling degrees, and each of the user satisfaction performance degrees, respectively determine the proficiency performance degree of the target customer service to handle each of the said problem types, including: For each of the said problem types, respectively perform the following steps: Using the said complexity and the user satisfaction performance degree of the target problem type, determine the first evaluation value of the target customer service to handle the target problem type, where the target problem type is any one of the said problem types; Using the said complexity and the said local business handling degree of the target problem type, determine the second evaluation value of the target customer service to handle the target problem type; Using the said first evaluation value and the said second evaluation value, determine the said proficiency performance degree of the target customer service to handle the target problem type.

6. The outbound transfer method for implementing segmented manual transfer according to claim 1, wherein Based on the said problem content and the said satisfaction score of each of the said manual transfer services within the historical time period, respectively determine the comprehensive language understanding degree of each of the said customer services, including: For each of the said customer services, respectively perform the following steps: Perform a semantic analysis on the said problem content of each of the said manual transfer services handled by the target customer service to obtain the clear expression degree of the requirements of each of the said manual transfer services, where the target customer service is any one of the said customer services; Using the said clear expression degree of the requirements of each of the said manual transfer services and the said satisfaction score, determine the local language understanding degree of the target customer service for each of the said manual transfer services; Perform a mean processing on each of the said local language understanding degrees to obtain the said comprehensive language understanding degree of the target customer service.

7. The outbound transfer method for implementing segmented manual transfer according to any one of claims 1-6, characterized in that, Performing customer service allocation for the target user based on the first fitness degree between each customer service and the target user includes: Screening out multiple candidate customer services with the first fitness degree greater than a preset fitness degree threshold from each of the customer services; Sorting each of the candidate customer services in descending order of the first fitness degree to obtain the arrangement order of each of the candidate customer services; Traversing each of the candidate customer services according to the arrangement order, and determining the target customer service as the candidate customer service whose first working state belongs to the idle state; Allocating the target customer service to the target user.

8. An outbound transfer system for implementing segmented manual transfer, characterized in that, The system includes: An information acquisition module, configured to acquire service processing information of each manual transfer service and current demand information of the target user within a historical time period, where the service processing information includes at least one of call duration, problem type, problem content, docking customer service, and satisfaction score; A capability determination module, configured to respectively determine the capability performance degree of each customer service based on the service processing information of each manual transfer service within the historical time period; The respectively determining the capability performance degree of each customer service based on the service processing information of each manual transfer service within the historical time period includes: Respectively determining the comprehensive service processing degree of each customer service based on the service processing information of each manual transfer service within the historical time period; Respectively determining the comprehensive language comprehension degree of each customer service based on the problem content and the satisfaction score of each manual transfer service within the historical time period; Using the comprehensive service processing degrees and the comprehensive language comprehension degrees to respectively determine the capability performance degree of each customer service; A fitness degree determination module, configured to respectively determine the first fitness degree between each customer service and the target user based on the current demand information and the capability performance degrees; The respectively determining the first fitness degree between each customer service and the target user based on the current demand information and the capability performance degrees includes: Based on the current demand information of the target user, determining at least one current problem type corresponding to the target user; Respectively determining the second fitness degree of each customer service for the current demand information based on the proficiency performance degree of each customer service in processing each current problem type; Using the second fitness degrees and the capability performance degrees to respectively determine the first fitness degree between each customer service and the target user; A customer service allocation module, configured to perform customer service allocation for the target user based on the first fitness degree between each customer service and the target user.

Images