CN111932144B

CN111932144B - Customer service agent distribution method and device, server and storage medium

Info

Publication number: CN111932144B
Application number: CN202010864488.8A
Authority: CN
Inventors: 殷腾飞
Original assignee: Tencent Technology Shanghai Co Ltd
Current assignee: Tencent Technology Shanghai Co Ltd
Priority date: 2020-08-25
Filing date: 2020-08-25
Publication date: 2023-09-19
Anticipated expiration: 2040-08-25
Also published as: CN111932144A

Abstract

The application provides a customer service agent distribution method, a customer service agent distribution device, a server and a storage medium, wherein the customer service agent distribution method, the server and the storage medium are used for receiving consultation problems sent by customers and determining text information representing the consultation problems; the text information is classified based on the pre-trained customer service agent allocation model to determine a candidate customer service mode for providing consultation service for customers, so that the problem of inaccurate customer service agent allocation caused by dependence on the corresponding relation between keywords and customer service agents in the prior art is avoided; and under the condition that the candidate customer service agents transfer customers to the target customer service agents, the customer service agent distribution model is automatically optimized based on the target customer service agents and the consultation problem, so that the accuracy of customer service agent distribution based on the customer service agent distribution model in the follow-up process is improved, accurate recommendation of the customer service agents is realized, and the convenience of post maintenance of the customer service agent distribution method is improved.

Description

Customer service agent distribution method and device, server and storage medium

Technical Field

The application relates to the technical field of artificial intelligence, in particular to a customer service agent distribution method, a customer service agent distribution device, a server and a storage medium.

Background

In instant messaging (Instant Messenger, abbreviated as IM) products, customer service agent allocation is a critical ring in a pre-sale marketing/after-sale service scenario, and the accuracy of customer service agent allocation directly affects the response efficiency of customer service agents to customer consultation problems.

The traditional customer service agent distribution mode is to distribute according to busy states of the customer service agents so as to balance the number of customers received by each customer service agent, and only the number of the customers received by the customer service agents is concerned in the mode, and the consultation problem of the customer service agents in good treatment is not considered, so that the situation that the customer service agents cannot solve the consultation problem often exists, and the accurate degree of the customer service agent distribution is low.

The existing customer service agent distribution mode is to preset the corresponding relation between keywords and customer service agents, acquire the keywords contained in the consultation questions sent by the clients, and distribute the customer service agents corresponding to the keywords in the consultation questions to the clients. The mode can exert the characteristics and different service familiarity of each customer service agent to a certain extent, but depends on the setting of the corresponding relation between the keywords and the customer service agents, and the distribution of the customer service agents is relatively dead, so that the later maintenance is inconvenient; in addition, if the problem of the customer service agent is changed and the corresponding relation between the keywords and the customer service agent is not updated in time, inaccurate distribution of the customer service agent is also caused; and the same meaning has a plurality of different description modes, and the keywords are difficult to be fully configured under the common condition, so that the precision degree of customer service agent allocation is also influenced.

Disclosure of Invention

In view of the above, the present application provides a customer service agent distribution method, device, server and storage medium, so as to improve the accuracy degree of customer service agent distribution and the convenience of post maintenance. The technical proposal is as follows:

a customer service agent distribution method comprises the following steps:

receiving a consultation problem sent by a client, and determining text information representing the consultation problem;

classifying the text information according to a pre-trained customer service agent distribution model to determine candidate customer service agents for providing consultation services for the customers in at least one customer service agent;

pushing the consultation questions to the candidate customer service agents;

responding to the feedback information of the candidate customer service agents to the consultation questions to determine target customer service agents for providing consultation services for the customers;

if the target customer service agent and the candidate customer service agent are not the same customer service agent, optimizing the customer service agent distribution model based on the consultation problem and the training sample constructed by the target customer service agent.

A customer service agent distribution device comprising:

the consultation problem receiving unit is used for receiving consultation problems sent by clients and determining text information representing the consultation problems;

The candidate customer service agent determining unit is used for classifying the text information according to a pre-trained customer service agent distribution model to determine candidate customer service agents for providing consultation services for the customers in at least one customer service agent;

the consultation problem pushing unit is used for pushing the consultation problem to the candidate customer service seat;

the target customer service agent determining unit is used for determining a target customer service agent for providing consultation service for the customer in response to the feedback information of the candidate customer service agent on the consultation problem;

and the optimizing unit is used for optimizing the customer service agent distribution model based on the consultation problem and the training sample constructed by the target customer service agent if the target customer service agent and the candidate customer service agent are not the same customer service agent.

A server, comprising: at least one memory and at least one processor; the memory stores a program, and the processor calls the program stored in the memory, wherein the program is used for realizing the customer service agent distribution method.

A computer-readable storage medium having stored therein computer-executable instructions for performing the customer service agent allocation method.

The embodiment of the application provides a customer service agent distribution method, a device, a server and a storage medium, which are used for receiving consultation problems sent by customers, determining text information representing the consultation problems, classifying the text information based on a pre-trained customer service agent distribution model to determine a candidate customer service mode for providing consultation services for the customers, and avoiding the problem of inaccurate customer service agent distribution caused by dependence on the corresponding relation between keywords and the customer service agents in the prior art; and under the condition that the candidate customer service agents transfer customers to the target customer service agents, the customer service agent distribution model is automatically optimized based on the target customer service agents and the consultation problem, so that the accuracy of customer service agent distribution based on the customer service agent distribution model in the follow-up process is improved, and the convenience of post maintenance of the customer service agent distribution method provided by the embodiment of the application is improved.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only embodiments of the present application, and that other drawings can be obtained according to the provided drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic diagram of a customer service agent distribution system according to an embodiment of the present application;

FIG. 2 is a schematic diagram of another customer service agent distribution system according to an embodiment of the present application;

FIG. 3 is a schematic diagram of a customer service agent distribution system according to an embodiment of the present application;

fig. 4 is a flowchart of a method for generating a customer service agent allocation model according to an embodiment of the present application;

fig. 5 is a schematic diagram of a customer service agent allocation model generation method according to an embodiment of the present application;

FIG. 6 is a schematic diagram of a customer service intelligent distribution process according to an embodiment of the present application;

FIG. 7 is a schematic diagram of a method for automatically correcting a customer service agent distribution model according to an embodiment of the present application;

fig. 8 is a flowchart of a customer service agent allocation method provided in an embodiment of the present application;

fig. 9 is a schematic diagram of a customer service agent allocation method according to an embodiment of the present application;

fig. 10 is a schematic diagram of another customer service agent allocation method according to an embodiment of the present application;

fig. 11 is a schematic structural diagram of a customer service agent distribution device according to an embodiment of the present application;

fig. 12 is a block diagram of a hardware structure of a server adapted to a customer service agent allocation method according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Examples:

in instant messaging IM customer service products, customer service distribution logic is a critical ring in pre-sales marketing/after-sales service scenarios, and the accuracy of customer service distribution directly affects the resolution efficiency of the final problem and the rate of sales.

Currently, the common customer service distribution systems in the industry include the following:

the earliest customer service distribution system is realized based on fixed distribution, and the fixed distribution mode can be as follows: the last portal in the instant messaging application corresponds to a fixed customer service seat, and the customer service is directly distributed no matter what customers come from. For example, referring to FIG. 1, regardless of which client clicks on the online customer service button of customer service agent A on the instant messaging application, customer service agent A is assigned to serve that client; no matter which client clicks the online customer service button of the customer service agent B on the instant messaging application, the customer service agent B is allocated to serve the client.

Conventional customer service agent distribution systems distribute, in connection with fig. 2, customer service agents according to their on-line status/reception status, when one customer service agent is not on-line or is servicing a plurality of customers, it may not be possible to normally receive the customer service agent to distribute new customers to him, and thus, customer service agents which are on-line and are not currently busy may be preferentially distributed.

The existing customer service agent allocation system allocates, according to the specified keyword, referring to fig. 3, by specifying a specific keyword to be allocated to a specific customer service agent, the keyword can be set according to the service that each customer service agent is good at, so that the customer is allocated to one customer service agent that is good at.

The customer service agent distribution system shown in fig. 1-2 only focuses on the problem that the reception quantity does not consider the service category or the adequacy of each customer service agent, the distribution accuracy is not ideal enough, and the distributed customer service agents cannot be guaranteed to solve the consultation problem of customers.

In the embodiment of the application, the accuracy of customer service agent allocation can be regarded as the accuracy of customer service agent allocation, and the better the customer service agent allocated to a customer is, the higher the accuracy of customer service agent allocation is; on the contrary, the less good the customer service agents allocated for the customer service are at handling the consultation problem raised by the customer, the lower the accuracy of the allocation of the customer service agents is.

The customer service agent distribution system shown in fig. 3 solves the problem that the distribution accuracy is not ideal to a certain extent through a keyword distribution mode, but keywords are difficult to set completely, particularly Chinese is profound, the expression mode is very rich, and the distribution accuracy is difficult to be realized through simple keyword matching. In addition, the customer service agent distribution system shown in fig. 3 relies on the setting of the corresponding relation between the keywords and the customer service agents, so that the distribution strategy is relatively dead, not flexible enough, one customer distributes a certain customer service agent, even if the customer service agent is distributed incorrectly, the customer service agent is distributed still in the next time, and automatic optimization cannot be realized through self-learning.

Artificial intelligence (Artificial Intelligence, AI) is the theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use the knowledge to obtain optimal results. In other words, artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. Artificial intelligence, i.e. research on design principles and implementation methods of various intelligent machines, enables the machines to have functions of sensing, reasoning and decision.

The artificial intelligence technology is a comprehensive subject, and relates to the technology with wide fields, namely the technology with a hardware level and the technology with a software level. Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning and other directions.

The embodiment of the application provides a method for carrying out Natural Language Processing (NLP) on consultation problems sent by clients, and a customer service seat which is most good at processing the consultation problems is found through an NLP classification model and is distributed. And when inaccurate customer service seat allocation occurs, the NLP classification model can be corrected according to the actual treatment person of the consultation problem (the customer service seat actually allocated to the customer), and the learning is retrained, so that the allocation accuracy is improved.

Natural language processing (Nature Language processing, NLP) is an important direction in the fields of computer science and artificial intelligence. It is studying various theories and methods that enable effective communication between a person and a computer in natural language. Natural language processing is a science that integrates linguistics, computer science, and mathematics. Thus, the research in this field will involve natural language, i.e. language that people use daily, so it has a close relationship with the research in linguistics. Natural language processing techniques typically include text processing, semantic understanding, machine translation, robotic questions and answers, knowledge graph techniques, and the like.

Furthermore, the customer service agent distribution method provided by the embodiment of the application can also adopt a voice processing technology, and when the consultation problem input by the customer is voice, the voice can be recognized by the voice processing technology to obtain text information of the consultation problem.

Key technologies to the speech technology (Speech Technology) are automatic speech recognition technology (ASR) and speech synthesis technology (TTS) and voiceprint recognition technology. The method can enable the computer to listen, watch, say and feel, is the development direction of human-computer interaction in the future, and voice becomes one of the best human-computer interaction modes in the future.

With research and advancement of artificial intelligence technology, research and application of artificial intelligence technology is being developed in various fields, such as common smart home, smart wearable devices, virtual assistants, smart speakers, smart marketing, unmanned, automatic driving, unmanned aerial vehicles, robots, smart medical treatment, smart customer service, etc., and it is believed that with the development of technology, artificial intelligence technology will be applied in more fields and with increasing importance value.

The scheme provided by the embodiment of the application is applied to the field of intelligent customer service, and is particularly used for realizing automatic distribution of customer service agents. The scheme provided by the embodiment of the application relates to artificial intelligence voice processing technology, natural language processing technology and other technologies, and is specifically described by the following embodiments:

The customer service agent distribution method provided by the embodiment of the application is applied to an instant communication system, the instant communication system is composed of a terminal and a server, wherein the terminal provides instant communication application, and the server provides technical support for the instant communication application on the terminal. The instant messaging application provided by the terminal is provided with a problem consultation entrance (such as an online customer service button), after a customer clicks the problem consultation entrance, a consultation window is displayed on the terminal, the customer can edit the consultation problem which the customer wants to input and input the consultation problem into the consultation window, the terminal transmits the consultation problem input into the consultation window to the server, and the server distributes customer service agents for the customer according to the consultation inquiry problem.

In the embodiment of the application, a pre-trained NLP classification model (also called as a customer service agent distribution model) is arranged in a server, and after receiving a consultation problem sent by a customer, the server converts the consultation problem into text information and inputs the text information into the customer service agent distribution model to obtain a customer service agent for distribution to the customer.

In order to provide good consultation service for clients, a plurality of customer service accounts can be set, and after the customer service logs in the instant messaging application through the customer service accounts, the consultation service can be provided for the clients based on the customer service accounts. Accordingly, a client can log in the instant messaging application through the client account, the consultation problem is sent through the client account, the server carries out classified calculation on the consultation problem sent by the client based on the customer service distribution model to determine a customer service account providing consultation service for the client, communication connection between the client account and the determined customer service account is established, information interaction between the client currently logged in the client account and the customer service of the customer service account determined by current login is achieved, and the purpose that the customer service provides consultation service for the client is achieved.

The customer service agent distribution method provided by the embodiment of the application relates to a customer service agent distribution model training process, a customer service agent intelligent distribution process and a customer service agent distribution automatic correction process.

The customer service agent allocation model training process is mainly used for initializing training sample data required by obtaining an NLP classification model through first training, wherein the training sample data can be the problem of previous customer consultation and the customer service agent corresponding to the problem to be solved, the training sample data comprises at least one training sample, and the NLP classification model to be trained is trained based on the training sample to obtain the NLP classification model used in the subsequent actual allocation process.

The intelligent customer service seat distribution process is used for an actual customer service seat distribution process, and after a customer is accessed, text information of consultation problems sent to the customer is input into an NLP classification model to obtain a customer service seat which is required to be distributed for the customer.

The automatic correction process of the customer service agent distribution model is self-learning logic in the customer service agent distribution method provided by the embodiment of the application, when the customer service agent feels that distributed customers are not right, customer transfer can be carried out, customers are distributed to other customer service agents, correspondingly, the automatic correction process can automatically correct the NLP classification model by using the new customer service agent after transfer and the consultation problem sent when the user accesses, re-learn is carried out to obtain an updated NLP classification model, and the subsequent customer service agent intelligent distribution process can use the updated NLP classification model to ensure that accurate customer service agents are distributed to the customers who send the same consultation problem subsequently.

The customer service agent allocation model training process is mainly used for generating an NLP classification model, namely, a customer service agent allocation model is generated, and a specific generation mode of the customer service agent allocation model is shown in FIG. 4.

As shown in fig. 4, the method includes:

s401, acquiring a training sample set formed by a plurality of training samples, wherein the training samples indicate historical consultation problems sent by a client and customer service agents actually distributed to the client according to the historical consultation problems;

the customer service agent distribution model training process can be regarded as a process of initializing and training the customer service agent distribution model, the process is only needed to be executed once in the whole life cycle of the customer service agent distribution model, the customer service agent distribution model is generated through the customer service agent distribution model training process, the customer service agent can be distributed for customers through the customer service agent distribution model, and further, the customer service agent distribution model generated in the customer service agent distribution model training process can be optimized through an automatic correction process of the customer service agent distribution model.

The training samples of the customer service agent allocation model training process may be questions from previous customer consultations (may also be referred to as historical consultation questions) and customer service agents that actually solve the historical consultation questions, such as the following formats:

Consultation questions	Customer service seat
		What is the play object asking me to buy did not receive it?	Customer service seat A
I want to complain that I charge does not go to account-!	Customer service seat A
		I report someone using the plug-in-!	Customer service seat B
Your game today's activities appear to be problematic.	Customer service seat C
		……	……

According to the embodiment of the application, the customer service account number comprises an account name and a password, one customer service account number can be regarded as one customer service seat, and different customer service account numbers can be regarded as different customer service seats.

The account names of different customer service accounts are different, the customer service seat can be represented by the account name of the customer service account, and the account name can be the customer service work number. In this case, the training sample may consist of consultation questions and customer service numbers.

S402, updating the historical consultation problems in the training sample into text information representing the historical consultation problems to obtain a target training sample;

referring to fig. 5, all training samples in the collected training sample set may be input to an NLP classification algorithm (e.g., textCNN, fastText, etc.) for training to generate a customer service agent allocation model. The NLP classification algorithm may be considered as an NLP classification model to be trained, i.e., a customer service agent allocation model to be trained.

It should be noted that, the consultation questions input by the clients may be text, voice, text and voice. Therefore, taking a training sample as an example, before the training sample is input into the customer service agent allocation model to be trained, the embodiment of the application can also determine the text information for representing the historical consultation problem in the training sample, further update the historical consultation problem in the training sample into the text information for representing the historical consultation problem to obtain a target training sample, and input the target training sample into the customer service agent allocation model to be trained.

S403, training the customer service seat distribution model to be trained based on all target training samples to obtain the customer service seat distribution model.

According to the embodiment of the application, the target training sample is input into the customer service agent distribution model to be trained, the result of classifying the text information in the target training sample by the customer service agent distribution model to be trained approaches to the customer service agent indicated by the target training sample as the training target, and the customer service agent distribution model to be trained is trained to obtain the customer service agent distribution model.

And (3) aiming at each training sample in the training sample set, a corresponding target training sample can be obtained, all obtained target training samples are input into the customer service seat distribution model to be trained, and the customer service seat distribution model to be trained is trained based on all target training samples so as to obtain the customer service seat distribution model.

In the embodiment of the application, the customer service agent distribution model consists of a feature extraction model and a classification model, wherein the feature extraction model is used for extracting feature information of text information in a target training sample, and the feature information can be a feature vector; the feature extraction model inputs the feature information of the text information in the extracted target training sample to the classification model, and the classification model classifies the consultation problems in the target training sample characterized by the feature information to determine a customer service agent (candidate customer service agent) for providing consultation services for the customer who sends the consultation problems.

Referring to fig. 6, a schematic diagram of an intelligent distribution process of customer service agents is provided in an embodiment of the present application. The intelligent customer service agent allocation process is used for an actual customer service agent allocation process, text information representing the consultation problem is automatically determined after a customer accesses (i.e. after the customer sends the consultation problem), then the text information is input into the generated customer service agent allocation model (i.e. NLP classification model), which customer service agent the consultation problem should be allocated to is calculated by combining with the trained customer service agent allocation model to be processed (for convenience of distinguishing, the customer service agent determined by the customer service agent allocation model is called as a candidate customer service agent), and finally the accessed customer is allocated to the corresponding customer service agent.

Referring to fig. 7, a schematic diagram of a method for automatically correcting a customer service agent allocation model is shown, after a customer allocates a candidate customer service agent, if the candidate customer service agent feels that the allocation is incorrect, the customer cannot process a consultation problem sent by the customer, and the customer can be manually transferred to other customer service agents capable of processing the consultation problem. At this time, the self-learning function of the customer service agent distribution method is triggered, the customer service agent distribution method can re-input the converted new customer service agent and the consultation problem when the customer is accessed into the NLP classification model for correction, the new NLP classification model is re-trained, the new NLP classification model (i.e. the optimized NLP classification model) is used for subsequent customer distribution, and therefore the purpose of self-learning can be achieved, and re-mistakes are avoided.

The following describes a customer service agent distribution method provided by the embodiment of the present application in detail from an actual customer service agent distribution process and an automatic correction customer service agent distribution model process, and specifically please refer to fig. 8.

As shown in fig. 8, the method includes:

s801, receiving a consultation problem sent by a client, and determining text information representing the consultation problem;

in the embodiment of the application, a client can open an instant messaging application on a terminal, register a user aiming at the instant messaging application to obtain a registered account, wherein the registered account comprises an account name and a password (for convenience of distinguishing, the registered account is called a client account), the client logs in the instant messaging application on the terminal through the client account, a consultation entrance (such as an online customer service button) is arranged on the instant messaging application, the client clicks the consultation entrance, the instant messaging application displays a consultation interface, the terminal can send the consultation problem to a server after the client inputs the consultation problem in the consultation interface, and the server can determine text information representing the consultation problem after receiving the consultation problem.

The consultation questions input by the clients can be text, voice or both text and voice, so that the server needs to determine text information representing the consultation questions after receiving the consultation questions sent by the clients through the terminals.

Taking a consultation problem as an example, the manner of determining text information representing the consultation problem may be: detecting whether text content exists in the consultation problem, and if so, extracting the text content in the consultation problem (for convenience of distinguishing, the text content in the consultation problem is called as first text content); detecting whether voice information exists in the consultation problem, and if so, identifying the voice information in the consultation problem to obtain text content for representing the voice information (for convenience of distinguishing, the text content for representing the voice information is called second text content); the first text content and the second text content are spliced into text information that characterizes the consultation problem.

Further, if text content is not present but voice information is present in the counseling problem, the second text content may be determined as text information characterizing the counseling problem; if text content is present in the counseling problem but no speech information is present, the first text content may be determined as text information characterizing the counseling problem.

For example, the consultation question input by the client is composed of a section of text content and a section of voice information, and in the consultation question, the text content in the consultation question is located before the voice information, and can be extracted and determined to be a first text content, the voice information in the consultation question is subjected to voice recognition to obtain a second text content representing the voice information, the first text content and the second text content are spliced to form text information representing the consultation question, and the first text content is located before the second text content in the text information.

The above provides a preferred manner of determining text information characterizing a consultation problem for an embodiment of the present application, and the manner is equally applicable to a process of determining text information characterizing a historical consultation problem. That is, in the embodiment of the present application, the manner of determining the text information characterizing the historical consultation problem is the same as the manner of determining the text information characterizing the consultation problem, so that the accuracy of the agent allocation model can be ensured.

S802, classifying the text information according to a pre-trained customer service agent distribution model to determine candidate customer service agents for providing consultation services for customers in at least one customer service agent;

In the embodiment of the application, after receiving the consultation problem sent by the client through the terminal and determining the text information representing the consultation problem, the server inputs the text information into the pre-trained customer service seat distribution model, and the customer service seat distribution model classifies the consultation problem according to the text information to obtain the customer service seat for providing the consultation service for the client (for convenience of distinguishing, the customer service seat is a candidate customer service seat).

S803, pushing consultation questions to the candidate customer service agents;

in the embodiment of the application, after receiving the consultation problem sent by the client through the terminal and determining the candidate customer service seat for providing the consultation service for the client according to the customer service seat distribution model, the server can push the consultation information to the candidate customer service seat, taking the customer service seat as a customer service account number as an example, and after pushing the consultation information to the customer service account number of the candidate customer service seat, the customer service currently using the customer service account number can return the feedback information of the consultation problem to the server according to whether the customer service is good at handling the consultation problem.

As a preferred implementation manner of the embodiment of the application, if the candidate customer service agent directly solves the consultation problem to the customer after receiving the consultation information, the feedback information characterization of the candidate customer service agent on the consultation problem can be determined to be good at processing.

As another preferred implementation of the embodiment of the present application, if the candidate customer service agent clicks the tamper button from the tamper button and the tamper button after receiving the consultation information, it may be determined that the feedback information of the candidate customer service agent to the consultation problem characterizes the tamper.

As a preferred implementation manner of the embodiment of the present application, if the candidate customer service agent transfers the consultation problem to other customer service agents after receiving the consultation information, it may be determined that the candidate customer service agent is not good at handling the consultation problem, and the transferred customer service agent may be regarded as a target customer service agent good at handling the consultation problem.

As another preferred implementation of the embodiment of the present application, if the candidate customer service agent clicks the disagreement processing button from the disagreement processing button and the disagreement processing button after receiving the consultation information, it may be determined that the candidate customer service agent characterizes the disagreement processing of the feedback information of the consultation problem, and further, the candidate customer service agent may be required to transfer the consultation problem to the target customer service agent that is good at processing the consultation problem.

S804, determining a target customer service agent for providing the customer with the consultation service in response to the feedback information of the candidate customer service agent on the consultation problem;

In the embodiment of the application, the feedback information of the candidate customer service agents on the consultation problems represents good treatment/poor treatment, and when the feedback information of the candidate customer service agents on the consultation problems represents poor treatment, the candidate customer service agents are not good at treating the consultation problems; and when the feedback information of the candidate customer service agents on the consultation questions indicates that the candidate customer service agents are good at handling the consultation questions, indicating that the candidate customer service agents are good at handling the consultation questions.

The mode of transferring the consultation problem to the target customer service agent by the candidate customer service agent can be as follows: the server pushes a currently online customer service agent list in at least one customer service agent to the candidate customer service agents, the candidate customer service agents manually select target customer service agents which are good at handling the consultation problem from the customer service agent list, and the consultation problem is transferred to the target customer service agents.

S805, detecting whether the target customer service agent and the candidate customer service agent are the same customer service agent; if the target customer service agent and the candidate customer service agent are not the same customer service agent, executing step S806;

in the embodiment of the application, if the feedback information of the candidate customer service agents on the consultation problem represents good treatment, the candidate customer service agents are used as the target customer service agents which are finally determined and used for providing consultation services for customers; if the feedback information representation of the candidate customer service agent on the consultation problem is not good at processing, the candidate customer service agent is used as the final determined target customer service agent for providing the consultation service for the customer to the other customer service agent to which the consultation problem is transferred.

S806, optimizing a customer service agent distribution model based on the training sample constructed by the consultation problem and the target customer service agent.

According to the embodiment of the application, if the target customer service agent and the candidate customer service agent are not the same customer service agent, the fact that the customer service agent distribution result of the consultation problem sent by the customer by the current customer service agent distribution model is inaccurate is indicated, at the moment, a training sample is required to be constructed according to the consultation problem and the target customer service agent, and the current customer service agent distribution model is optimally trained based on the training sample to obtain the optimized customer service agent distribution model, so that the customer service agent distribution accuracy of the customer service agent distribution model is further improved.

Furthermore, in the customer service agent distribution method provided by the embodiment of the application, if the target customer service agent and the candidate customer service agent are the same customer service agent, the customer service agent distribution model is not required to be optimized according to the consultation problem sent by the customer service and the target customer service agent.

In the embodiment of the application, the customer service agent distribution model is composed of a feature extraction model and a classification model, wherein the feature extraction model is used for extracting feature information of text information representing the consultation problem, and the classification model is used for classifying the consultation problem according to the feature information to obtain candidate customer service agents for processing the consultation problem. Specifically, after receiving a consultation problem sent by a client, determining text information representing the consultation problem, inputting the text information into a feature extraction model, extracting feature information of the text information by the feature extraction model, inputting the extracted feature information into a classification model, and classifying the consultation problem by the classification model according to the feature information to obtain a customer service seat (candidate customer service seat) for providing consultation service for the client sending the consultation problem.

The text information is classified according to a pre-trained customer service agent allocation model, and the determination of the candidate customer service agent for providing the customer with the consultation service in at least one customer service agent can be: the method comprises the steps of inputting text information representing consultation problems input by customers into a customer service agent distribution model, classifying the consultation problems according to the text information by the customer service agent distribution model, and outputting matching information of each customer service agent in at least one customer service agent relative to the consultation problems, namely outputting one matching information for each customer service agent in the at least one customer service agent. The matching information characterizes the extent of tampering; and selecting one customer service agent from the at least one customer service agents as a candidate customer service agent by utilizing the matching information of each customer service agent in the at least one customer service agent relative to the consultation problem.

Specifically, text information representing a consultation problem input by customer service is input into a feature extraction model, feature information of the text information is extracted by the feature extraction model, the feature information is input into a classification model, the classification model classifies the consultation problem according to the feature information to obtain matching information of each customer service seat in at least one customer service seat relative to the consultation problem, and then matching information of each customer service seat in the at least one customer service seat relative to the consultation problem is utilized to select one customer service seat from the at least one customer service seat as a candidate customer service seat.

In the embodiment of the application, at least one customer service agent is all customer service agents indicated by training samples for training the candidate customer service agents.

As a preferred implementation manner of the embodiment of the present application, matching information of each customer service agent in at least one customer service agent with respect to the consultation problem may be used to select a customer service agent with the highest treatment degree on the consultation problem from at least one customer service agent, as a candidate customer service agent for providing a consultation service for a customer who sends the consultation problem.

As another implementation manner of the embodiment of the application, the idle state information of each customer service agent in at least one customer service agent can be obtained, and one customer service agent is selected from the at least one customer service agents to serve as a candidate customer service agent for providing consultation service for a customer sending the consultation problem by combining the idle state information of each customer service agent in the at least one customer service agent and the matching information of each customer service agent relative to the consultation problem.

The idle state information of the customer service agents characterizes the customer reception quantity of the customer service agents. The method comprises the steps of firstly determining each customer service agent with the treatment degree of the relative consultation problem meeting the preset condition from at least one customer service agent, selecting one customer service agent with the highest treatment degree of the relative consultation problem from all the determined unselected customer service agents, judging whether the customer reception quantity of the current selected customer service agent reaches the preset upper limit value, and determining the current selected customer service agent as a candidate customer service agent if the customer reception quantity of the current selected customer service agent does not reach the preset upper limit value; if the customer reception number of the currently selected customer service agents reaches a preset upper limit value, returning to execute a process of selecting a customer service agent with highest treatment degree relative to the consultation problem from the unselected customer service agents in all the determined customer service agents until the candidate customer service agents are selected.

According to the embodiment of the application, if the treatment degree of the customer service agent relative to the consultation problem reaches the preset standard degree, the customer service agent is determined to meet the preset condition; if the treatment degree of the customer service seat relative to the consultation problem does not reach the preset standard degree, determining that the customer service seat does not meet the preset condition.

If the customer reception number of each customer service agent meeting the preset condition determined from the at least one customer service agent reaches the preset upper limit value, each customer service agent meeting the preset condition determined from the at least one customer service agent can be monitored, and once the customer service agent with the customer reception number not reaching the preset upper limit value is monitored, the monitored customer service agent is determined to be a candidate customer service agent.

And distributing the clients sending the consultation problems to the candidate customer service agents, if the candidate customer service agents are not good at handling the consultation problems, pushing a customer service agent list to the candidate customer service agents, wherein the customer service agent list indicates the currently online customer service agents except the candidate customer service agents in at least one customer service agent, and the candidate customer service agents can select one customer service agent good at handling the consultation problems from the customer service agent list as a target customer service agent and transfer the clients to the target customer service agents.

Further, the customer service agent list provided by the embodiment of the application not only indicates all the currently online customer service agents except the candidate customer service agents in at least one customer service agent, but also each customer service agent indicated by the customer service agent list carries a recommendation index, and for one customer service agent, the recommendation index of the customer service agent is related to the matching information of the customer service agent relative consultation problem and the idle state information of the current customer service agent.

The recommendation index of the customer service agent is positively correlated with the treatment degree of the tampering degree of the matching information characterization of the relative consultation problem of the customer service agent; the better the customer service agent is at handling the consultation problem, the higher the recommendation index of the customer service agent.

The recommendation index of the customer service agents is inversely related to the customer reception quantity represented by the idle state information of the current customer service agents; the more customers of the current customer service agents receive, the lower the recommendation index of the customer service agents.

The customer service agent distribution method provided by the embodiment of the application not only can distribute the customer service agents based on the treatment degree of the customer service agents on consultation problems, but also can be combined with other distribution logics, for example, the customer service agents are already receiving a plurality of customers, and when reaching the upper limit of reception, the current customers can be distributed to other not very busy customer service agents by combining with other distribution logics.

The foregoing is merely a preferred manner of other allocation logic provided by the embodiments of the present application, and the other allocation logic is not limited to the specific manner related to the customer reception amount, and the inventor can set the other allocation logic according to his own needs, which is not limited herein.

The customer service agent distribution method provided by the embodiment of the present application will be described in detail with reference to a schematic diagram of the customer service agent distribution method provided by the embodiment of the present application, and in particular, fig. 9 is shown.

As shown in fig. 9, after a client accesses to send a self-consultation problem, the method for distributing customer service agents provided by the embodiment of the present application automatically determines text information representing the self-consultation problem sent by the client, and performs NLP processing on the text information, converts the text information into numerical vectors (feature vectors) that can be operated by a machine learning classification algorithm (i.e., a classification model), and then uses a classification model that has been trained in advance to conduct classification to obtain a customer service agent distribution result (candidate customer service agents).

Fig. 10 is a schematic diagram of a customer service agent allocation method according to an embodiment of the present application. As shown in fig. 10, after the user accesses, if the service agent allocation result (candidate service agent) of the current classification model is inaccurate, the candidate service agent manually transfers the customer to another service agent (target service agent) that is good at handling. According to the customer service agent distribution method provided by the embodiment of the application, when the switching operation is detected, the NLP classification model is automatically corrected by the correct target customer service agent after switching, and retraining is performed. After the new NLP classification model is provided, the user consults similar problems and is distributed to the correct customer service agents.

According to the customer service agent distribution method provided by the embodiment of the application, the NLP classification algorithm can select textCNN, fastText text vector conversion and classification integrated algorithm, and can also select a text feature vector extraction algorithm (for example doc2 vec) together with a machine learning multi-class classification algorithm (for example a decision tree) to realize the process in two steps, and in this case, the text feature extraction algorithm and the classification algorithm are used by training a model. In addition, the scheme can be further matched with other distribution schemes to achieve a more ideal effect.

According to the customer service agent distribution method provided by the embodiment of the application, the consultation problems sent by the customers are intelligently identified and classified through the NLP technology, so that the problems of dead plates and difficult setting of complete keyword matching are avoided, the method is more intelligent and more accurate than the conventional keyword matching, the customers can be more accurately distributed to the customer service agents capable of solving other problems, the customer service agent efficiency is higher, the customer consultation problem solving efficiency is higher, and the identification hit rate and the accuracy of the consultation problems sent by the customers are greatly improved.

In addition, when the customer service agent finds that the accessed customer is not a customer service agent which can be processed by the customer service agent, the customer service agent can be transferred to the customer service agent which can be processed really. By utilizing the self-learning capability of the AI, when the customer service agent is transferred to the customer, automatic correction logic is triggered, the NLP classification model is retrained, and the distribution accuracy is continuously and self-improved, so that the customer service agent distribution method provided by the embodiment of the application is more intelligent.

Fig. 11 is a schematic diagram of a customer service agent distribution device according to an embodiment of the present application.

As shown in fig. 11, the apparatus includes:

a consultation problem receiving unit 111, configured to receive a consultation problem sent by a client, and determine text information characterizing the consultation problem;

a candidate customer service agent determining unit 112, configured to perform classification processing on the text information according to a pre-trained customer service agent allocation model, to determine a candidate customer service agent for providing a customer with a consultation service in at least one customer service agent;

a consultation problem pushing unit 113, configured to push consultation problems to the candidate customer service agents;

a target customer service agent determining unit 114, configured to determine a target customer service agent that provides a customer with a consultation service in response to feedback information of the candidate customer service agent on the consultation problem;

and an optimizing unit 115, configured to optimize the customer service agent allocation model based on the training sample constructed by the consultation problem and the target customer service agent if the target customer service agent and the candidate customer service agent are not the same customer service agent.

In the embodiment of the present application, preferably, the consultation problem receiving unit includes:

the consultation problem receiving subunit is used for receiving consultation problems sent by clients;

a text extraction unit for extracting a first text content in the consultation problem;

The voice recognition unit is used for recognizing the voice information in the consultation problem to obtain second text content used for representing the voice information;

and the text information generation unit is used for splicing the first text content and the second text content into text information which is characterized by the consultation problem.

In the embodiment of the present application, preferably, the customer service agent allocation model is composed of a feature extraction model and a classification model, and the candidate customer service agent determination unit includes:

the feature extraction unit is used for inputting the text information into the feature extraction model to extract the feature information of the text information;

the classification unit is used for classifying the characteristic information by using the classification model to determine candidate customer service agents for providing consultation services for the customers in the at least one customer service agent.

In the embodiment of the present application, preferably, the target customer service agent determining unit includes:

the feedback information receiving unit is used for receiving feedback information of the candidate customer service agents on the consultation problem;

the judging unit is used for judging whether the feedback information characterizes the consultation problem to be treated;

a first determining unit for determining a target customer service agent for providing a customer with a consultation service, excluding the candidate customer service agent, from at least one customer service agent if the feedback information indicates that the consultation problem is not good;

And the second determining unit is used for determining the candidate customer service agents as target customer service agents if the feedback information characterizes that the consultation problem is good at being processed.

In an embodiment of the present application, preferably, the classification unit includes:

the matching information generating unit is used for inputting the characteristic information into the classification model and outputting matching information of each customer service agent in at least one customer service agent relative to the consultation problem, and the matching information characterizes the good treatment degree of the consultation problem;

the state information acquisition unit is used for acquiring idle state information of each customer service agent in the current at least one customer service agent;

and a third determining unit configured to determine, from the at least one customer service agent, a candidate customer service agent for providing a consultation service to the customer based on the idle state information of each of the at least one customer service agent and the matching information of each of the customer service agents with respect to the consultation problem.

In an embodiment of the present application, preferably, the first determining unit includes:

a customer service agent list determining unit, configured to determine a customer service agent list, where the customer service agent list indicates each customer service agent currently on-line except for the candidate customer service agent in at least one customer service agent;

the customer service agent list pushing unit is used for pushing the customer service agent list to the candidate customer service agents;

And the selecting unit is used for responding to the selecting operation of the candidate customer service agents on each customer service agent indicated by the customer service agent list and determining the customer service agent selected by the candidate customer service agent as the target customer service agent.

In the embodiment of the application, preferably, each customer service agent indicated by the customer service agent list carries a recommendation index, and the recommendation index of the customer service agent is related to idle state information of the current customer service agent and matching information of the relative consultation problem of the customer service agent.

Cloud technology (Cloud technology) refers to unifying serial resources such as hardware, software, network and the like in a wide area network or a local area network to realize calculation of data,(Storage)Processing andsharingIs a kind of (a) and (b) of the productEscrow tubeTechniques.

Cloud technology (Cloud technology) is based on the general terms of network technology, information technology, integration technology, management platform technology, application technology and the like applied by Cloud computing business models, and can form a resource pool, so that the Cloud computing business model is flexible and convenient as required. Cloud computing technology will become an important support. Background services of technical networking systems require a large amount of computing, storage resources, such as video websites, picture-like websites, and more portals. Along with the high development and application of the internet industry, each article possibly has an own identification mark in the future, the identification mark needs to be transmitted to a background system for logic processing, data with different levels can be processed separately, and various industry data needs strong system rear shield support and can be realized only through cloud computing.

Cloud computing (clouding) is a computing model that distributes computing tasks across a large pool of computers, enabling various application systems to acquire computing power, storage space, and information services as needed. The network that provides the resources is referred to as the "cloud". Resources in the cloud are infinitely expandable in the sense of users, and can be acquired at any time, used as needed, expanded at any time and paid for use as needed.

As a basic capability provider of cloud computing, a cloud computing resource pool (cloud platform for short, generally referred to as IaaS (Infrastructure as a Service, infrastructure as a service) platform) is established, in which multiple types of virtual resources are deployed for external clients to select for use.

According to the logic function division, a PaaS (Platform as a Service ) layer can be deployed on an IaaS (Infrastructure as a Service ) layer, and a SaaS (Software as a Service, software as a service) layer can be deployed above the PaaS layer, or the SaaS can be directly deployed on the IaaS. PaaS is a platform on which software runs, such as a database, web container, etc. SaaS is a wide variety of business software such as web portals, sms mass senders, etc. Generally, saaS and PaaS are upper layers relative to IaaS.

Cloud computing (closed computing) refers to the delivery and usage mode of an IT infrastructure, meaning that required resources are obtained in an on-demand, easily scalable manner through a network; generalized cloud computing refers to the delivery and usage patterns of services, meaning that the required services are obtained in an on-demand, easily scalable manner over a network. Such services may be IT, software, internet related, or other services. Cloud Computing is a product of fusion of traditional computer and network technology developments such as Grid Computing (Grid Computing), distributed Computing (distributed Computing), parallel Computing (Parallel Computing), utility Computing (Utility Computing), network storage (Network Storage Technologies), virtualization (Virtualization), load balancing (Load balancing), and the like.

With the development of the internet, real-time data flow and diversification of connected devices, and the promotion of demands of search services, social networks, mobile commerce, open collaboration and the like, cloud computing is rapidly developed. Unlike the previous parallel distributed computing, the generation of cloud computing will promote the revolutionary transformation of the whole internet mode and enterprise management mode in concept.

The cloud call center ((Cloud Call Center)) is a call center system built based on a cloud computing technology, enterprises can rapidly own the call center without purchasing any software and hardware systems and only by having basic conditions such as personnel, places and the like, and software and hardware platforms, communication resources, daily maintenance and services are provided by server providers. The method has the characteristics of short construction period, less investment, low risk, flexible deployment, strong system capacity scalability, low operation and maintenance cost and the like; whether a telephone marketing center or a customer service center, an enterprise can establish a call center system which has comprehensive, stable and reliable functions and can distribute the call access nationwide all over the country only by renting the service as required.

The instant communication system is composed of a terminal and a server. The terminal may be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, etc. The server can be service equipment for providing service for the user at the network side, can be an independent physical server, can be a server cluster or a distributed system formed by a plurality of physical servers, and can also be a cloud server for providing cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, basic cloud computing service such as big data and artificial intelligent platform. The terminal and the server may be directly or indirectly connected through wired or wireless communication, and the present application is not limited herein.

In the embodiment of the application, the server provides the customer service agent distribution method based on the cloud call center, and in order to facilitate understanding, the customer service agent distribution method provided by the embodiment of the application is described in detail from the perspective of the server. Fig. 12 is a block diagram of a hardware structure of a server according to an embodiment of the present application. Referring to fig. 12, the hardware structure of the server may include: processor 121, communication interface 122, memory 123, and communication bus 124;

In the embodiment of the present invention, the number of the processor 121, the communication interface 122, the memory 123 and the communication bus 124 may be at least one, and the processor 121, the communication interface 122 and the memory 123 complete communication with each other through the communication bus 124;

processor 121 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of the present invention, etc.;

memory 123 may comprise high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), etc., such as at least one disk memory;

wherein the memory stores a program, and the processor is operable to invoke the program stored in the memory, the program being operable to:

classifying the text information according to a pre-trained customer service agent distribution model to determine candidate customer service agents for providing consultation services for customers in at least one customer service agent;

pushing consultation questions to candidate customer service agents;

responding the feedback information of the candidate customer service agents to the consultation problems and determining a target customer service agent for providing consultation services for the customers;

If the target customer service agent and the candidate customer service agent are not the same customer service agent, optimizing a customer service agent distribution model based on a training sample constructed by the consultation problem and the target customer service agent.

Alternatively, the refinement function and the extension function of the program may be described with reference to the above.

Still further, an embodiment of the present application further provides a computer readable storage medium, where computer executable instructions are stored, where the computer executable instructions are configured to execute the above customer service agent allocation method.

Alternatively, the refinement and expansion functions of the computer-executable instructions may be described with reference to the foregoing.

The above describes in detail a customer service agent distribution method, device, server and storage medium provided by the present invention, and specific examples are applied to illustrate the principles and embodiments of the present invention, and the above description of the examples is only used to help understand the method and core ideas of the present invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.

It should be noted that, in the present specification, each embodiment is described in a progressive manner, and each embodiment is mainly described as different from other embodiments, and identical and similar parts between the embodiments are all enough to be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant points refer to the description of the method section.

It is further noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include, or is intended to include, elements inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. The customer service agent distribution method is characterized by comprising the following steps of:

pushing the consultation questions to the candidate customer service agents;

receiving feedback information of the candidate customer service agents on the consultation questions;

if the feedback information indicates that the consultation problem is not good at being processed, determining a customer service agent list, wherein the customer service agent list indicates all the customer service agents which are on line currently except the candidate customer service agent in the at least one customer service agent;

Pushing the customer service agent list to the candidate customer service agents;

responding to the selection operation of the candidate customer service agents on each customer service agent indicated by the customer service agent list, and determining the customer service agent selected by the candidate customer service agents as a target customer service agent; the target customer service agent is another customer service agent which is good at processing the consultation problem and to which the candidate customer service agent transfers the consultation problem;

if the target customer service agent and the candidate customer service agent are not the same customer service agent, responding to the switching operation of the candidate customer service agent, triggering a training sample constructed based on the consultation problem and the switched target customer service agent to automatically optimize the customer service agent distribution model, and using the optimized customer service agent distribution model for subsequent customer service agent distribution.

2. The method of claim 1 wherein the receiving the consultation questions from the clients and determining text information characterizing the consultation questions comprises:

receiving a consultation problem sent by a client;

extracting first text content in the consultation questions;

identifying the voice information in the consultation questions to obtain second text content for representing the voice information;

And splicing the first text content and the second text content into text information representing the consultation problem.

3. The method of claim 1, wherein the customer service agent allocation model is composed of a feature extraction model and a classification model, wherein classifying the text information according to the pre-trained customer service agent allocation model determines candidate customer service agents for providing consultation services to the customer in at least one customer service agent, comprising:

inputting the text information into the feature extraction model to extract feature information of the text information;

and classifying the characteristic information by using the classification model to determine candidate customer service agents for providing consultation services for the customers in at least one customer service agent.

4. The method of claim 1, wherein after the receiving feedback information of the candidate customer service agent for the advisory problem, the method further comprises:

and if the feedback information characterizes that the consultation problem is good at being processed, determining the candidate customer service agent as a target customer service agent.

5. The method of claim 3, wherein classifying the feature information using the classification model to determine candidate customer service agents for providing advisory services to the customer in at least one customer service agent comprises:

Inputting the characteristic information into the classification model, and outputting matching information of each customer service agent in at least one customer service agent relative to the consultation problem, wherein the matching information characterizes the degree of treatment of the consultation problem;

acquiring idle state information of each customer service agent in the at least one customer service agent at present;

and determining candidate customer service agents for providing consultation services for the customers from the at least one customer service agent based on the idle state information of each customer service agent in the at least one customer service agent and the matching information of each customer service agent relative to the consultation questions.

6. The method of claim 1, wherein each of the customer service agents indicated by the list of customer service agents carries a recommendation index, the recommendation index of the customer service agent being related to current idle state information of the customer service agent and matching information of the customer service agent to the consultation questions.

7. A customer service agent distribution device, characterized by comprising:

the target customer service agent determining unit is used for receiving feedback information of the candidate customer service agents on the consultation problems; if the feedback information indicates that the consultation problem is not good at being processed, determining a customer service agent list, wherein the customer service agent list indicates all the customer service agents which are on line currently except the candidate customer service agent in the at least one customer service agent; pushing the customer service agent list to the candidate customer service agents; responding to the selection operation of the candidate customer service agents on each customer service agent indicated by the customer service agent list, and determining the customer service agent selected by the candidate customer service agents as a target customer service agent; the target customer service agent is another customer service agent which is good at processing the consultation problem and to which the candidate customer service agent transfers the consultation problem;

The optimizing unit is used for triggering a training sample constructed automatically based on the consultation problem and the target customer service agent after switching to optimize the customer service agent distribution model if the target customer service agent and the candidate customer service agent are not the same customer service agent, and using the optimized customer service agent distribution model for subsequent customer service agent distribution.

8. A server, comprising: at least one memory and at least one processor; the memory stores a program, and the processor invokes the program stored in the memory, where the program is configured to implement the customer service agent allocation method according to any one of claims 1 to 6.

9. A computer-readable storage medium having stored therein computer-executable instructions for performing the customer service agent allocation method of any one of claims 1-6.