CN119830024A

CN119830024A - Customer service agent matching method and device, electronic equipment and storage medium

Info

Publication number: CN119830024A
Application number: CN202411675145.1A
Authority: CN
Inventors: 潘明杰
Original assignee: iFlytek Co Ltd
Current assignee: iFlytek Co Ltd
Priority date: 2024-11-21
Filing date: 2024-11-21
Publication date: 2025-04-15

Abstract

The invention provides a customer service agent matching method, a customer service agent matching device, electronic equipment and a storage medium, wherein the method comprises the steps of obtaining current dialogue data of a user, analyzing the current dialogue data based on a trained large language model to obtain user key characteristics of the user, and determining customer service agents matched with the user based on the user key characteristics and current index values of agent key indexes of the customer service agents, wherein the current index values are updated based on historical service records of the customer service agents. The customer service agent matching method, the customer service agent matching device, the electronic equipment and the storage medium provided by the invention utilize the strong data processing and learning capabilities of the trained large-scale language model to accurately capture the user key characteristics of the user, update the current index values of the agent key indexes of the customer service agent in time, comprehensively evaluate the multidimensional data of the user key characteristics and the agent key indexes, realize optimal matching and promote the intelligent and personalized development of customer service distribution.

Description

Customer service agent matching method and device, electronic equipment and storage medium

Technical Field

The present invention relates to the technical field of customer service centers, and in particular, to a customer service agent matching method, a device, an electronic device, and a storage medium.

Background

Along with the diversification of customer service demands, how to quickly and accurately match customer appeal with the most suitable customer service agents becomes a key for improving service quality.

The existing customer service distribution system mostly adopts a static or simple dynamic strategy, only data (such as waiting time) with a single dimension is often concerned, and the actual capability and real-time state of a customer service representative are not fully considered, so that a customer request cannot be always distributed to the most suitable customer service seat, thereby not only reducing service efficiency, but also influencing customer satisfaction.

Disclosure of Invention

The invention provides a customer service agent matching method, a customer service agent matching device, electronic equipment and a storage medium, which are used for solving the defects of poor service efficiency and customer satisfaction caused by a static or simple and dynamic customer service allocation strategy in the prior art.

The invention provides a customer service agent matching method, which comprises the following steps:

Acquiring current dialogue data of a user;

analyzing the current dialogue data based on the trained large language model to obtain user key characteristics of the user;

and determining the customer service agents matched with the user based on the user key characteristics and the current index values of the agent key indexes of the customer service agents, wherein the current index values are updated based on the historical service records of the customer service agents.

According to the customer service agent matching method provided by the invention, the current dialogue data is analyzed based on the trained large language model to obtain the user key characteristics of the user, and the method comprises the following steps:

Generating analysis input text based on analysis prompt information and the current dialogue data, wherein the analysis prompt information comprises an instruction for determining key characteristics of a user;

And inputting the analysis input text into the large language model to obtain the user key characteristics output by the large language model.

According to the customer service agent matching method provided by the invention, the analysis prompt information also comprises question information and expected output aiming at the key characteristics of the user.

According to the customer service agent matching method provided by the invention, the step of determining the customer service agent matched with the user based on the key characteristics of the user and the current index values of the agent key indexes of the customer service agents comprises the following steps:

determining a matching score between the client and each customer service agent based on the similarity between the user key features and the agent key indexes and the current index values;

And determining a customer service agent matched with the user based on the matching score.

According to the customer service agent matching method provided by the invention, the step of determining the matching score between the customer and each customer service agent based on the similarity between the key characteristics of the customer and the key indexes of the agent and the current index value comprises the following steps:

Determining an initial matching score between a client and each customer service agent based on the weight value of each user key feature, the similarity and the current index value, wherein the weight value represents the importance degree of the user key feature to agent matching;

And determining the matching score based on the current index value of the service satisfaction index in the agent key index and the initial matching score.

According to the customer service agent matching method provided by the invention, after the customer service agent matched with the user is determined, the method further comprises the following steps:

receiving service evaluation of the user for the customer service agent;

And updating the current index value of the seat key index of the customer service seat based on the service evaluation.

According to the customer service agent matching method provided by the invention, the key characteristics of the user comprise at least one of user appeal, emotion state and emergency degree, and the key indexes of the agent comprise at least one of skill level, compression resistance, emotion response capability and service satisfaction rate.

The invention also provides a customer service seat matching device, which comprises:

The data acquisition unit is used for acquiring the current dialogue data of the user;

the data analysis unit is used for analyzing the current dialogue data based on the trained large language model to obtain the user key characteristics of the user;

And the agent matching unit is used for determining the customer service agents matched with the user based on the key characteristics of the user and the current index values of the agent key indexes of the customer service agents, and the current index values are updated based on the history service records of the customer service agents.

The invention also provides an electronic device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the customer service agent matching method according to any one of the above when executing the program.

The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements a customer service agent matching method as described in any of the above.

The invention also provides a computer program product comprising a computer program which when executed by a processor implements a customer service agent matching method as described in any one of the above.

The customer service agent matching method, the customer service agent matching device, the electronic equipment and the storage medium can accurately capture the user key characteristics of the user by utilizing the strong data processing and learning capabilities of the trained large-scale language model, compare with the simple key word matching to determine the service requirement of the user, accurately extract the user key characteristics in multiple dimensions, comprehensively evaluate the user key characteristics and the multidimensional data of the agent key indexes by updating the current index values of the agent key indexes of the customer service agent in time, realize optimal matching, and promote customer service distribution to intelligent and personalized development.

Drawings

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

Fig. 1 is a schematic flow chart of a customer service agent matching method provided by the invention.

Fig. 2 is a flowchart of step 120 in the customer service agent matching method provided by the present invention.

Fig. 3 is a schematic flow chart of step 130 in the customer service agent matching method provided by the present invention.

Fig. 4 is a second flowchart of step 130 in the customer service agent matching method provided by the present invention.

Fig. 5 is a second flow chart of the customer service agent matching method provided by the invention.

Fig. 6 is a schematic structural diagram of a customer service agent matching device provided by the invention.

Fig. 7 is a schematic structural diagram of an electronic device provided by the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, 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.

In the current conventional customer service agent matching process, the main steps generally include the following aspects:

① And the user requests access, namely when the user needs service, the user submits the service request to the customer service center through telephone, online chat or other communication modes.

② And the basic problem identification is that the system primarily analyzes the request of the user and determines the basic information and service requirement of the user through simple keyword matching or preset question-answer pairs.

③ And the static allocation strategy is to allocate the user request to the corresponding customer service agent according to the basic information of the user and preset rules (such as service type, time and the like). This strategy does not take into account the specific capabilities and current state of customer service.

④ Dynamic adjustment in some systems, if it is found that the first allocated customer service cannot effectively process the request, a simple dynamic adjustment may be performed, such as reallocation according to the current idle state of the customer service.

⑤ And the service processing is that after the customer service agent receives the allocated user request, the customer service agent starts to process the problem of the user, and provides solutions or performs further service operation.

⑥ Service end and feedback after service completion, the system may invite the user to evaluate the service to collect service quality feedback.

In summary, the customer service agent matching process in the existing customer service system comprises the steps of static allocation, dynamic allocation, service processing, service feedback and the like, and aims to improve service efficiency, reduce the burden of manual customer service and ensure that users obtain satisfactory service experience.

However, the prior art often only focuses on data in a single dimension (such as waiting time), and ignores other important factors such as skills, experience, and historic performance of agents, so that customer requests cannot always be allocated to the most appropriate customer service agents. This one-sided and lack of personalized assessment mechanisms limits the improvement of dispensing efficiency and quality.

Aiming at the problems, the embodiment of the invention provides a customer service agent matching method, which utilizes the strong data processing and learning capabilities of a trained large language model to accurately capture the user key characteristics of a user, updates the current index value of the agent key index of the customer service agent in time, comprehensively evaluates the multi-dimensional data of the user key characteristics and the agent key index, realizes optimal matching, and promotes customer service distribution to develop to intellectualization and individuation.

The embodiment of the invention can be applied to the scene needing customer service agent matching. The execution main body of the method can be electronic equipment such as terminal equipment, computers, servers, server clusters or specially designed customer service agent matching equipment, and also can be a customer service agent matching device arranged in the electronic equipment, and the customer service agent matching device can be realized by software, hardware or a combination of the two.

Fig. 1 is one of flow diagrams of a customer service agent matching method provided by the present invention, as shown in fig. 1, the method includes the following steps:

step 110, obtaining current dialogue data of a user;

Step 120, analyzing the current dialogue data based on the trained large language model to obtain the user key characteristics of the user;

And 130, determining the customer service agents matched with the user based on the key characteristics of the user and the current index values of the agent key indexes of the customer service agents, wherein the current index values are updated based on the historical service records of the customer service agents.

In particular, the current dialog data may be man-machine dialog data, such as dialog data generated by a user interacting with an AI intelligent customer service. For example, the user interacts with the customer service system through the terminal equipment (mobile phone, computer, etc.), and can generate dialogue data through chat application, voice call, video call, etc., and the data acquisition layer captures the data in real time as current dialogue data. Current dialog data includes, but is not limited to, text, voice, images, and the like. The captured dialog data may be further pre-processed, such as to remove noise, standardized formats, text segmentation, etc., for subsequent analysis.

Here, the large language model (Large Language Model, LLM) is simply referred to as a large model, which refers to a natural language processing (Natural Language Processing, NLP) model with huge parameter scale, the number of model parameters and/or the complexity of model structure exceed a preset threshold, and the model processes large-scale text data during training, and has the capability of understanding and generating natural language. For example, the large language model may include a star fire large model, or the like.

Further, the large language model herein may be a large language model in a general field, or may be a large language model in a customer service field obtained by fine tuning data in a customer service center field, for example, the large language model may be obtained by fine tuning by applying a large amount of sample dialogue data and key feature description information on the basis of a large language model trained in advance.

On the basis, the current dialogue data can be analyzed through a large language model, and the user key characteristics of the user are obtained. For example, the current dialogue data may be included in an input text of a large language model, the input text is input into the large language model, the dialogue content is analyzed by the large language model, and key features such as emotional tendency (such as positive, negative, neutral), question type (such as technical consultation, complaint, advice, etc.), key requirements (such as solving a question, seeking advice, etc.) and the like of the user are extracted. Here, the dimensions of the key features of the user may be multidimensional, and key information expressed by the user in the current dialogue data can be comprehensively reflected.

Here, the key index of the customer service agent is also multidimensional, and reflects the personal characteristics and advantages of the customer service agent from multiple dimensions. The current index value is the value corresponding to the key index of the customer service seat at the current moment, and the current index value is updated along with the historical service record of the customer service seat. The historical service records of the customer service agents are obtained from the database, and the current key index values of the agents, such as average response time, problem resolution, customer satisfaction index and the like, are calculated by utilizing big data analysis technology, such as statistics or machine learning algorithm.

After the key characteristics and the current index values of the user are obtained, a matching algorithm (such as multi-objective optimization, distance measurement and the like) can be used for determining customer service agents matched with the user. In the matching process, priorities can be set according to service demands, such as priority matching of agents with high satisfaction and low complaint rate, customer service with skill level matched with the complaint of the user can be preferentially matched according to the complaint type of the user, customer service with strong concentricity can be preferentially matched with emotion of the user, and the matching degree between key characteristics of the user and the current index value of each customer service agent can be comprehensively considered, so that a comprehensive matching degree score is calculated, and customer service with highest priority matching score is calculated.

For example, assuming a customer service sheetlet, he handled a number of issues regarding product return in past service records and obtained a high degree of customer evaluation. Through large model analysis, the small sheets can be found to be excellent in processing the return goods problem, and the small sheets are mild in communication style and durable, so that the discontent customer emotion can be effectively calmed. Thus, when the system receives a request for product refund, such a request may be preferentially assigned to the sheetlet.

The method provided by the embodiment of the invention can accurately capture the user key features of the user by utilizing the strong data processing and learning capabilities of the trained large language model, can accurately extract the user key features in multiple dimensions compared with the simple key word matching to determine the service requirements of the user, and comprehensively evaluate the user key features and the multidimensional data of the seat key indexes by updating the current index values of the seat key indexes of the customer service seat in time so as to realize optimal matching and promote the customer service distribution to develop intelligently and individually.

The method provided by the embodiment of the invention comprises the steps of firstly, more accurately knowing the actual capability and the current state of each customer service through comprehensive evaluation of multi-dimensional data, secondly, carrying out deep analysis on the appeal and emotion of the user by utilizing a large model to enable the matching process to be more accurate and personalized, and finally adopting an intelligent matching algorithm to dynamically adjust the agent allocation strategy so as to adapt to the continuously changing service requirements and the customer emotion, thereby remarkably improving the quality and the efficiency of the service, and better meeting the customer requirements and enhancing the user experience compared with a simple or static allocation strategy.

In some embodiments, the large language model is derived based on sample dialogue data through unsupervised pre-training and instruction fine-tuning. The method comprises the steps of acquiring multi-round dialogue texts in the field from the Internet, removing privacy information (names, addresses, telephones and the like) of users in a keyword mode and the like, and cleaning open source data for unsupervised pre-training.

The whole model structure can adopt a Decoder-only structure, takes tarnsformer structure as a main network, and the network module mainly comprises a Multi-head technology and a fully-connected network.

Training criteria:

The large language model can predict the next word according to the input of the first k words, and measures the quality of the trained language model according to indexes such as confusion (Perplexity). And then, carrying out instruction fine adjustment, using a large amount of man-machine conversation and customer service seat conversation data, carrying out conversation operation regularity, facilitating user appeal analysis and emotion analysis and carding, enabling a large model to fully learn characteristic capture of a user, and inputting information of characteristic key description into the large model, so as to give data more conforming to the characteristics.

In some embodiments, the user key features include at least one of user appeal, emotional state, urgency, and agent key indicators include at least one of skill level, stress-resistance, emotional response capability, and service satisfaction.

Specifically, a user complaint refers to a main complaint expressed by a current dialog user. Through natural language processing technology, semantic analysis is carried out on the consultation content of the user, specific requirements and problems of the user are identified, and the user requests are divided into different categories, such as product consultation, technical support, account problems, complaint feedback, service improvement suggestions and the like.

Emotion analysis is an integral part of understanding the needs of users. By analyzing the user's linguistic words, intensity of mood, etc., the emotional state of the user may be inferred. The current emotional states are classified into five categories, calm, slightly dissatisfied, apparently dissatisfied, anger, and despair.

And distinguishing emergency and non-emergency situations according to the emergency degree of the problem judged by the words and the language of the customer, so as to ensure that the emergency problem can be responded quickly.

It is contemplated that in conventional systems, the skills and experience of customer service agents are often underutilized. Due to the lack of effective assessment and matching mechanisms, highly skilled and experienced customer service agents may not fully exploit their potential, while novice or poorly performing agents may be overused. The agent key indicators in this embodiment may include skill levels. Skill level refers to the expertise and ability of a customer service agent to solve a problem in a particular area (e.g., technical support, account management, etc.). This includes knowledge of the extent of the product or service, industry knowledge, and experience with handling a particular type of problem. The skill level directly influences whether the customer service agent can accurately and rapidly solve the problem of customers, is one of important indexes for measuring the customer service quality, and can ensure that customer requests are distributed to the most suitable agents by evaluating the skill expertise of the customer service agent, thereby improving the efficiency and quality of solving the problem.

The emotion of the customer is one of the important factors affecting the quality of service. However, the related art has obvious shortcomings in understanding and coping with the emotion of the client, which may cause misunderstanding or conflict in the service process, further affecting the client experience. The key indicators of the seat in this embodiment may include emotion response capability. Emotion response capability refers to the ability to understand and respond to a customer's emotion. This includes concentricity, communication skills, and conflict resolution capabilities. Emotional response capability is critical to establishing a good customer relationship. The effective emotion management not only can improve customer satisfaction, but also can help to relieve the tense service scene and reduce negative feedback. By analyzing the emotion response capability of the customer service agents, the customer requests which need high concentricity and patience can be better matched, so that the overall service quality is improved.

The real-time status of the customer service agents (e.g., workload, pressure level, etc.) has a direct impact on the quality of service. The key indicators of the seat in this embodiment may include pressure resistance. The compressive resistance refers to the coping strategies and mental adjustment capabilities of customer service agents in the face of complaints, emergencies or other high-pressure situations. In customer service, challenging and stressful situations are often encountered. The customer service seat can be kept cool by having good stress coping capability, so that the problem is effectively solved, and emotional reaction is avoided. Assessing the coping pressure capability of customer service agents helps to ensure that complex or urgent customer requests are handled quickly and properly, ensuring continuity and stability of service.

Based on any of the above embodiments, fig. 2 is a schematic flow chart of step 120 in the customer service agent matching method provided by the present invention, as shown in fig. 2, based on a trained large language model, the present dialogue data is analyzed to obtain the user key features of the user, that is, step 120 specifically includes:

step 121, generating analysis input text based on analysis prompt information and current dialogue data, wherein the analysis prompt information comprises an instruction for determining key characteristics of a user;

Step 122, inputting the analysis input text into the large language model to obtain the user key characteristics output by the large language model.

Specifically, in order to extract user key features through a large language model, an input for the large language model may be constructed based on current dialogue data, and a Prompt (Prompt) of the large language model. In the embodiment of the invention, a prompt for generating key characteristics of a user aiming at a large language model is recorded as analysis prompt information, and input aiming at the large language model is recorded as analysis input text.

Here, the analysis hint information is used to specify the generation requirements of the user key features at the input stage of the large language model, i.e. the analysis hint information is used to instruct the large language model to determine the user key features. Thus, in the analysis prompt, instructions for instructing the large language model to generate the key features of the user may be included, where the instructions are in the form of natural language text, such as "please extract the key features of the user in the following dialog content", or "suppose you are a language analysis expert, please analyze the main appeal of the user based on the following dialog content, and extract the key features of the user".

After the current dialog data is obtained, an analysis input text may be generated based on the analysis prompt message and the current dialog data. For example, the analysis prompt information and the current dialogue data may be spliced to obtain an analysis input text, or the analysis prompt information and the current dialogue data may be filled into a preset input text template to obtain an analysis input text, which is not particularly limited in the embodiment of the present invention.

After the analysis input text is obtained, the analysis input text can be input into a large language model, the dialogue content is understood by the large language model, the main appeal of the user is identified, the emotion of the user is analyzed, the emergency degree of the problem is judged, and therefore key characteristics of the user are extracted.

According to the method provided by the embodiment of the invention, the analysis input text which is input as the large language model is generated by analyzing the prompt information and the current dialogue data, so that the automatic extraction of the key features of the user is realized by calling excellent semantic understanding and analysis capability of the large language model.

Based on any of the above embodiments, the analysis hint information further includes questioning information and expected output for key features of the user.

Specifically, the analysis hint information may include, in addition to instructions to determine the user key features to instruct the large language model to generate the user key features, question information and expected output for the user key features to facilitate directing the large language model to generate user key features that conform to the user's expectations.

The question information is a question set by a pointer to the type of the user key feature desired to be output, and the desired output is a desired answer to the question information by the pointer. For example, the user key characteristics of the expected output comprise the type of the complaint, and question information' what is the main category of the user complaint? as another example, the user key features of the desired output include the emotional state of the user, and the question information "what level the user is currently in.

And the large language model outputs the key characteristics of the user meeting the requirements through understanding and analyzing the analysis prompt information.

In some embodiments, analyzing the input text may be shown as follows:

"you are a linguistic analysis expert, knowing that the user AI customer service dialogue is as follows:

[ user ]: feed, hello | how much I last call cost is |!

AI customer service, you good, i are intelligent customer service. Please provide your cell phone number.

[ User ] I want to know why my telephone charge is so expensive

AI customer service brings inconvenience to you due to the very sorry. To better help you solve the problem, we suggest that you first provide some basic information so we can find the problem faster. If you insist to transfer manual customer service, I will record and schedule for you as soon as possible.

[ User ]: i no matter how much i now turn to manual | my credit, I all quickly go crazy of-!

AI customer service, which is very well understood to your emotion, we will switch to manual customer service for you as soon as possible. Let you wait a little bit and we will immediately handle you.

Please combine the above dialog content, analyze and answer the following:

1. What are the main classifications of user appeal? complaint feedback, product consultation, technical support, account problems and service improvement;

2. What is the current mood of the user? calm, slight dissatisfaction, obvious dissatisfaction, anger and despair;

3. What is the level of urgency of the current user's demand.

Through analysis, the large language model outputs the key characteristics of the user:

Complaint feedback.

Anger.

Is very urgent.

That is, based on the current dialogue content, the type of the user's complaint is complaint feedback, the current emotional state of the user is anger, and the emergency degree of the current user's demand is very urgent.

Based on any of the above embodiments, fig. 3 is one of the flow diagrams of step 130 in the customer service agent matching method provided by the present invention, as shown in fig. 3, based on the key characteristics of the user and the current index values of the agent key indexes of the customer service agents, the determining a customer service agent matched with the user, that is, step 130 specifically includes:

Step 131, determining a matching score between the customer and each customer service agent based on the similarity between the user key features and the agent key indexes and the current index values;

step 132, determining a customer service agent matching the user based on the matching score.

Specifically, matching is performed for the user and the customer service agents, and the current index value of the agent key index of each customer service agent can be determined by comprehensively considering the key characteristics of the user.

Firstly, calculating the similarity between the key features of the user and the key indexes of the seat, wherein common similarity measurement methods comprise cosine similarity, pearson correlation coefficient, spearman correlation coefficient and the like. One or more of the methods may be selected for use in combination according to the actual situation.

For each user key feature, the similarity between the user key feature and all the agent key indexes is calculated. The user key features and the agent key indicators can be regarded as vectors, and then the cosine value of the included angle or other similarity indicators between the vectors are calculated. It can be understood that the higher the similarity is, the greater the association between the user key features and the agent key indexes is, whereas the lower the similarity is, the smaller the association between the user key features and the agent key indexes is. And further, according to the calculated similarity value, establishing an association relationship between the key characteristics of the user and the key indexes of the seat. The user key features with higher similarity values can be associated with the agent key indexes to form an association pair.

Then, based on the current index value and the similarity, a matching score between the customer and each customer service agent is calculated. And carrying out weighted summation on the similarity and the current index value to obtain a matching score. And selecting the customer service agent with the highest score as the agent matched with the user according to the matching score.

If there are multiple agents with matching scores that are very close, further screening may be performed in consideration of other factors (e.g., agents' experience, skills, etc.).

According to the method provided by the embodiment of the invention, the matching score between the client and each customer service agent is determined based on the similarity between the key characteristics of the user and the key indexes of the agents and the current index value, and the customer service agent matched with the user is determined based on the matching score, so that the user request can be optimally distributed to the most suitable customer service agent.

Based on any of the above embodiments, fig. 4 is a second flowchart of step 130 in the customer service agent matching method provided by the present invention, as shown in fig. 4, the step 131 of determining a matching score between a customer and each customer service agent based on a similarity between a user key feature and an agent key index and a current index value includes:

step 131-1, determining an initial matching score between the client and each customer service agent based on the weight value, the similarity and the current index value of each user key feature, wherein the weight value represents the importance degree of the user key feature on agent matching;

Step 131-2, determining a matching score based on the current index value of the service satisfaction index in the agent key index and the initial matching score.

Specifically, the weight value of each user key feature may be preset, so as to reflect the importance degree of each user key feature to the seat matching. The higher the importance level is, the larger the weight value is, and the lower the importance level is, the smaller the weight value is. The weight values of the user key features may be shown as shown in table 1.

TABLE 1

The weight value of the key characteristics of the user, the similarity between the key indexes of the agents and the current index value can be multiplied for each customer service agent, and then weighted summation is carried out to obtain an initial matching score. And then obtaining a matching score based on the current index value of the service satisfaction index in the seat key index and the initial matching score.

In some embodiments, assume that U represents a set of user appeal types, C represents a set of customer service capabilities, for each user appealAnd each customer serviceDefining a matching functionCan be expressed as:

Wherein the method comprises the steps of Is the weight value of the kth user key feature,Measuring user appealAnd customer serviceSimilarity in the kth feature, T is the current index value.

Eventually, the system will choose to makeMaximum customer serviceDistribution to users。

For example, the weight values of the key features of the user are shown in table 2 below:

TABLE 2

The current coordinate values of the customer service agents are shown in the following table 3:

TABLE 3 Table 3

Calculating a matching score between the current customer and each customer service using the above formula:

The matching score of the customer and customer service X is (1×90+0.8×65+0.8×70) ×8.85=198×8.85= 175.23

The matching score of the customer and customer service Y is (1×80+0.8×75+0.8×90) ×0.912=212×0.912= 193.344

From the calculation result, it can be seen that the matching score of the customer a and the customer service Y is highest, and therefore the customer can be assigned to the customer service Y.

According to the method provided by the embodiment of the invention, the weight value is given to each user key feature through the importance degree of the user key feature on the seat matching, so that the matching algorithm is further optimized, and the accuracy of customer service seat allocation is improved.

Based on any of the above embodiments, after determining the customer service agent matched with the user, further includes:

Receiving service evaluation of a user for a customer service seat;

based on the service evaluation, the current index value of the agent key index of the customer service agent is updated.

Specifically, the present embodiment further includes a dynamic scoring mechanism for service satisfaction based on user service evaluation to ensure that optimal allocation decisions can be made at any given moment, taking into account real-time state changes of customer service agents.

After the customer service agent and the user are in communication, the user can evaluate the satisfaction degree of the overall quality of the current customer service, score the customer service, and divide the customer service into satisfaction and dissatisfaction. After receiving the service evaluation of the user for the customer service agent, the service satisfaction rate= (satisfaction score number/total score number) of the customer service agent may be updated in real time. The service satisfaction index is an important index for customer service scoring and is used in the next round of matching calculation of users and customer service agents.

Based on any of the above embodiments, fig. 5 is a second flowchart of a customer service agent matching method provided by the present invention, as shown in fig. 5, where the method includes:

S1, collecting information such as skills, experience, historical performances and the like of customer service agents, collecting historical interaction records of users, and determining current index values of agent key indexes of the customer service agents. And selecting training samples, cleaning and labeling, and performing fine tuning training on the large language model.

And S2, acquiring current dialogue data of the user, and analyzing the current dialogue data based on the trained large language model to obtain user key characteristics of the user, wherein the user key characteristics comprise at least one of user appeal, emotion state and emergency degree.

And S3, determining a matching score between the client and each customer service agent based on the similarity between the key features of the user and the key indexes of the agents and the current index value, and determining the customer service agent matched with the user based on the matching score. The seat key index comprises at least one of skill level, compression resistance, emotion response capability and service satisfaction rate.

And S4, receiving service evaluation of the user for the customer service agent, and updating the current index value of the agent key index of the customer service agent based on the service evaluation.

The embodiment of the invention utilizes the powerful data processing and learning capabilities of the large model to accurately capture the user's appeal and emotion change, and combines the actual capability and state of the customer service agent to realize accurate matching. Through deep analysis of user complaints and moods, a more personalized service scheme can be provided, so that user satisfaction is remarkably improved. The intelligent matching algorithm reduces unnecessary switching and waiting time and improves processing speed and overall service efficiency. The matching method can be adjusted according to different scenes and requirements, and has strong flexibility and expansibility. And the real-time state of the user demand and customer service is deeply analyzed by utilizing a large model technology, so that highly accurate seat matching is realized. And providing customized service according to emotion and appeal of the user, and enhancing user satisfaction. The intelligent algorithm reduces manual intervention, quickens response speed and improves service flow efficiency.

According to the embodiment of the invention, the capability of comprehensively evaluating each seat is ensured by comprehensively considering the multiple dimension data such as the skill level, the compressive capacity, the emotion response capability and the like of the customer service seat. The user demand is more accurately understood by analyzing the characteristics of the user such as appeal, emotion and the like, and more careful service is provided. And optimally distributing the user request to the most suitable customer service agents according to the comprehensive evaluation result and the user characteristics by utilizing an advanced matching algorithm.

The customer service agent matching device provided by the invention is described below, and the customer service agent matching device described below and the customer service agent matching method described above can be correspondingly referred to each other.

Based on any of the above embodiments, fig. 6 is a schematic structural diagram of a customer service agent matching device provided by the present invention, and as shown in fig. 6, the device includes:

a data acquisition unit 610 for acquiring current dialogue data of a user;

A data analysis unit 620, configured to analyze the current dialogue data based on the trained large language model, so as to obtain user key features of the user;

The agent matching unit 630 is configured to determine a customer service agent matched with the user based on the user key feature and a current index value of an agent key index of each customer service agent, where the current index value is updated based on a history service record of the customer service agent.

Based on any of the above embodiments, the data analysis unit is specifically configured to:

Based on any of the above embodiments, the analysis prompt further includes questioning information and expected output for the user key features.

Based on any of the above embodiments, the agent matching unit is specifically configured to:

Based on any one of the above embodiments, the method further includes an index updating unit configured to:

receiving service evaluation of the user for the customer service agent;

Fig. 7 illustrates a physical schematic diagram of an electronic device, which may include a processor (processor) 710, a communication interface (Communications Interface) 720, a memory (memory) 730, and a communication bus 740, where the processor 710, the communication interface 720, and the memory 730 communicate with each other via the communication bus 740, as shown in fig. 7. The processor 710 may invoke logic instructions in the memory 730 to perform a customer service agent matching method including obtaining current dialogue data of a user, analyzing the current dialogue data based on a trained large language model to obtain user key features of the user, determining customer service agents matching the user based on the user key features and current index values of agent key indicators of the customer service agents, the current index values being updated based on historical service records of the customer service agents.

Further, the logic instructions in the memory 730 described above may be implemented in the form of software functional units and may be stored in a computer readable storage medium when sold or used as a stand alone product. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. The storage medium includes a U disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, an optical disk, or other various media capable of storing program codes.

In another aspect, the invention further provides a computer program product, the computer program product comprises a computer program, the computer program can be stored on a non-transitory computer readable storage medium, when the computer program is executed by a processor, the computer can execute the customer service agent matching method provided by the methods, the method comprises the steps of obtaining current dialogue data of a user, analyzing the current dialogue data based on a trained large language model to obtain user key features of the user, and determining customer service agents matched with the user based on the user key features and current index values of agent key indexes of the customer service agents, wherein the current index values are updated based on historical service records of the customer service agents.

In still another aspect, the present invention further provides a non-transitory computer readable storage medium, on which a computer program is stored, the computer program being implemented when executed by a processor to perform the customer service agent matching method provided by the above methods, the method comprising obtaining current dialogue data of a user, analyzing the current dialogue data based on a trained large language model to obtain user key features of the user, determining a customer service agent matching the user based on the user key features and current index values of agent key indexes of the customer service agents, the current index values being updated based on historical service records of the customer service agents.

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

It should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present invention, and not for limiting the same, and although the present invention has been described in detail with reference to the above-mentioned embodiments, it should be understood by those skilled in the art that the technical solution described in the above-mentioned embodiments may be modified or some technical features may be equivalently replaced, and these modifications or substitutions do not make the essence of the corresponding technical solution deviate from the spirit and scope of the technical solution of the embodiments of the present invention.

Claims

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

Acquiring current dialogue data of a user;

2. The customer service agent matching method according to claim 1, wherein the analyzing the current dialogue data based on the trained large language model to obtain the user key feature of the user comprises:

3. The customer service agent matching method of claim 2, wherein the analysis prompt message further comprises question information and expected output for the user key features.

4. The customer service agent matching method according to claim 1, wherein the determining the customer service agent matched with the user based on the user key feature and the current index value of the agent key index of each customer service agent comprises:

5. The customer service agent matching method according to claim 4, wherein the determining a matching score between the customer and each of the customer service agents based on the similarity between the user key feature and the agent key index and the current index value includes:

6. The customer service agent matching method according to any one of claims 1 to 5, wherein after the customer service agent matching the user is determined, further comprising:

receiving service evaluation of the user for the customer service agent;

7. The customer service agent matching method according to any one of claims 1 to 5, wherein the user key features include at least one of user appeal, emotional state, urgency, and at least one of skill level, compression resistance, emotional response capability, and service satisfaction rate.

8. A customer service agent matching apparatus, comprising:

9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the customer service agent matching method of any one of claims 1 to 7 when the program is executed by the processor.

10. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements the customer service agent matching method of any of claims 1 to 7.