Customer service scheduling method and device, storage medium and electronic equipment

Abstract

The invention discloses a customer service scheduling method and device, a storage medium and electronic equipment, and the method comprises the steps: determining the current real-time service information, and employing a pre-trained consultation amount prediction model according to the real-time service information, and determining the predicted consultation amount. And obtaining the current real-time consultation amount, and determining the fluctuation rate according to the real-time consultation amount and the predicted consultation amount. And when the fluctuation rate reaches a preset threshold value, generating a customer service scheduling instruction, and scheduling the customer service according to the customer service scheduling instruction. Through the consultation amount prediction model, the current predicted consultation amount can be predicted more accurately, and the fluctuation rate is determined according to the predicted consultation amount and the real-time consultation amount, so that the fluctuation condition of the consultation amount of the user can be determined quickly and accurately. When the fluctuation rate reaches the preset threshold value, the customer service scheduling instruction is automatically generated, the customer service is scheduled in time, long-time waiting of the user is avoided, better experience is brought to the user, and therefore the satisfaction degree of the user is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method, device, storage medium and electronic device for customer service scheduling. Background Art

[0002] With the continuous development of science and technology, online customer service has received widespread attention and has been widely used in all walks of life.

[0003] Currently, users can consult online customer service based on the problems they encounter in the process of executing business. Online customer service can respond to users' inquiries in a timely manner, determine the corresponding replies to the inquiries, and return them to users to solve the problems encountered by users, thereby improving user satisfaction. However, due to the increase in the number of users and the diversification of consultation content, the number of existing customer service staff is insufficient, making it difficult to quickly respond to users' inquiries. Therefore, how to dispatch customer service staff to quickly respond to users' inquiries is a very important issue.

[0004] Based on this, this specification provides a method for customer service scheduling. Summary of the invention

[0005] This specification provides a customer service scheduling method, device, storage medium and electronic device to partially solve the above-mentioned problems existing in the prior art.

[0006] This manual adopts the following technical solutions:

[0007] This specification provides a customer service scheduling method, including:

[0008] Determine current real-time business information;

[0009] According to the real-time business information, a pre-trained consultation volume prediction model is used to determine the predicted consultation volume;

[0010] Obtaining a current real-time consultation volume, and determining a volatility based on the real-time consultation volume and the predicted consultation volume;

[0011] When the volatility reaches a preset threshold, a customer service dispatch instruction is generated;

[0012] Dispatch customer service according to the customer service dispatch instruction.

[0013] Optionally, determining a volatility according to the real-time consultation volume and the predicted consultation volume specifically includes:

[0014] Determining the difference between the real-time consultation volume and the predicted consultation volume;

[0015] The volatility is determined based on the difference and the predicted consultation amount.

[0016] Optionally, when the volatility reaches a preset threshold, a customer service dispatch instruction is generated, specifically including:

[0017] Get all the consultation contents currently entered by all users;

[0018] Generate a consultation ranking according to the consultation contents;

[0019] A customer service dispatch instruction is generated based on the consultation ranking.

[0020] Optionally, generating a consultation ranking according to the consultation contents specifically includes:

[0021] Constructing a vocabulary according to each consultation content, wherein the vocabulary includes a plurality of first words;

[0022] According to the consultation contents, for each first word, determining the consultation content including the first word as the first content;

[0023] Determining a first frequency of the first word according to the number of the first content and the number of each consultation content;

[0024] For each consultation content, determine the first word frequency corresponding to the first vocabulary according to the consultation content;

[0025] Determining the first importance of the first vocabulary in the consultation content according to the first frequency and the first word frequency;

[0026] The consultation contents are sorted according to the first importance respectively corresponding to the first words in the consultation contents to obtain a consultation ranking.

[0027] Optionally, the consultation contents are sorted according to the first importance respectively corresponding to the first words in the consultation contents to obtain the consultation ranking, which specifically includes:

[0028] Determining a target word corresponding to the consultation content according to the first importances respectively corresponding to the first words of the consultation content;

[0029] According to the target words corresponding to the respective consulting contents, the consulting contents are grouped to obtain consulting groups, wherein the consulting contents included in each consulting group have the same target words;

[0030] Determining a first quantity of consulting contents included in each consulting group;

[0031] The consultation contents are sorted in descending order of the first quantity to obtain a consultation ranking.

[0032] Optionally, dispatching customer service according to the customer service dispatch instruction specifically includes:

[0033] Generate dispatch information according to the customer service dispatch instruction;

[0034] Determine a communication mode corresponding to the scheduling information and use it as a target communication mode;

[0035] The scheduling information is sent to the operator according to the target communication method.

[0036] Optionally, pre-training a consultation volume prediction model specifically includes:

[0037] Obtain historical consultation information from the customer service consultation database;

[0038] Determine historical business information based on the consulting information and use it as a training sample;

[0039] Determine the consultation volume in the historical time corresponding to the training sample and use it as the label of the training sample;

[0040] The consultation volume prediction model to be trained is trained according to the training samples and the annotations.

[0041] This specification provides a customer service scheduling device, including:

[0042] A first determination module, used to determine current real-time business information;

[0043] A prediction module, used to determine the predicted consultation volume based on the real-time business information using a pre-trained consultation volume prediction model;

[0044] A second determination module is used to obtain the current real-time consultation volume and determine the volatility according to the real-time consultation volume and the predicted consultation volume;

[0045] A generation module, used for generating a customer service dispatch instruction when the volatility reaches a preset threshold;

[0046] The scheduling module is used to schedule customer service according to the customer service scheduling instruction.

[0047] This specification provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned customer service scheduling method is implemented.

[0048] This specification provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned customer service scheduling method when executing the program.

[0049] At least one of the above technical solutions adopted in this specification can achieve the following beneficial effects:

[0050] The customer service scheduling method provided in this specification determines the current real-time business information, and uses a pre-trained consultation volume prediction model to determine the predicted consultation volume based on the real-time business information. Then the current real-time consultation volume is obtained, and the volatility is determined based on the real-time consultation volume and the predicted consultation volume. When the volatility reaches a preset threshold, a customer service scheduling instruction is generated, and the customer service is scheduled according to the customer service scheduling instruction.

[0051] It can be seen from the above method that when scheduling customer service, the server of this application can determine the current real-time business information, and based on the real-time business information, use the pre-trained consultation volume prediction model to determine the predicted consultation volume. Then obtain the current real-time consultation volume, and determine the volatility based on the real-time consultation volume and the predicted consultation volume. When the volatility reaches the preset threshold, a customer service scheduling instruction is generated, and the customer service is scheduled according to the customer service scheduling instruction. The consultation volume prediction model can more accurately predict the current predicted consultation volume, and determine the volatility based on the predicted consultation volume and the real-time consultation volume, so that the fluctuation of the user's consultation volume can be determined quickly and accurately. When the volatility reaches the preset threshold, a customer service scheduling instruction is automatically generated, and the customer service is dispatched in time to avoid users waiting for a long time, bring better experience to users, and thus improve user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The drawings described herein are used to provide a further understanding of this specification and constitute a part of this specification. The illustrative embodiments and descriptions of this specification are used to explain this specification and do not constitute an improper limitation on this specification. In the drawings:

[0053] Figure 1 A flowchart of a customer service scheduling method provided in this specification;

[0054] Figure 2 A schematic diagram of a customer service dispatching device structure provided in this specification;

[0055] Figure 3 A method corresponding to the Figure 1 Schematic diagram of the structure of an electronic device. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solutions and advantages of this specification more clear, the technical solutions of this specification will be clearly and completely described below in combination with the specific embodiments of this specification and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this specification.

[0057] Currently, users may encounter problems in the process of performing business. When encountering problems, users can consult online customer service. Online customer service can respond to users' inquiries in a timely manner, determine the corresponding reply to the inquiry, and return it to the user, thereby solving the problems encountered by the user and improving user satisfaction. However, there may be a sudden increase in the number of inquiries in a certain period of time, that is, a large number of users are consulting. At this time, the number of existing customer service staff is insufficient, and it is difficult to respond to users' inquiries in a timely manner, which brings bad health to users and reduces user satisfaction.

[0058] Based on this, this specification provides a method, device, storage medium and electronic device for customer service scheduling. The following is a detailed description of the technical solutions provided by each embodiment of this specification in conjunction with the accompanying drawings.

[0059] Figure 1 A flowchart of a customer service scheduling method provided in this specification includes the following steps:

[0060] S100: Determine current real-time business information.

[0061] S102: According to the real-time business information, a pre-trained consultation volume prediction model is used to determine the predicted consultation volume.

[0062] In this specification, the device for scheduling customer service can first determine the current real-time business information, and then determine the predicted consultation volume using a pre-trained consultation volume prediction model based on the real-time business information. Among them, the device for scheduling customer service can be a server, or an electronic device such as a desktop computer, a laptop computer, etc. For the convenience of description, the following only takes the server as the execution subject to illustrate the method of customer service scheduling provided in this specification. The above-mentioned real-time business information may include time information corresponding to the current moment, and may also include time information corresponding to the current time period, which is not specifically limited in this specification. The time information may include date and time, and may also include information such as week, solar terms, and festivals. For example, the time information may include x year x month x day x hour x minute, week x, Qingming Festival, and Qingming Festival. In addition, in addition to time information, the above-mentioned real-time business information may also include business information. The business may be a game, and the business information may be activity information and user information. The activity information is information about the current real-time activity, and the activity information may include the number of users participating in the game activity, the game activity time, the game activity content, and the game activity reward. The user information may include the number of users who have logged into the game and the number of historical consultations of the user. It should be noted that the above only takes the game business as an example, and the business can also be the business of other online customer service applications.

[0063] In this specification, the above-mentioned consultation volume prediction model may be a model pre-trained by the server, and the consultation volume prediction model may be an XGBoost (eXtreme Gradient Boosting) model. The consultation volume prediction model may be based on the gradient boosting (Gradient Boosting) framework, and is a strong prediction model composed of multiple weak prediction models (i.e., decision trees). Of course, the consultation volume prediction model may also be a model of the GPT (Generative Pre-trained Transformer) series, or a traditional machine learning model, i.e., a linear regression model, a logistic regression model, a support vector machine (SVM) model, etc., and a deep learning model, i.e., a feedforward neural network, a Transformer structure model, etc., which is not specifically limited in this specification. In the above step S102, the server may input real-time business information into the pre-trained consultation volume prediction model to determine the predicted consultation volume. The predicted consultation volume is the number of user consultations predicted by the consultation volume prediction model. In addition, in order to ensure that the format of the real-time business information is consistent with the format of the data input into the consultation volume prediction model, the server may determine the input format of the consultation volume prediction model, and adjust the real-time business information according to the input format, and input the adjusted real-time business information into the pre-trained consultation volume prediction model to determine the predicted consultation volume. The input format may be pre-set by the server, and the input format may be an array format or a string format, which is not specifically limited in this specification.

[0064] S104: Obtain the current real-time consultation volume, and determine the volatility based on the real-time consultation volume and the predicted consultation volume.

[0065] In this specification, the server may first obtain the current real-time consultation volume, and determine the volatility based on the real-time consultation volume and the predicted consultation volume. The real-time consultation volume is the number of current real-time user consultations. The volatility characterizes the fluctuation of the user's consultation volume, that is, the difference between the real-time consultation volume and the predicted consultation volume. The larger the volatility, the greater the difference between the real-time consultation volume and the predicted consultation volume, and the greater the fluctuation of the user's consultation volume. Conversely, the smaller the volatility, the smaller the difference between the real-time consultation volume and the predicted consultation volume, and the smaller the fluctuation of the user's consultation volume. When determining the volatility based on the real-time consultation volume and the predicted consultation volume, the server may determine the difference between the real-time consultation volume and the predicted consultation volume. Then determine the volatility based on the difference and the predicted consultation volume. Specifically, the server may determine the ratio of the difference to the predicted consultation volume, and use it as the volatility.

[0066] S106: When the volatility reaches a preset threshold, a customer service dispatch instruction is generated.

[0067] S108: Dispatching customer service according to the customer service dispatching instruction.

[0068] In this description, when the volatility reaches a preset threshold, the server can generate a customer service dispatch instruction. Then dispatch the customer service according to the customer service dispatch instruction. Among them, the above preset threshold is a value preset by the server. When the volatility reaches the preset threshold, it means that the current user's consultation volume fluctuates greatly, and the number of existing customer services is insufficient, and other customer services need to be dispatched to avoid users waiting for a long time. However, when the volatility has not reached the preset threshold, it means that the current user's consultation volume fluctuates little, and there is no need to dispatch other customer services at present. The above customer service dispatch instruction is used to instruct the server to dispatch customer service. It should be noted that the customer service is manual customer service and online customer service. In addition, since the business can be a game, the customer service can be a game customer service.

[0069] When dispatching customer service according to the customer service dispatch instruction, the server can directly send the customer service dispatch instruction to the operator, so that the operator can quickly respond to the user's consultation after receiving the customer service dispatch instruction, solve the user's questions in time, and ensure that the problem is solved in time. The operator can be a customer service staff who is not currently working, or a customer service supervisor who manages the customer service staff, and this manual does not make specific restrictions. Of course, the operator as the customer service supervisor can dispatch the customer service according to the customer service dispatch instruction, that is, the customer service supervisor can dispatch the selected customer service to perform customer service work according to the customer service dispatch instruction after receiving the customer service dispatch instruction. As for how many customer service personnel to dispatch, it can be determined by the operator based on the real-time consultation volume and the predicted consultation volume, so the above-mentioned customer service dispatch instruction can also include the real-time consultation volume and the predicted consultation volume. When sending the customer service dispatch instruction to the operator, the server can use any communication method such as text message, email, and telephone to send the customer service dispatch instruction to the operator.

[0070] In addition, when dispatching customer service according to the customer service dispatch instruction, the server can generate dispatch information according to the customer service dispatch instruction. Then determine the communication method corresponding to the dispatch information, and use it as the target communication method, and then send the dispatch information to the operator according to the target communication method. Among them, the above dispatch information represents the dispatch customer service, and the dispatch information can be the dispatch customer service. The above communication method may include multiple communication methods such as text messages, emails, and telephone calls. The above target communication method is one of the above multiple communication methods. The specific communication method can be randomly determined by the server, and can also be determined according to the dispatch information. This specification does not make specific limitations.

[0071] Based on this, taking the target communication mode determined by the scheduling information as an example, the above scheduling information may also include a volatility. When determining the communication mode corresponding to the scheduling information and using it as the target communication mode, the server may determine the difference between the volatility and the above preset threshold. If the difference is greater than the preset value, the first communication mode is used as the target communication mode corresponding to the scheduling information. If the difference is not greater than the preset value, the second communication mode is used as the target communication mode corresponding to the scheduling information. Among them, the preset value may be preset by the server. The larger the difference between the above volatility and the above preset threshold, the more it means that the number of customer services currently required is greater, the more it is necessary to dispatch customer services as soon as possible to avoid users waiting, and the more it is necessary to notify the operator as soon as possible. Therefore, it is necessary to select a communication mode that can quickly notify the operator and enable the operator to respond quickly, that is, the first communication mode. The first communication mode may be a telephone. Of course, the telephone may be an application of the telephone function provided by the mobile phone itself, or a third-party application, which may provide a telephone function. On the contrary, the smaller the difference between the above volatility and the above preset threshold, the smaller the number of customer services currently required is, so the second communication method can be selected to notify the operator. The response time of the second communication method may be longer than the response time of the first communication method. The response time may be the time for the operator to dispatch customer service according to the scheduling information. The second communication method may be SMS and email.

[0072] It can be seen from the above method that when scheduling customer service, the server of this application can determine the current real-time business information, and based on the real-time business information, use the pre-trained consultation volume prediction model to determine the predicted consultation volume. Then obtain the current real-time consultation volume, and determine the volatility based on the real-time consultation volume and the predicted consultation volume. When the volatility reaches the preset threshold, a customer service scheduling instruction is generated, and the customer service is scheduled according to the customer service scheduling instruction. The consultation volume prediction model can more accurately predict the current predicted consultation volume, and determine the volatility based on the predicted consultation volume and the real-time consultation volume, so that the fluctuation of the user's consultation volume can be determined quickly and accurately. When the volatility reaches the preset threshold, a customer service scheduling instruction is automatically generated, and the customer service is dispatched in time to avoid users waiting for a long time, bring better experience to users, and thus improve user satisfaction.

[0073] In this specification, in order to better understand the content of user consultation, quickly respond to questions raised by users, help customer service personnel solve user problems more efficiently, and improve user satisfaction. In the above step S106, the server can also obtain the consultation content input by all current users, and generate a consultation ranking based on each consultation content. According to the consultation ranking, a customer service scheduling instruction is generated. Among them, the consultation content is the content entered by the user when raising a consultation. The above customer service scheduling instruction can also include a consultation ranking.

[0074] When generating the consulting ranking according to each consulting content, the server can determine the similarity between any two consulting contents, and group each consulting content according to the similarity between each consulting content to obtain each first group. Determine the number of consulting contents included in each first group, sort each consulting content in descending order according to the number, and obtain the consulting ranking. Among them, the server can use any existing algorithm or machine learning model for calculating similarity to determine the similarity between any two consulting contents, and the server can also use a pre-set algorithm for calculating similarity or a pre-trained machine learning model for calculating similarity to determine the similarity between any two consulting contents.

[0075] The similarity of the consulting contents in each of the above-mentioned first groups is relatively high. Therefore, when the consulting contents are grouped according to the similarity between the consulting contents to obtain each first group, the server can determine other consulting contents whose similarity with the consulting content reaches the first value for each consulting content in turn according to the similarity between the consulting contents, and divide the consulting content and other consulting contents into one group. Determine each first group, and the similarity between all consulting contents included in the first group reaches the first value, and the first value can be a value preset by the server. For example, assuming that there are consulting contents 1 to 4, the similarity between consulting content 1 and consulting content 2 is 90%, the similarity between consulting content 1 and consulting content 3 is 20%, and the similarity between consulting content 1 and consulting content 4 is 93%. The similarity between consulting content 2 and consulting content 3 is 10%, and the similarity between consulting content 2 and consulting content 4 is 94%. The similarity between consulting content 3 and consulting content 4 is 30%. Assuming that the first value is 80%, the server can determine other consulting contents with a similarity of 50% with consulting content 1, namely consulting content 2 and consulting content 4, for each consulting content in turn according to the similarity between the consulting contents. Therefore, the server can group consulting content 1, consulting content 2 and consulting content 4 into one group. Similarly, other consulting contents that have not been divided can also be divided in the above manner, so the server can obtain the first group 1 and the first group 2. The first group 1 can include consulting content 1, consulting content 2 and consulting content 4, and the first group 2 can include consulting content 3.

[0076] When the consultation contents are sorted in the order of the number from large to small to obtain the consultation ranking, the server may first sort the first groups in the order of the number from large to small to obtain the first sequence, and then sort the consultation contents in the order of the first sequence to obtain the consultation ranking. It should be noted that the order of all the consultation contents included in each first group may be randomly determined by the server, and the above consultation ranking may be a sequence composed of the consultation contents.

[0077] Continuing with the above example, the number of consulting contents included in the first group 1 is 3, and the number of consulting contents included in the first group 2 is 1. The first groups are sorted in order of the number from large to small to obtain the first sequence, that is, the first group 1 and the first group 2. Then, the consulting contents are sorted in the order of the first sequence to obtain the consulting ranking, and the order of all consulting contents included in each first group can be randomly determined by the server. Therefore, the server can randomly determine the order of consulting content 1, consulting content 2 and consulting content 4 included in the first group 1. Assuming the order is consulting content 4, consulting content 1, consulting content 2, the above consulting ranking can be "consulting content 4, consulting content 1, consulting content 2, consulting content 3".

[0078] In addition, when generating the consulting ranking according to each consulting content, the server may also use the TF-IDF (TermFrequency-Inverse Document Frequency) algorithm to generate the consulting ranking according to each consulting content. Specifically, the server may first construct a vocabulary list according to each consulting content. The vocabulary list includes a number of first words, and the first words may be obtained based on each consulting content. Specifically, the server may perform a word segmentation operation on each consulting content to obtain a number of second words, and determine each first word according to each second word. Among them, when determining each first word according to each second word, the server may directly use each second word as each first word. The server may also group each second word to obtain each second group, and each second group includes the same second word. For each second group, a second word is randomly selected from the second group and used as the first word.

[0079] Then, the server may determine the consultation content including the first word for each first word according to each consultation content, and use it as the first content. Then, according to the number of first contents and the number of each consultation content, the first frequency of the first word is determined. When determining the first frequency of the first word, the server may use the following formula for calculation:

[0080]

[0081] Among them, t represents the first word, IDF t represents the first frequency of t, N represents the number of each consulting content, and df(t) represents the number of first contents.

[0082] Afterwards, the server can determine the first word frequency corresponding to the first word for each consultation content according to the consultation content. Specifically, the server can determine the words included in the consultation content for each consultation content, and use them as the third words, determine the number of words in the third words that are the same as the first word, and use them as the target number, that is, how many words in the third words are the same as the first word. Then determine the ratio of the target number to the number of the third words, and use it as the first word frequency corresponding to the first word. Specifically, the following calculation formula can be used for calculation:

[0083]

[0084] Among them, s represents the number of the third word, n t Indicates the number of targets, TF t Represents the first word frequency corresponding to t.

[0085] Then, the server may determine the first importance of the first word in the consultation content according to the first frequency and the first word frequency. Specifically, the server may determine the product of the first frequency and the first word frequency as the first importance of the first word in the consultation content, which may be calculated using the following formula:

[0086] TF-IDF=TF t ×IDF t

[0087] Among them, TF-IDF represents the first importance.

[0088] Finally, the server may sort each consulting content according to the first importance corresponding to each first word in each consulting content to obtain a consulting ranking. Specifically, the server may determine the target word corresponding to the consulting content according to the first importance corresponding to each first word of the consulting content. According to the target words corresponding to each consulting content, each consulting content is grouped to obtain each consulting group. The first number of consulting contents included in each consulting group is determined. According to the order of the first number from large to small, each consulting content is sorted to obtain a consulting ranking. Among them, when determining the target word corresponding to the consulting content according to the first importance corresponding to each first word of the consulting content, the server may determine the first word corresponding to the largest first importance according to the first importance corresponding to each first word included in the consulting content, and use it as the target word of the consulting content. The target words of the consulting content included in each consulting group are the same. The process of sorting each consulting content in the order of the first number from large to small to obtain a consulting ranking is similar to the process of sorting each consulting content in the order of the number from large to small to obtain a consulting ranking, and will not be repeated here.

[0089] In addition, when the consulting contents are sorted according to the first importance corresponding to each first word in each consulting content to obtain the consulting ranking, the server can also determine the first importance corresponding to each first word in each consulting content for each first word, and use it as the second importance. Determine the mean of each second importance. Sort the first words in order of the mean from large to small to obtain a second sequence. Sort the consulting contents according to the second sequence and the first importance corresponding to each first word in each consulting content to obtain the consulting ranking. Among them, when the consulting contents are sorted according to the second sequence and the first importance corresponding to each first word in each consulting content to obtain the consulting ranking, the server can first determine the consulting content including the first word for each first word in the second sequence in turn, and use it as the first consulting content, and sort the first consulting content according to the first importance of the first word in the first consulting content from large to small. Afterwards, continue with the next first word of the first word, and determine the consulting content including the next first word from other consulting contents except the above-mentioned first consulting content, and re-use it as the first consulting content, and continue to sort the newly determined first consulting content until all consulting contents are sorted, and obtain the consulting ranking composed of each consulting content.

[0090] For example, assuming that there are first words 1 to 3 and consulting contents 1 to 4, and the second sequence is first word 1, first word 3, and first word 2. The server may first determine the consulting content including the first word 1 for each first word in the second sequence, and assume that the consulting content 1 and consulting content 2 both include the first word 1, and the first importance of the first word 1 in consulting content 1 is 0.9, and the first importance of the first word 1 in consulting content 2 is 0.7, so the first consulting content is consulting content 1 and consulting content 2. Afterwards, the first consulting contents are sorted in descending order according to the first importance of the first word 1 in the first consulting content (i.e., consulting content 1 and consulting content 2), i.e., consulting content 1 and consulting content 2. Afterwards, continue with the next first word of the first word, that is, first word 3, and determine the consulting content including the next first word from other consulting contents (that is, consulting content 3 and consulting content 4) except the above-mentioned first consulting content, and re-use it as the first consulting content. Assuming that consulting content 4 includes the first word 3, continue to sort the newly determined first consulting content (that is, consulting content 4) until all consulting contents are sorted, and obtain the consulting ranking composed of each consulting content, that is, consulting content 1, consulting content 2, consulting content 4, and consulting content 3.

[0091] In this specification, in order to display the consultation ranking more clearly and intuitively, so that the operation personnel can better understand the hot consultation content of the user, after obtaining the consultation ranking, in addition to generating customer service scheduling instructions based on the consultation ranking, the server can also use data visualization technology to visualize the consultation ranking to the operation personnel. The data visualization technology can be any existing data visualization technology, specifically a data visualization algorithm or a machine learning model, etc., which is not specifically limited in this specification. The visual display can be in the form of a chart, so the server can use data visualization technology to generate a chart corresponding to the consultation ranking, and visualize the chart to the operation personnel.

[0092] In this specification, the above-mentioned consultation volume prediction model may be a model pre-trained by the server, so when pre-training the consultation volume prediction model, the server may determine the historical consultation information, and determine the historical business information based on the consultation information, and use it as a training sample. Determine the consultation volume within the historical time corresponding to the training sample, and use it as a label for the training sample. According to the training samples and the labels, the consultation volume prediction model to be trained is trained. Among them, the above-mentioned historical consultation information may be consultation information within a certain time or a certain time period in history, and the consultation information may include business information and the user's consultation volume within a certain time or a certain time period in history. When the consultation volume prediction model to be trained is trained according to the training samples and the labels, the server may input the training samples into the consultation volume prediction model to be trained, determine the prediction results, and determine the training loss based on the prediction results and the labels, and train the consultation volume prediction model to be trained based on the training loss.

[0093] In addition, the above historical consultation information can be stored in the customer service consultation database, so the server can also obtain the historical consultation information from the customer service consultation database. According to the consultation information, the historical business information is determined and used as a training sample. The consultation volume in the historical time corresponding to the training sample is determined and used as the annotation of the training sample. According to the training sample and the annotation, the consultation volume prediction model to be trained is trained.

[0094] In this specification, after the consultation volume prediction model is trained, the server can create a call interface for calling the consultation volume prediction model based on the trained consultation volume prediction model, and save it. Wherein, the call interface can be created by the server in any existing way, and the call interface can be an API (Application Programming Interface) interface. Of course, when creating a call interface for calling the consultation volume prediction model, the server can also create a Flask application to load the consultation volume prediction model and provide an API interface for calling the consultation volume prediction model, and the Flask application provides an API interface for calling the consultation volume prediction model. Therefore, in the above-mentioned step S102, the server can determine the call interface for calling the pre-trained consultation volume prediction model, and then input the real-time business information into the consultation volume prediction model through the call interface to determine the predicted consultation volume output by the consultation volume prediction model. Wherein, the call interface can be an API interface for calling the consultation volume prediction model, and can also be an API interface for calling the consultation volume prediction model provided by the pre-created Flask application.

[0095] In this specification, the server can determine whether to dispatch customer service according to a preset period, so in the above step S100, when the preset period is reached, the server can determine the current real-time business information. The preset period can be pre-set by the server, and the preset period can be 2 minutes or other time periods, which is not specifically limited in this specification.

[0096] Based on this, when the volatility does not reach the preset threshold, the server can wait for a preset period of time, and when the current time reaches the preset period, re-set the volatility according to the preset threshold. Figure 1 The customer service scheduling method shown in the figure continues to schedule customer service.

[0097] In this specification, the server can periodically train the above-mentioned consultation volume prediction model. Specifically, the server can continue to train the above-mentioned consultation prediction model according to a specified period. The specific training process is similar to the training process of the above-mentioned pre-trained consultation volume prediction model, and will not be repeated here. The training samples and annotations of the above-mentioned training consultation prediction model can be the above-mentioned Figure 1 The customer service scheduling method shown is generated by the process of customer service scheduling.

[0098] The above are one or more implementation methods of this specification. Based on the same idea, this specification also provides a corresponding customer service scheduling device, such as Figure 2 shown.

[0099] Figure 2 A schematic diagram of a customer service dispatching device provided in this specification includes:

[0100] A first determination module 200, used to determine current real-time service information;

[0101] Prediction module 202, used to determine the predicted consultation volume based on the real-time business information using a pre-trained consultation volume prediction model;

[0102] The second determination module 204 is used to obtain the current real-time consultation volume and determine the volatility according to the real-time consultation volume and the predicted consultation volume;

[0103] A generating module 206, configured to generate a customer service dispatch instruction when the volatility reaches a preset threshold;

[0104] The scheduling module 208 is used to schedule customer service according to the customer service scheduling instruction.

[0105] Optionally, the second determination module 204 is specifically used to determine the difference between the real-time consultation volume and the predicted consultation volume; and determine the volatility based on the difference and the predicted consultation volume.

[0106] Optionally, the generating module 206 is specifically used to obtain various consultation contents currently input by all users; generate a consultation ranking according to the various consultation contents; and generate a customer service dispatch instruction according to the consultation ranking.

[0107] Optionally, the generation module 206 is specifically used to construct a vocabulary list according to the various consulting contents, wherein the vocabulary list includes a plurality of first words; according to the various consulting contents, for each first word, determine the consulting content including the first word and use it as the first content; according to the number of the first contents and the number of the various consulting contents, determine the first frequency of the first word; for each consulting content, determine the first word frequency corresponding to the first word according to the consulting content; according to the first frequency and the first word frequency, determine the first importance of the first word in the consulting content; according to the first importance corresponding to each first word in the various consulting contents, sort the various consulting contents to obtain a consulting ranking.

[0108] Optionally, the generation module 206 is specifically used to determine the target vocabulary corresponding to the consulting content according to the first importance corresponding to each first vocabulary of the consulting content; group the consulting contents according to the target vocabulary corresponding to each consulting content to obtain consulting groups, wherein the target vocabulary of the consulting contents included in each consulting group is the same; determine the first number of consulting contents included in each consulting group; and sort the consulting contents in descending order of the first number to obtain a consulting ranking.

[0109] Optionally, the scheduling module 208 is specifically used to generate scheduling information according to the customer service scheduling instruction; determine the communication method corresponding to the scheduling information and use it as the target communication method; and send the scheduling information to the operator according to the target communication method.

[0110] Optionally, the device further comprises:

[0111] The training module 210 is used to obtain historical consultation information from the customer service consultation database; determine historical business information based on the consultation information and use it as a training sample; determine the consultation volume within the historical time corresponding to the training sample and use it as a label for the training sample; and train the consultation volume prediction model to be trained based on the training sample and the label.

[0112] This specification also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 A customer service scheduling method is provided.

[0113] This manual also provides Figure 3 The one shown corresponds to Figure 1 A schematic diagram of the electronic device. Figure 3 As shown, at the hardware level, the electronic device includes a processor, an internal bus, a network interface, a memory, and a non-volatile memory, and may also include other hardware required for the business. The processor reads the corresponding computer program from the non-volatile memory into the memory and then runs it to achieve the above Figure 1 The method of customer service scheduling.

[0114] Of course, in addition to software implementation, this specification does not exclude other implementation methods, such as logic devices or a combination of software and hardware, etc., that is to say, the executor of the following processing flow is not limited to each logic unit, but can also be hardware or logic devices.

[0115] In the 1990s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0116] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.

[0117] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0118] For the convenience of description, the above device is described in various units according to their functions. Of course, when implementing this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0119] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems, or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0120] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0121] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0122] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0123] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0124] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0125] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0126] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0127] Those skilled in the art will appreciate that the embodiments of this specification may be provided as methods, systems or computer program products. Therefore, this specification may take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0128] This specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. This specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0129] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0130] The above description is only an embodiment of the present specification and is not intended to limit the present specification. For those skilled in the art, the present specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present specification shall be included in the scope of the claims of the present specification.

Claims

1. A method for customer service scheduling, characterized in that: include: Determine current real-time business information; According to the real-time business information, a pre-trained consultation volume prediction model is used to determine the predicted consultation volume; Obtaining a current real-time consultation volume, and determining a volatility based on the real-time consultation volume and the predicted consultation volume; When the volatility reaches a preset threshold, a customer service dispatch instruction is generated; Dispatch customer service according to the customer service dispatch instruction.

2. The method according to claim 1, characterized in that Determining a volatility according to the real-time consultation volume and the predicted consultation volume specifically includes: Determining the difference between the real-time consultation volume and the predicted consultation volume; The volatility is determined based on the difference and the predicted consultation amount.

3. The method according to claim 1, characterized in that When the volatility reaches a preset threshold, a customer service dispatch instruction is generated, specifically including: Get all the consultation contents currently entered by all users; Generate a consultation ranking according to the consultation contents; A customer service dispatch instruction is generated based on the consultation ranking.

4. The method according to claim 3, characterized in that According to the above-mentioned consultation contents, a consultation ranking is generated, including: Constructing a vocabulary according to each consultation content, wherein the vocabulary includes a plurality of first words; According to the consultation contents, for each first word, determining the consultation content including the first word as the first content; Determining a first frequency of the first word according to the number of the first content and the number of each consultation content; For each consultation content, determine the first word frequency corresponding to the first vocabulary according to the consultation content; Determining the first importance of the first vocabulary in the consultation content according to the first frequency and the first word frequency; The consultation contents are sorted according to the first importance respectively corresponding to the first words in the consultation contents to obtain a consultation ranking.

5. The method according to claim 4, characterized in that According to the first importance corresponding to each first word in each consultation content, the consultation contents are sorted to obtain a consultation ranking, which specifically includes: Determining a target word corresponding to the consultation content according to the first importances respectively corresponding to the first words of the consultation content; According to the target words corresponding to the respective consulting contents, the consulting contents are grouped to obtain consulting groups, wherein the consulting contents included in each consulting group have the same target words; Determining a first quantity of consulting contents included in each consulting group; The consultation contents are sorted in descending order of the first quantity to obtain a consultation ranking.

6. The method according to claim 1, characterized in that Dispatching customer service according to the customer service dispatch instruction, specifically including: Generate dispatch information according to the customer service dispatch instruction; Determine a communication mode corresponding to the scheduling information and use it as a target communication mode; The scheduling information is sent to the operator according to the target communication method.

7. The method according to claim 1, characterized in that Pre-training the consultation volume prediction model, including: Obtain historical consultation information from the customer service consultation database; Determine historical business information based on the consulting information and use it as a training sample; Determine the consultation volume in the historical time corresponding to the training sample and use it as the label of the training sample; The consultation volume prediction model to be trained is trained according to the training samples and the annotations.

8. A customer service dispatching device, characterized in that: include: A first determination module, used to determine current real-time business information; A prediction module, used to determine the predicted consultation volume based on the real-time business information using a pre-trained consultation volume prediction model; A second determination module is used to obtain the current real-time consultation volume and determine the volatility according to the real-time consultation volume and the predicted consultation volume; A generation module, used for generating a customer service dispatch instruction when the volatility reaches a preset threshold; The scheduling module is used to schedule customer service according to the customer service scheduling instruction.

9. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 7 is implemented.

Images