Customer service support system and customer service support program

Abstract

To allow for prompt optimization of customer service.SOLUTION: An operation terminal 2 used by each communicator C who responds to a telephone from a customer and a support server 3 are connected so as to be communicable. Each operation terminal 2 transmits a response time and a response content for each telephone response to the support server 3. The support server 3 evaluates telephone response skills of each communicator C on the basis of the response time, and generates a talk script for the telephone response on the basis of the response content of communicators C with high evaluated skills.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a customer support system and a customer support program for assisting in telephone response to customers.

Background Art

[0002] For example, when a new service or a fee menu is introduced in an electric power company, inquiries about the service and the like from consumers pour into the customer center / call center of the electric power company. In order to appropriately answer and respond to such inquiries, the electric power company prepares answers (talk scripts, Q&A) for assumed inquiries in advance. However, there are cases where unexpected inquiries are received from consumers, or it takes time to satisfy consumers with the prepared answers, and it is necessary to revise the talk script.

[0003] On the other hand, a system is known in which troubles that occur during a telephone response to one customer and countermeasures therefor are reflected in telephone responses to other customers (see, for example, Patent Document 1). After detecting the end of a telephone response to a customer by an operator who makes a telephone response, this system inquires the operator about the troubles that occurred during the telephone response, and collects trouble information regarding troubles that may occur during the telephone response. Then, the collected trouble information is analyzed and classified into customer-specific trouble information for which individual countermeasures should be found for each customer and general trouble information for which countermeasures applicable to all customers should be constructed.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] By the way, conventionally, when revising a talk script, it has been necessary to listen to the conversation content from each communicator / operator at the customer center and the requests from customers, etc., and consider what kind of answers would be optimal, which has required a great deal of time and effort. In particular, because it takes time, there has been a risk that prompt response and improvement may not be carried out. On the other hand, in the system of Patent Document 1, countermeasures are not constructed unless trouble occurs. Therefore, in cases where customers are not satisfied even though no trouble occurs, countermeasures are not constructed. Moreover, after the end of the telephone response, in order to inquire the operator about the trouble that occurred during the telephone response and collect trouble information, not only is the burden on the operator large, but it also takes time to collect information and construct countermeasures.

[0006] Therefore, an object of the present invention is to provide a customer response support system and a customer response support program that enable prompt optimization of responses to customers.

Means for Solving the Problems

[0007] In order to solve the above problems, the invention according to claim 1 is characterized in that an operation terminal used by each communicator who conducts a telephone response to a customer and a support server are communicably connected, and each of the operation terminals transmits the response time and response content for each telephone response to the support server, and the support server includes skill evaluation means for evaluating the telephone response skills of each communicator based on the response time, and script generation means for generating a talk script for telephone response based on the response content of a communicator with high skills evaluated by the skill evaluation means.

[0008] The invention according to claim 2 is characterized in that, in the customer response support system according to claim 1, the skill evaluation means periodically evaluates the telephone response skills of each communicator, and the script generation means periodically generates the talk script.

[0009] The invention according to claim 3 is the customer support system according to claim 1, wherein the skill evaluation means evaluates the skill for each title indicating the subject matter of the inquiry from the customer, and the script generation means generates the talk script for each title.

[0010] The invention according to claim 4 is the customer support system according to claim 1, wherein the skill evaluation means evaluates the skill based on the number of telephone responses to the customer within a predetermined time for each communicator.

[0011] The invention according to claim 5 is the customer support system according to claim 1, wherein the script generation means uses a learning model for a script learned based on past performance data so that the talk script is output when the response content of a communicator with high skill is input.

[0012] The invention according to claim 6 is a customer support program characterized by causing a computer to function as skill evaluation means for evaluating the telephone response skill of each communicator who makes a telephone response to a customer based on the response time for each telephone response, and script generation means for generating a talk script for telephone response based on the response content for each telephone response of a communicator with high skill evaluated by the skill evaluation means.

[0013] The invention according to claim 7 is the customer support program according to claim 6, wherein the skill evaluation means evaluates the skill for each title indicating the subject matter of the inquiry from the customer, and the script generation means generates the talk script for each title.

[0014] The invention according to claim 8 is the customer support program according to claim 6, wherein the skill evaluation means evaluates the skill based on the number of telephone responses to the customer within a predetermined time for each communicator.

[0015] The invention according to claim 9 is characterized in that, in the customer support program according to claim 6, the script generation means uses a learning model for scripts that is machine-learned based on past performance data so that the talk script is output when the response content of a communicator with high skills is input.

Advantages of the Invention

[0016] According to the inventions described in claims 1 and 6, a talk script for telephone response is generated based on the response content of a communicator (operator) with high telephone response skills (using the response content as a model). Therefore, by all communicators responding to calls according to this talk script, it becomes possible to optimize the response to customers. Moreover, since the talk script for telephone response is automatically generated based on the response time and response content of each communicator, it becomes possible to optimize the response to customers quickly and without burdening the communicators.

[0017] In addition, a communicator with a short response time for telephone response is considered to have high telephone response skills, etc. Therefore, it becomes possible to appropriately evaluate the skills by evaluating the telephone response skills of communicators based on the response time for each telephone response. Furthermore, by appropriately evaluating and selecting communicators with high skills, it becomes possible to generate an appropriate talk script.

[0018] According to the invention described in claim 2, since the talk script is generated regularly, it becomes possible to continuously optimize the response to customers by generating an appropriate and up-to-date talk script corresponding to the telephone responses of each communicator performed daily.

[0019] According to the inventions described in claims 3 and 7, since the talk script is generated for each title indicating the subject matter of an inquiry from a customer, it becomes possible to optimize the response to customers according to the title.

[0020] In addition, a communicator who handles a large number of phone calls to customers within a predetermined time is considered to have high phone call handling skills. According to the inventions described in claim 4 and claim 8, since the phone call handling skills are evaluated based on the number of phone calls made to customers by each communicator within a predetermined time, it becomes possible to appropriately evaluate the skills. Furthermore, by appropriately evaluating and selecting communicators with high skills, it becomes possible to generate an appropriate talk script.

[0021] According to the inventions described in claim 5 and claim 9, since a talk script is output using a learning model for the machine-learned script, it becomes possible to generate a more appropriate talk script and more appropriately optimize the response to customers.

Brief Description of the Drawings

[0022]

Figure 1

Figure 2

Figure 3

Figure 4

Embodiments for Carrying Out the Invention

[0023] Hereinafter, this invention will be described based on the illustrated embodiments.

[0024] FIG. 1 is a schematic configuration diagram showing a customer response support system 1 according to an embodiment of this invention. This customer response support system 1 is a system that supports phone calls to customers. In this embodiment, the case of supporting each communicator (operator) C at a customer center operated by an electric power company will be mainly described.

[0025] The customer support system 1 is communicably connected to a plurality of operation terminals 2 respectively used by each communicator C who conducts telephone correspondence with a customer (a person in need), and a support server 3.

[0026] The operation terminal 2 is a communication terminal and computer used by the communicator C to conduct telephone correspondence with the customer, has functions equivalent to those of a general operation terminal in a call center, and also has functions such as transmitting the response time and response content for each telephone correspondence (each call) to the support server 3. That is, as shown in FIG. 2, it mainly includes a headset 21, an input unit 22, a display unit 23, a communication unit 24, a counter 25, a memory 26, and a central processing unit 27 that controls these components.

[0027] The headset 21 includes headphones (earphones) and a microphone, and enables the communicator C to have a conversation with the customer by wearing it on the head. Here, the operation terminal 2 is provided with a response button for responding to an incoming call, etc. By turning on the response button, etc., the headphones and the microphone are turned on and the call starts. Also, when the response button is turned off, etc., or when the customer ends the conversation (when the end tone is detected), the headphones and the microphone are turned off and the call ends. And the time from the start of this call to the end of the call becomes the response time described later.

[0028] The input unit 22 is an interface for inputting various information, commands, etc., and is composed of a keyboard, a mouse, etc. Specifically, it inputs the general outline of the conversation content and response content with the customer, and things that are particularly concerning in the conversation with the customer (claims, new questions, etc.). Here, a response sheet for inputting such information and the identification information of the communicator C, etc. is displayed on the display unit 23 described later, and input and entry are made according to this response sheet.

[0029] At this time, a title indicating the subject matter of the inquiry from the customer is entered in the response sheet. This title is a name for identifying services, systems, matters proposed by the customer, etc. that the customer wants to know, and is determined in advance. For example, in the case of a time-limited discount service described later, it is set as "Highlights (7)" or the like.

[0030] The display unit 23 is a display for displaying various data, information, etc. Specifically, it displays a response sheet as described above, or displays a talk script for telephone response described later. Here, the talk script corresponding to the title input by the input unit 22 is displayed. That is, when a title is input, the title is transmitted to the support server 3, and the talk script corresponding to this title is transmitted from the support server 3 and displayed on the display unit 23.

[0031] The communication unit 24 is an interface for communicating with the outside via the Internet network, telephone communication network, etc. Specifically, it transmits the response time and response content for each telephone response to the support server 3, and receives a talk script from the support server 3. The counter 25 is an instrument for measuring time, and the memory 26 is a storage unit for storing various data, information, voices, etc.

[0032] And in this embodiment, the time from the start to the end of the call is measured by the counter 25, and this time is transmitted to the support server 3 as the response time for each telephone response. Also, the response sheet filled in by the communicator C is transmitted to the support server 3 as the response content.

[0033] On the other hand, by directly transmitting (streaming) the voice of the call / conversation to the support server 3, the response time and response content for each phone call handling may be transmitted to the support server 3. In this case, the voice call (recording) time becomes the response time, and the recording itself, or its analysis result / summary, becomes the response content. Also, in this embodiment, for each phone call handling, that is, for each call end, the response time and response content (response sheet including the identification information of the communicator C) are transmitted to the support server 3, but they may be stored and accumulated in the memory 26 and periodically transmitted to the support server 3 in a batch.

[0034] The support server 3 is a server computer that generates and updates the talk script for phone call handling for customers in order to support each communicator C. As shown in FIG. 3, it mainly includes an input unit 31, a display unit 32, a communication unit 33, a storage unit 34, a skill evaluation task (skill evaluation means) 35, a script generation task (script generation means) 36, a learning task 37, and a central processing unit 38 that controls these.

[0035] The input unit 31 is an interface for inputting various information, commands, etc. Specifically, it inputs activation commands for the skill evaluation task 35 and the script generation task 36, or inputs characters for modifying the talk script. The display unit 32 is a display for displaying various data, information, etc. Specifically, it displays the output results by the skill evaluation task 35 and the script generation task 36. The communication unit 33 is an interface for communicating with the outside via the Internet network, telephone communication network, etc. Specifically, it receives the response time and response content for each phone call handling from each operation terminal 2, or transmits the talk script to each operation terminal 2.

[0036] The memory unit 34 mainly includes a communicator database (corresponding information storage means) 341, a learning model for scripts 342, and a performance database for scripts 343. Here, the communicator database 341 will be described, and the learning model for scripts 342 and the performance database for scripts 343 will be described later.

[0037] The communicator database 341 is a database that stores information related to the telephone correspondence of each communicator C. That is, based on the response time and response sheet received from each operation terminal 2, for each communicator C, the response time and response content for each telephone correspondence of each title are stored. Further, it stores the skills (skill values) for each title evaluated in the skill evaluation task 35 described later, the evaluation date and time, the number of telephone correspondence cases, and so on.

[0038] Here, the number of telephone correspondence cases is the number of telephone correspondences to customers within a predetermined time, and in this embodiment, it is calculated by the central processing unit 38. That is, every time a response sheet is received from the operation terminal 2, the number of processed cases of the corresponding communicator C is counted, and the number of processed cases within a predetermined time (for example, one day, one week, one month) is calculated as the number of telephone correspondence cases. Also, the number obtained by averaging the number of telephone correspondence cases is calculated as the average number of telephone correspondence cases.

[0039] Also, in this embodiment, the number of processed cases and the number of telephone correspondence cases are counted for each title, the average number of telephone correspondence cases for each title is calculated, and it is stored in the communicator database 341. Here, the predetermined time may be a set value / fixed value, may be made changeable, or may be set differently depending on the title.

[0040] The skill evaluation task 35 is a task program that evaluates the telephone response skills of each communicator C based on the response time for each telephone call and the number of telephone calls made to customers within a predetermined time. That is, it is started periodically or when a start command is input at the input unit 31, and based on information stored in the communicator database 341, etc., for each title indicating the subject matter of an inquiry from a customer, the telephone response skills of each communicator C are evaluated.

[0041] Specifically, first, for each communicator C and for each title, the average value of the response time required for telephone response is calculated as the average response time. Next, based on the average response time and the average number of telephone responses for each title of each communicator C, the skill value for each title of each communicator C is evaluated and calculated. For example, the shorter the average response time, the higher the points are set, and the more the average number of telephone responses, the higher the points are set, and points are assigned to each communicator C. Then, the communicator C with a high score is evaluated and set with a high skill value, and the lower the score, the lower the skill value is evaluated and set. Then, the evaluated skill value for each title of each communicator C is stored in the communicator database 341.

[0042] In this embodiment, such a skill evaluation task 35 periodically evaluates the skills of each communicator C. Here, "periodically" means repeating at a predetermined timing period, and it may be at regular intervals or at variable intervals. For example, the period may be changed for each title. If the title is related to a new service, the period may be shortened during the service introduction period and then lengthened later.

[0043] Also, when evaluating skills repeatedly in this way, the average response time and the average number of telephone responses in only the new period, excluding the previously evaluated period, may be calculated to evaluate the skills. That is, the skills in the latest period may be evaluated sequentially.

[0044] The script generation task 36 is a task / program that automatically generates and creates a talk script for telephone correspondence based on the corresponding content of communicator C with high skills evaluated in the skill evaluation task 35. That is, it is started periodically or when a start command is input at the input unit 31, and automatically generates a talk script for telephone correspondence for each title with reference to the corresponding content of communicator C with high skills in each title.

[0045] Here, at the time of startup, the title can be specified. If the title is not specified, talk scripts for all titles are generated. If the title is specified, only the talk script for that title is generated. The processing procedure when the title is specified is described below, but the same processing procedure is applied to all titles even when talk scripts are generated for all titles.

[0046] First, obtain the skill values of each communicator C for the title to be generated from the communicator database 341, and obtain the corresponding content / corresponding sheet of communicator C with high skill values (for example, the top several names or a value equal to or higher than a predetermined value). At this time, the corresponding content of multiple communicators C may be obtained. Next, based on the corresponding content, extract the corresponding content to be reflected in the talk script (such as the content of the reply to the customer) and generate the talk script. For example, extract the reply content common to communicators C with high skill values, or extract the reply content for which the customer is satisfied / convincing with the reply, and generate the talk script.

[0047] Furthermore, in this embodiment, a talk script (first draft) is prepared in advance for each title and sent to each operation terminal 2 and each communicator C. Also, when the generation of the talk script by the script generation task 36 is the second time, there is a first talk script. Therefore, the talk script is generated and updated based on the earliest talk script as well. For example, when the response to a question from a customer that was not in the earliest talk script is included in the corresponding content, the question and the response content are added to generate the talk script. Also, when additional explanations are given in the response content of the earliest talk script without giving a reply that satisfies or convinces the customer, a talk script including the explanations in the response content is generated.

[0048] Specifically, for example, if the title is related to a time-limited discount service and in the earliest talk script, in response to the question "What criteria are used to set the discount period and unit price?", it is described that the answer is "Regarding the discount period, it is determined considering ○○. Regarding the discount unit price, it is determined considering ○○, etc.".

[0049] In contrast, based on the corresponding content of the communicator C with a high skill value, a talk script is generated that answers "Regarding the discount period, it is determined considering ××. Regarding the discount unit price, it is determined considering ××, etc.".

[0050] Furthermore, in the corresponding content of the communicator C with a high skill value, following this, in response to the question "Shouldn't it be discounted throughout the year instead of at specific dates and times?", if the answer is "Considering △△, the service is set to be at specific dates and times", and a reply that satisfies and convinces the customer is given, then this question and the response content are added to generate the talk script.

[0051] For such a script generation task 36, a learning model 342 for scripts, which is machine-learned based on past performance data, is used so that a talk script is output when the response content of a highly skilled communicator C (the response content of an excellent person) is input. This learning model 342 for scripts is created by a learning task 37.

[0052] That is, the learning task 37 creates a learning model 342 for scripts by using a known machine learning algorithm such as a neural network with the past performance data recorded and stored in a performance database 343 for scripts. This performance database 343 for scripts is a database in which performance data including talk scripts of telephone responses created by experts and experienced persons related to the creation of talk scripts in a customer center based on the response content of excellent persons as input information is recorded and stored. Note that the past performance data includes, in addition to the actual response content of excellent persons and data created based on the talk scripts of telephone responses actually created by experts and experienced persons related to the creation of talk scripts, data created by pre-training and the like.

[0053] As shown in FIG. 4, this learning task 37 uses machine learning and deep learning using a neural network, and based on the performance data recorded in the performance database 343 for scripts, for example, a neural network is created with the response content of excellent persons as the input layer, the talk script of telephone response as the output layer, and the analysis process from the input layer to the output layer as the intermediate layer. Then, the learning task 37 uses the performance data of the learning model 342 for scripts as learning data to perform learning on various parameters in the intermediate layer. That is, the learning task 37 performs learning on various parameters in the intermediate layer so that a talk script is properly output based on the response content of excellent persons.

[0054] Then, the generated talk script is transmitted to each operation terminal 2. At this time, it may be transmitted to all the operation terminals 2 simultaneously at the time of generation, or as described above, only the talk script corresponding to the title input and transmitted from the operation terminal 2 may be transmitted to the operation terminal 2.

[0055] Such a script generation task 36 is basically started in conjunction with the activation of the skill evaluation task 35. That is, as described above, every time the skill evaluation task 35 is periodically activated, the script generation task 36 is activated following the skill evaluation task 35. On the other hand, the skill evaluation task 35 and the script generation task 36 may be activated at arbitrary timings respectively.

[0056] Next, the operation and the operation method of the customer support system 1 with such a configuration, and the customer support method by the customer support system 1 will be described.

[0057] First, when each communicator C answers a call from a customer and inputs a title indicating the subject of the inquiry on the response sheet on the operation terminal 2, a talk script corresponding to this title is transmitted from the support server 3 and displayed on the operation terminal 2. Then, the communicator C conducts a telephone response according to this talk script (for reference), and inputs the outline of the conversation content and response content with the customer, and the matters that were particularly concerning in the conversation with the customer, etc. into the response sheet. After that, when the telephone response is completed, the response sheet, the response content, and the response time are transmitted to the support server 3.

[0058] Next, in the support server 3, a skill evaluation task 35 is activated at a predetermined time, and as described above, the skills of telephone response for each title of each communicator C are evaluated. Subsequently, a script generation task 36 is activated, and as described above, based on the response content of the communicator C with high skills evaluated in the skill evaluation task 35, a talk script for telephone response is automatically generated and updated for each title. Then, the generated and updated talk script is transmitted to each operation terminal 2 at a predetermined timing, and each communicator C makes a telephone response according to the latest talk script.

[0059] As described above, according to this customer response support system 1 and customer response support method, based on the response content of the communicator C with high skills in telephone response (using the response content as a model), a talk script for telephone response is generated. Therefore, by having all communicators C make telephone responses according to this talk script, it is possible to optimize the response to the customers. Moreover, since the talk script for telephone response is automatically generated based on the response time and response content of each communicator C, it is possible to optimize the response to the customers quickly and without burdening the communicator C.

[0060] Also, a communicator C with a short response time for telephone response or a communicator C with a large number of telephone responses to customers within a predetermined time is considered to have high skills in telephone response. And since the skills of telephone response are evaluated based on the response time for each telephone response of each communicator C and the number of telephone responses to customers within a predetermined time, it is possible to evaluate the skills appropriately. Furthermore, by appropriately evaluating and selecting communicators C with high skills, it is possible to generate an appropriate talk script.

[0061] Also, since a talk script is generated for each title indicating the subject matter of the inquiry from the customer, it is possible to optimize the response to the customer according to the title.

[0062] In addition, since the talk script is output using the learned model 342 for the machine-learned script, it is possible to generate a more appropriate talk script and optimize the response to the customer.

[0063] Moreover, since the talk script is generated periodically, it is possible to generate an appropriate and up-to-date talk script corresponding to the daily telephone response of each communicator C and continuously optimize the response to the customer.

[0064] As described above, the embodiments of the present invention have been described in detail. However, the specific configuration is not limited to this embodiment, and even if there are design changes and the like within the scope not departing from the gist of the present invention, they are included in the present invention. For example, in the skill evaluation task 35 of the above embodiment, the skill is mainly evaluated based on the response time and the number of telephone responses. However, the skill may also be evaluated based on the analysis of the response content and the satisfaction of the customer (the degree to which the customer is satisfied and convinced by the reply). In addition, although the case of supporting each communicator C in the customer center operated by the power company has been described, it is also applicable to the case of supporting other communicators C.

[0065] On the other hand, by installing the following customer response support program on a general-purpose computer, the above-described customer response support system 1 and support server 3 may be configured.

[0066] That is, the computer is made to function as skill evaluation means (skill evaluation task 35) for evaluating the telephone response skills of each communicator C who conducts telephone correspondence with customers based on the response time for each telephone correspondence of each communicator C and the number of telephone correspondences to customers within a predetermined time, and script generation means (script generation task 36) for generating a telephone response talk script based on the response content for each telephone correspondence of the communicator C with high skills evaluated by the skill evaluation means. The skill evaluation means evaluates skills for each title indicating the subject matter of inquiries from customers, the script generation means generates a talk script for each title, and further, the script generation means uses a learning model 342 for a script that has been machine-learned based on past performance data so that a talk script is output when the response content of the communicator C with high skills is input.

Explanation of Signs

[0067] 1 Customer support system 2 Operation terminal 3 Support server 342 Learning model for script 343 Performance database for script 35 Skill evaluation task (skill evaluation means) 36 Script generation task (script generation means) C Communicator

Claims

1. An operation terminal used by each communicator who conducts telephone correspondence with customers is communicably connected to a support server, each of the operation terminals transmits the response time and response content for each telephone correspondence to the support server, to the support server, skill evaluation means for evaluating the telephone correspondence skills of each communicator based on the response time; and script generation means for generating a telephone correspondence script based on the response content of the communicator with high skills evaluated by the skill evaluation means. A customer response support system characterized by the above.

2. The skill evaluation means periodically evaluates the telephone correspondence skills of each communicator, and the script generation means periodically generates the script. The customer response support system according to claim 1, characterized by the above.

3. The skill evaluation means evaluates the skill for each title indicating the subject matter of the inquiry from the customer, and the script generation means generates the script for each title. The customer response support system according to claim 1, characterized by the above.

4. The skill evaluation means evaluates the skill based on the number of telephone correspondences to the customer within a predetermined time for each communicator. The customer response support system according to claim 1, characterized by the above.

5. When the response content of the communicator with high skills is input, the script generation means uses a learning model for the script learned based on past performance data so that the telephone correspondence script is output. The customer response support system according to claim 1, characterized by the above.

6. A computer, skill evaluation means for evaluating the telephone correspondence skills of each communicator based on the response time for each telephone correspondence of each communicator who conducts telephone correspondence with customers; and script generation means for generating a telephone correspondence script based on the response content for each telephone correspondence of the communicator with high skills evaluated by the skill evaluation means. A customer response support program characterized by causing the computer to function as such.

7. The skill evaluation means evaluates the skill for each title indicating the subject matter of the inquiry from the customer, and the script generation means generates the script for each title. The customer response support program according to claim 6, characterized by the above.

8. The skill evaluation means evaluates the skill based on the number of telephone responses to the customer within a predetermined time by each communicator. The customer response support program according to claim 6, characterized in that.

9. When the response content of a communicator with high skill is input, the script generation means uses a learning model for a script that is machine-learned based on past performance data so as to output the talk script. The customer response support program according to claim 6, characterized in that.

Images