JP6438674B2 - Response system, response method, and computer program - Google Patents

Description

  The present invention relates to a technique for automatically recognizing and responding to voice.

Conventionally, in a call center, a user selects a desired service menu by performing push dialing according to voice guidance.
On the other hand, and a technique for recognizing emotion of the speaker by analyzing the audio data, voice recognition techniques, such as converting the audio data into text data has been put into practical use (for example, see Patent Documents 1 and 2 .)
In recent years, there has been proposed a technique for assigning a veteran operator to a user whose emotion is angry by monitoring the emotion of the user during a call using voice recognition technology. In addition, a technique has been proposed in which information related to a call content is immediately provided to an operator by converting the call content into text data using a voice recognition technology.

JP 2012-113542 A JP 2012-47875 A

  However, in the conventional call center, even if the user's emotion is recognized by using the voice recognition technology as described above, it is an operator that takes into account the user's emotion, and such response is possible. There was a problem that human resources had to be secured.

  In view of the above circumstances, an object of the present invention is to provide a technique that can perform a response according to a user's emotion with less human resources in providing a call center service.

  One aspect of the present invention is an emotion recognition unit that acquires user emotion information representing a user's emotion based on voice data transmitted from a call device of a call center user, and the voice data is converted into text data. A response system comprising: a voice conversion unit for conversion; and a response unit that selects a message to be responded to a user from message candidates based on the text data and the user emotion information.

  One aspect of the present invention is the above-described response system, in which the response unit stores the response type information indicating the type of response to the user and text data in association with the text data. Based on the acquired response type information and the user emotion information, a message that responds to the user is selected from message candidates.

  One aspect of the present invention is the above-described response system, in which the response unit stores response content information representing the content of processing to be executed in response to a user and response type information in association with each other. The response content information is acquired based on the response type information, and a message that responds to the user is selected from message candidates based on the acquired response content information and the user emotion information. .

  One aspect of the present invention is the above-described response system, in which a message in which a call center operator responds to a user in the past is a message in which a user's emotion is changed is added to the message candidate or the message candidate An additional unit for deleting from the message candidates is further provided.

  One aspect of the present invention is the response system described above, wherein the additional unit can change a user's emotion into a better emotion among messages that the call center operator has responded to the user in the past. Added to the message candidates.

  One aspect of the present invention is the response system described above, wherein the response unit provides the user emotion information to an operator of a call center.

  One aspect of the present invention is an emotion recognition step of acquiring user emotion information representing the user's emotion based on voice data transmitted from a call device of a call center user, and the voice data is converted into text data. A response method comprising: a voice conversion step for conversion; and a response step for selecting a message to be responded to a user from among message candidates based on the text data and the user emotion information.

  One aspect of the present invention is an emotion recognition step of acquiring user emotion information representing the user's emotion based on voice data transmitted from a call device of a call center user, and the voice data is converted into text data. A computer program for causing a computer to execute a voice conversion step for conversion, and a response step for selecting a message to respond to a user from message candidates based on the text data and the user emotion information. .

  According to the present invention, in providing a call center service, it is possible to perform a response corresponding to a user's emotion with fewer human resources.

1 is a system configuration diagram showing a system configuration of an automatic voice response system 1 according to a first embodiment. The figure which shows the specific example of the audio | voice information table 310. FIG. The figure which shows the 1st specific example of the text information table 510. FIG. The figure which shows the specific example of the flow management table 710. FIG. The figure which shows the specific example of the action management table 720. FIG. The figure which shows the specific example of the response management table 730. The functional block diagram which shows the function structure of the automatic response apparatus 600. FIG. The flowchart which shows the flow of a process of the voice automatic response in the voice automatic response system. The figure which shows the specific example of the telephone call between the audio | voice automatic response system 1 and a user. The system block diagram which shows the system configuration | structure of the audio | voice automatic response system 1a of 2nd Embodiment. The functional block diagram which shows the function structure of the automatic response apparatus 600a. The figure which shows the specific example of the response log | history table 740. The flowchart which shows the flow of a template production | generation process in the audio | voice automatic response system 1a of 2nd Embodiment. The figure which shows the 2nd specific example of the text information table 510. FIG.

[First Embodiment]
FIG. 1 is a system configuration diagram showing a system configuration of an automatic voice response system 1 according to the first embodiment.
The telephones 2-1 to 2-3 are telephones used by the user to make calls to the call centers 3-1 to 3-3. The telephones 2-1 to 2-3 may be telephones that make a call by wireless communication such as a mobile phone or a smartphone, or may be telephones that make a call by wired communication such as a fixed telephone. The telephones 2-1 to 2-3 are connected to the network 4 and communicate with the automatic voice response system 1 via the network 4. In the following description, the telephone sets 2-1 to 2-3 are described as the telephone set 2 unless otherwise distinguished for the sake of simplicity. The number of telephones 2 may be different from that shown in FIG.

  The call centers 3-1 to 3-3 are offices and facilities provided for the business operators to specialize in the business of handling calls to customers. The call centers 3-1 to 3-3 may be call centers that the same business operator has at a plurality of locations, or may be call centers that are operated by different business operators. In the following description, in order to simplify the description, the call centers 3-1 to 3-3 are described as the call center 3 unless otherwise distinguished. An operator is arranged in the call center 3. The operator is a person who answers the user by telephone. Note that the number of call centers 3 may be different from that in FIG.

  The network 4 is a network for connecting the telephone 2 to the automatic voice response system 1. The network 4 may be composed of a plurality of types of networks depending on the communication method of the telephone 2. For example, the network 4 may be a mobile phone network, a fixed telephone line network, or a VoIP (Voice over Internet Protocol) network.

  The call center management devices 5-1 to 5-3 are devices that manage and control the call centers 3-1 to 3-3, respectively. For example, the call center management devices 5-1 to 5-3 are connected to an external system such as a customer management system and provide information to an operator of the call center 3. Further, for example, the call center management devices 5-1 to 5-3 relay voice communication to an operator of the call center 3 in response to a request from the voice automatic response system 1. In the following description, the call center management devices 5-1 to 5-3 are referred to as the call center management device 5 unless otherwise distinguished for the sake of simplicity. The number of call center management devices 5 may be different from that in FIG. 1 depending on the number of call centers 3. Further, the call center management device 5 may manage and control a plurality of call centers 3.

  The automatic voice response system 1 is between each call center 3 and each telephone 2 of the user, and makes a primary response to the user's call transmission. That is, the automatic voice response system 1 is a system that provides the call center 3 with a function of performing automatic voice response to a user's call. The automatic voice response system 1 transfers the user's call to the call center 3 as necessary.

The automatic voice response system 1 includes a relay device 100, an emotion recognition device 200, a voice DB 300, a voice conversion device 400, a text DB 500, an automatic response device 600, and a response DB 700.
The relay device 100 receives the audio signal transmitted from the telephone set 2. The relay device 100 separates the user's voice and the operator's voice included in the received voice signal. The relay device 100 relays voice data acquired by separating the voice signal to the emotion recognition device 200. Also, the relay device 100 transfers the user's call to the operator in response to a request from the automatic response device 600. The relay device 100 relays the voice of the operator and voice data generated by the automatic response function of the automatic voice response system 1 to the telephone set 2.

  The emotion recognition device 200 receives the voice data relayed by the relay device 100. The emotion recognition device 200 divides the received voice data for each utterance. One utterance represents a utterance made by one person before the other person starts speaking. The emotion recognition device 200 recognizes the user's emotion when the voice included in the voice data is emitted for each divided voice data. Any existing technique may be used as a method for recognizing a speaker's emotion from voice data. Emotion recognition apparatus 200 registers voice data and information representing user emotion recognized from the voice data (hereinafter referred to as “user emotion information”) in voice DB 300 in association with each other.

  The voice DB 300 is a database that stores voice data. The voice DB 300 holds a voice information table 310 for storing voice data. The voice DB 300 stores voice data and user emotion information in association with each other. The voice DB 300 is connected to the emotion recognition device 200 and the voice conversion device 400. The voice information table 310 is registered by the emotion recognition device 200.

FIG. 2 is a diagram showing a specific example of the audio information table 310.
The voice information table 310 has a voice information record for each voice ID. The voice information record has values of voice ID, call ID, call classification, voice information, and emotion information. The voice ID is identification information of voice data stored in the voice information table 310. The call ID is call identification information indicated by the voice ID. The call category is identification information that represents the subject that has produced the voice of the voice data indicated by the voice ID. Specifically, the call classification is information indicating whether the call is a call by a user, a call by an operator, or an automatic response by the voice automatic response system 1. The voice information is actual data of the voice data indicated by the voice ID. Emotion information is user emotion information recognized based on voice data indicated by a voice ID. The voice information record is registered in the voice information table 310 by the emotion recognition device 200.

Returning to the description of FIG.
The voice conversion device 400 refers to the voice information table 310 of the voice DB 300 and acquires voice data. The voice conversion device 400 converts voice included in the acquired voice data into text data. Any existing technique may be used as a method for converting voice data into text data.

  The voice conversion device 400 extracts keywords related to the service provided by the call center 3 from the acquired text data. Any existing technique may be used as a method for extracting a keyword from text data. For example, the voice conversion device 400 may hold information relating to the business or type of business of the business operator who operates the call center 3 in advance as dictionary information, and may extract words that are strongly related to the dictionary information as keywords. Further, for example, the voice conversion device 400 may be configured to learn dictionary information by performing statistical processing on the accumulated text data. The speech conversion apparatus 400 registers the identification information of the speech data and the keyword acquired from the speech data in the text DB 500 in association with each other.

  The text DB 500 is a database that accumulates text data acquired from voice data. The text DB 500 holds a text information table 510 for storing text data for each voice data. The text DB 500 stores voice data and text data acquired from the voice data in the text information table 510 in association with each other. The text DB 500 is connected to the voice conversion device 400 and the automatic response device 600. The text information table 510 is registered by the voice conversion device 400.

FIG. 3 is a diagram showing a first specific example of the text information table 510.
The text information table 510 has a text information record for each voice ID. The text information record has each value of voice ID, call ID, call classification, text information, keyword, and emotion information. The voice ID, the call ID, the call classification, and the emotion information are the same as the voice ID included in the voice information record of the voice information table 310. The text information is text data obtained by converting the voice data indicated by the voice ID. The keyword represents a keyword extracted by the speech conversion device 400 from text data indicated by the text information. The text information record is registered in the text information table 510 by the voice conversion device 400.

Returning to the description of FIG.
The automatic response device 600 selects a message to respond to the user based on the keyword and user emotion information acquired from the voice data. Specifically, the automatic response device 600 acquires a message that responds to the user from the response message template stored in the response DB 700 based on the keyword and the user emotion information. The automatic response device 600 converts the acquired response message into voice data. The automatic response device 600 transmits the voice data generated by the conversion of the response message to the user's telephone 2. Note that the automatic response device 600 may be configured to acquire voice data converted in advance, instead of acquiring text data indicating a response message.
Note that the automatic answering device 600 requests the relay device 100 to transfer the user's call to the operator of the call center 3 that is the destination when the text data indicating the message used for the response to the user cannot be acquired.

  The response DB 700 is a database that holds information necessary for the automatic response device 600 to respond automatically. For example, the response DB 700 holds a flow management table 710, an action management table 720, and a response management table 730. The response DB 700 is connected to the automatic response device 600. The response management table 730 is updated by the automatic response device 600.

FIG. 4 is a diagram illustrating a specific example of the flow management table 710.
The flow management table 710 is a table in which services provided to the user by the automatic voice response system 1 are defined. The flow management table 710 has a flow management record for each flow ID. The flow management record has values of flow ID, related information, and service information. The flow ID is service identification information and corresponds to a service provided to the user by the automatic voice response system 1. The related information has text data related to the service indicated by the flow ID as a value. The service information represents the content of the service indicated by the flow ID. The flow management record is registered in advance in the flow management table 710 by the administrator of the automatic voice response system 1.

FIG. 5 is a diagram illustrating a specific example of the action management table 720.
The action management table 720 is a table in which a response flow to be performed by the automatic response device 600 is defined for each service provided to the user. Below, each response defined for each service is called an action. The action management table 720 has an action management record for each action ID. The action management record has values of flow ID, action ID, action information, and execution flag. The flow ID is the same as the flow ID that the flow management record of the flow management table 710 has. The action ID is identification information of an action executed by the automatic response device 600 in the selected service. The action information represents the content of the action indicated by the action ID. The execution flag is information indicating an action to be executed next by the automatic response device 600 in the selected service.

  The action management record is registered in advance in the action management table 720 as a master by the administrator of the automatic voice response system 1. The action management table 720 is generated with the execution flag value initialized for each service provided to the user. For example, the example of FIG. 5 represents the action management table 720 generated for the service indicated by the flow ID value “F001” in FIG.

FIG. 6 is a diagram illustrating a specific example of the response management table 730.
The response management table 730 stores a message template (an example of a message candidate) that responds for each action ID registered in the action management table 720. The response management table 730 has a response history record for each response ID. The response history record has each value of response ID, action ID, action information, response information, and emotion information. The response ID is identification information of a response history record. The action ID is the same as the action ID that the action management record of the action management table 720 has. The action information is the same as the action information included in the action management record of the action management table 720. The response information represents a response message indicated by the response ID. That is, the response information is text data serving as a template for generating voice data used for the action indicated by the action ID. The emotion information is the same as the emotion information that the voice information record of the voice information table 310 and the text information record of the text information table 510 have. The response history record is registered in advance in the response management table 730 by the administrator of the automatic voice response system 1.

FIG. 7 is a functional block diagram showing a functional configuration of the automatic response device 600.
The automatic response device 600 includes a CPU (Central Processing Unit), a memory, an auxiliary storage device, and the like connected by a bus, and executes an automatic response program. The automatic response device 600 functions as a device including a communication unit 610, a response determination unit 620, and a response generation unit 630 (an example of a response unit) by executing an automatic response program.
All or some of the functions of the automatic response device 600 may be realized using hardware such as an application specific integrated circuit (ASIC), a programmable logic device (PLD), or a field programmable gate array (FPGA). . The automatic response program may be recorded on a computer-readable recording medium. The computer-readable recording medium is, for example, a portable medium such as a flexible disk, a magneto-optical disk, a ROM, a CD-ROM, or a storage device such as a hard disk built in the computer system. The automatic response program may be transmitted via a telecommunication line.

  The communication unit 610 is configured using a communication interface such as a LAN (Local Area Network). The communication unit 610 communicates with the relay device 100, the emotion recognition device 200, the voice conversion device 400, and the call center management device 5.

  The response determination unit 620 determines the type of response to be answered to the user based on the keyword acquired from the voice data. For example, the response determination unit 620 determines a service desired by the user. The response determination unit 620 outputs information indicating the type of response determined based on the keyword (hereinafter referred to as “response type information”) to the response generation unit 630.

  Based on the response type information output from the response determination unit 620, the response generation unit 630 acquires information indicating an action to respond (hereinafter referred to as “response content information”). The response generation unit 630 acquires a response message used for a response to the user from the response DB 700 based on the response content information and the user emotion information. For example, the response generation unit 630 selects an action to respond from the action management table 720 based on the response type information output from the response determination unit 620. The response determination unit 620 acquires the identification information of the selected action as response content information. The response generation unit 630 acquires text data of a message that responds to the user from the response management table 730 based on the response content information and the user emotion information. The response generation unit 630 converts the acquired text data into voice data. The response generation unit 630 transmits voice data to the user's telephone 2.

FIG. 8 is a flowchart showing the flow of automatic voice response processing in the automatic voice response system 1.
First, a user operates the telephone 2 to make a call to the call center 3. Relay device 100 receives a call from telephone 2. The relay device 100 acquires voice data from the incoming call. The relay device 100 relays voice data to the emotion recognition device 200. The emotion recognition apparatus 200 acquires the relayed voice data (step S101). The emotion recognition device 200 generates a voice ID that is identification information of the voice data. In addition, the emotion recognition device 200 generates a call ID that is identification information of the call of the relayed voice data.

  For example, the call ID may be generated based on information such as the telephone number and time of the caller telephone 2. The emotion recognition device 200 recognizes the user's emotion when the voice included in the voice data is emitted from the voice data (step S102). Emotion recognition device 200 outputs voice ID, call ID, voice data, and user emotion information to voice DB 300. The voice DB 300 generates a voice information record having information output from the emotion recognition device 200 as a value. The voice DB 300 registers the generated voice information record in the voice information table 310. The emotion recognition device 200 notifies the speech conversion device 400 that a new speech information record has been registered.

  When receiving the notification that a new voice information record is registered by the emotion recognition apparatus 200, the voice conversion device 400 refers to the voice information table 310 and acquires each value of the newly registered voice information record. The voice conversion device 400 converts voice included in the acquired voice data into text data. The voice conversion device 400 extracts keywords from the acquired text data. Then, the voice conversion device 400 outputs the value of the voice ID acquired from the voice information record, the value of emotion information acquired from the voice information record, and the text data and keywords acquired from the voice data to the text DB 500. The text DB 500 generates a text information record having information output from the speech conversion device 400 as a value. The text DB 500 registers the generated text information record in the text information table 510. The voice conversion device 400 notifies the automatic response device 600 that a new text information record has been registered.

  When receiving a notification that a new text information record has been registered by the voice conversion device 400, the response determination unit 620 of the automatic response device 600 refers to the text information table 510, and each value of the newly registered text information record. To get. The response determination unit 620 refers to the flow management table 710 and selects a flow management record that has text data that is strongly related to the acquired text information value as the related information value. The response determination unit 620 acquires the value of the flow ID from the selected flow management record. The response determination unit 620 outputs the acquired flow ID value to the response generation unit 630 as response type information. As described above, the response determination unit 620 determines the service desired by the user based on the keyword acquired from the voice data (step S103).

  Next, the response generation unit 630 selects the action management table 720 corresponding to the service indicated by the response type information output from the response determination unit 620. The response generation unit 630 acquires the value of the execution flag from the selected action management table 720. Based on the value of the execution flag, the response generation unit 630 determines an action to be executed next in the selected service (step S104).

  Next, the response generation unit 630 refers to the response management table 730 and selects a response management record having response content information indicating the action determined in step S104 as an action ID value. Furthermore, the response generation unit 630 selects, from the selected response management records, a response management record having the user emotion information determined in step S102 as the emotion information value. The response generation unit 630 acquires the value of response information from the selected response management record. The response generation unit 630 converts the acquired response information value into audio data. The response generation unit 630 transmits the voice data to the telephone 2 of the caller user. As described above, the response generation unit 630 responds to the user with a response message selected based on the user's emotion (step S105).

  Next, the response generation unit 630 determines whether or not all actions have been performed in the action flow in the selected service (step S106). When there is an action that has not been executed (step S106: NO), the response determination unit 620 moves the process to step S101 and continues the call with the user. On the other hand, when all responses have been made (step S106—YES), the response determination unit 620 ends the call with the user.

FIG. 9 is a diagram illustrating a specific example of a call between the automatic voice response system 1 and a user.
In FIG. 9, the calls shown in the calls 1, 3, 5, 7 and the call 9 represent calls made by the user to the automatic voice response system 1. Calls shown as calls 2, 4, 6 and 8 represent responses of the automatic voice response system 1 to the user. The call shown in FIG. 9 is exchanged between the user and the automatic voice response system 1 in the order of the call 1 to the call 9. Note that the description in parentheses in the user's call represents the user's feelings when the call is made.

  When the user's call has the contents shown in the example of FIG. 9, for example, the contents shown in FIG. 2 are registered in the voice information table 310 by the emotion recognition device 200. The voice conversion device 400 converts voice data held as voice information of the voice information record into text data. The voice conversion device 400 extracts keywords from text data obtained by converting voice data. As a result, for example, the contents shown in FIG. 3 are registered in the text information table 510 by the voice conversion device 400.

  When the flow management table 710 is registered with the contents shown in FIG. 4, the response determination unit 620 of the automatic response device 600 uses, for example, the voice data having “V001” as the voice ID value. Recognizes that they want to “restart credit card usage”. Further, the automatic response device 600 refers to the text information table 510, and the emotion of the user in the call is “angry” based on the emotion information value of the text information record having “V001” as the voice ID value. Recognize that. When the action management table 720 is registered with the content shown in FIG. 5, the automatic response device 600 determines that the action information value “customer information value” of the action management record having “A001” as the action ID value. Recognize that “confirm” is the action to be performed on the selected service.

  When the response management table 730 is registered with the contents shown in FIG. 6, the automatic response device 600 has a response history having “A001” as the action ID value and “anger” as the emotion information value. It is recognized that the value of the response information of the response history record having the record, that is, “R001” as the response ID value is a message responding to the user. At this time, assuming that the emotion of the user is “normal”, the value of the response information in the response history record having “R002” as the response ID value is selected as a message to respond to the user. That is, even if the message has the same content, a message suitable for the user's emotion at that time is selected. Thus, the automatic response device 600 selects a message to respond to the user according to the user's emotion.

In the automatic voice response system 1 according to the first embodiment configured as described above, the automatic response device 600 selects a message to respond to the user based on information representing the emotion of the user acquired from the voice data. . The automatic voice response system 1 including the automatic response device 600 responds with a more polite wording to a user having an emotion of “anger” and responds with a shorter word to a user having an emotion of “impression”, for example. It becomes possible to do.
Therefore, the voice automatic response system 1 according to the embodiment enables the service provider of the call center 3 to perform a response corresponding to the emotion of the user with less human resources in providing the call center service.

  In addition, the automatic response device 600 according to the first embodiment recognizes a service desired by a user based on a keyword extracted from voice data, and generates a response message corresponding to the service. Therefore, the automatic voice response system 1 including the automatic response device 600 can reduce the burden of the user's input operation for information transmission, and can smoothly exchange information with the user.

  By providing the automatic answering system 1 of the first embodiment with the automatic answering device 600 described above, the service provider of the call center 3 provides the call center service to the user even in the absence of an operator such as at night. It becomes possible. Even in a call center capable of handling nighttime, there are cases where the number of operators arranged at night is less than in daytime. As a result, when the call center is used at night, there is a possibility that the waiting time until the user receives the service becomes long. Further, even in daytime use, in a situation where the call center is congested, the user has to make a call again after a while and wait until the congestion is resolved. When the automatic voice response system 1 according to the first embodiment includes the automatic response device 600 described above, the operator of the call center 3 can provide a service to the user with a shorter waiting time.

  Further, the automatic voice response system 1 of the first embodiment makes use of the waiting time until an operator can respond by performing an automatic response according to the embodiment in a situation where the call center is congested. It may be configured to collect. The information collected in this way is taken over by an operator who can respond, so that in the automatic voice response system 1 of the first embodiment, it is possible to shorten the operator's response time.

<Modification>
When the actual data of the audio data is stored other than the audio information table 310, the value of the audio information registered in the audio information table 310 may be information indicating the storage location of the actual data of the audio data.
In the automatic response device 600, the response determination unit 620 and the response generation unit 630 may be configured as one functional unit having both functions.

  The response type information output by the response determination unit 620 may be used for purposes other than acquiring the next action (response content information) in response to the user. In the present embodiment, the response type information represents the type of service provided to the user. Therefore, for example, the automatic response device 600 may be configured to acquire the usage count for each service by providing a counter function. By configuring the automatic answering apparatus 600 in this way, the service provider of the call center 3 can analyze the statistics of the number of times of using the service and connect it to service improvement activities. In addition, the automatic answering device 600 may be configured to acquire the number of times the service is used for each user by using the call ID. By configuring the automatic answering apparatus 600 in this way, the service provider of the call center 3 can analyze the statistics of the number of times the service is used for each user and monitor, for example, unauthorized use.

  In step S105, when the response message used for the response to the user cannot be acquired, the automatic response device 600 transmits the voice data including the voice that the operator has responded to the other user in the past. A response may be made. For example, in this case, the automatic response device 600 may select a text information record having a text information value that is strongly related to the keyword value of the newly registered text information record. The automatic response device 600 acquires the value of the text information from the selected text information record. The automatic response device 600 converts the value of the acquired text information into voice data. The automatic response device 600 transmits the converted voice data to the telephone 2 of the user.

  When personal data such as name, address, and credit card number is included in voice data including voices that the operator has responded to other users in the past, the automatic response device 600 excludes such personal information. The voice data may be transmitted in place of the personal information of the user to be responded to. In addition, when personal information is included in voice data including voices that the operator has responded to other users in the past, the automatic answering device 600 does not automatically answer but transfers the user's call to the operator of the call center 3. To the relay device 100.

  When the user's emotion at the start of the call is in an “anger” state or when the user's emotion deteriorates due to the automatic response, the automatic answering device 600 does not perform an automatic response and does not perform the automatic response. It may be configured to request the relay apparatus 100 to transfer to the third operator.

  The automatic answering device 600 may provide the operator with information exchanged with the user between the start of the call and the transfer when the user's call is transferred to the operator. With this function, for example, the operator does not need to ask the user again about the information heard from the user, so that the service efficiency can be improved and the user's feelings can be prevented from deteriorating. it can.

  The automatic voice response system 1 may use the response message selected by the automatic response device 600 for purposes other than the automatic response. For example, the automatic response device 600 may be configured to provide the selected response message to the operator as an example of a response. By configuring the automatic response device 600 in this way, for example, even an inexperienced operator can perform a more appropriate response. Furthermore, the automatic response device 600 may provide the user emotion information used for selecting the response message to the operator. For example, by providing user emotion information, the operator can know the emotions of users in other countries that are difficult to grasp.

  The automatic voice response system 1 may be configured to determine a user's language type based on the voice data and select a response message according to the user's language. In this case, for example, the emotion recognition device 200 determines the language type of the user. The voice DB 300 and the text DB 500 store voice data identification information and a language type in association with each other. The response DB 700 stores the identification information of the response message and the language type in association with each other. The automatic response device 600 acquires a response message from the response DB 700 based on the response content information, emotion information, and language type.

  Note that the response message acquired in accordance with the language type of the user may be used for an automatic response to the user, or may be provided to the operator as an example of the response as described above. By providing the operator with a response message corresponding to the language type, it is possible to improve the quality of response of the operator who corresponds in a language other than the native language.

  The automatic voice response system 1 may have a function of translating a user's call into the operator's native language. In this case, for example, the voice conversion device 400 translates the user's call converted into text data into the native language of the operator. Text DB 500 stores identification information of speech data and translated text data in association with each other. The automatic answering device 600 provides the operator's call translated into the native language to the operator. By translating the user's call into the native language and providing it to the operator, the operator corresponding to the language other than the native language can understand the content of the user's call more accurately.

[Second Embodiment]
FIG. 10 is a system configuration diagram showing a system configuration of the automatic voice response system 1a of the second embodiment.
The automatic voice response system 1a differs from the automatic voice response system 1 of the first embodiment in that an automatic response device 600a is provided instead of the automatic response device 600 and a response DB 700a is provided instead of the response DB 700. Other configurations of the automatic voice response system 1 a are the same as those of the automatic voice response system 1. In FIG. 10, the same components as those in the automatic voice response system 1 are denoted by the same reference numerals as those in FIG.

FIG. 11 is a functional block diagram showing a functional configuration of the automatic response device 600a.
The automatic response device 600a is different from the automatic response device 600 of the first embodiment in that a response generation unit 630a is provided instead of the response generation unit 630 and a template generation unit 640 is further provided. The other configuration of the automatic response device 600a is the same as that of the automatic response device 600. In FIG. 11, functional units having the same functions as those of the automatic response apparatus 600 are denoted by the same reference numerals as those in FIG.

  In addition to the function of the response generation unit 630, the response generation unit 630a has a function of registering a history of responses to the user in the response DB 700a. With this function, the response generation unit 630a registers the identification information of the voice data of the operator and the response content information indicating the action executed by the operator in the response DB 700a.

  The template generation unit 640 refers to the history of past responses by the operator, and adds a high-quality response to a response message template used for an automatic response. For example, the template generation unit 640 extracts a response that can change the user's emotion to a better emotion as a high-quality response. For example, the template generation unit 640 may extract a response that can change the emotion of the user who was in an “anger” state at the start of the call to “joy” at the end of the call. The template generation unit 640 adds the extracted response as a new template to the response management table 730.

The response DB 700a is different from the response DB 700 of the first embodiment in that it further includes a response history table 740.
FIG. 12 is a diagram illustrating a specific example of the response history table 740.
The response history table 740 is a table that holds a history of actions executed in response to a user. The response history table 740 has a response history record for each call ID. The response history record has each value of call ID, voice ID, and action ID. The call ID and the voice ID are the same as the call ID and the voice ID included in the voice information record and the text information record. The action ID is the same as the action ID included in the action management record. The response history record is registered in the response history table 740 by the response generation unit 630a when an action of response to the user is executed.

FIG. 13 is a flowchart showing a flow of template generation processing in the automatic voice response system 1a of the second embodiment.
First, when the user makes a call with the operator, the emotion recognition apparatus 200 performs the same processing as step S102 and recognizes the user's emotion from the voice data. In addition, the voice conversion device 400 performs the same processing as step S103, and converts voice data into text data. The response determination unit 620 of the automatic response device 600 performs the same process as in step S103, and determines the service desired by the user. The response determination unit 620 outputs the identification information of the service determined as the service desired by the user to the response generation unit 630a as response type information. The response generation unit 630a refers to the action management table 720 corresponding to the response type information, and determines an action corresponding to the operator's response. The response generation unit 630a generates a response history record having the identification information of the target call, the identification information of the voice data, and the response content information indicating the response action as values. The response generation unit 630a registers the generated response history record in the response history table 740.

The template generation unit 640 refers to the text information table 510 and selects the response history table 740 corresponding to the call ID in which the emotion information at the end of the call has changed to a better emotion than the emotion information at the start of the call in the past call. (Step S201).
The template generation unit 640 acquires the value of the voice ID of the response history record registered in the selected response history table 740. The template generation unit 640 refers to the text information table 510 and selects a text information record having the voice ID value acquired from the response history record as the voice ID value. The template generation unit 640 acquires the value of text information and the value of emotion information from the selected text information record. As described above, the template generation unit 640 acquires a message that the operator has responded to based on the response history of the operator (step S202).

  The template generation unit 640 outputs the text information and emotion information values acquired from the text information record to the response DB 700a. The response DB 700a generates a response management record having the value output from the template generation unit 640. The response DB 700a registers the generated response management record in the response management table 730 (step S203).

FIG. 14 is a diagram showing a second specific example of the text information table 510.
FIG. 14 shows that a call indicated by the text information table 510 in the example of FIG. 3 is performed between the user and the operator.
For example, the template generation unit 640 is shown in the example of FIG. 12 as a table in which the emotion information at the end of the call becomes a better emotion than the emotion information at the start of the call in the response history table 740 in the past call. The content response history table 740 is selected. The template generation unit 640 acquires each value of the response history record from the selected response history table 740.

  The template generation unit 640 refers to the text information table 510 and selects a text information record having the voice ID and call ID values acquired from the response history record as the voice ID and call ID values. The template generation unit 640 acquires the value of text information and the value of emotion information from the selected text information record. The template generation unit 640 outputs the text information acquired from the text information record and the value of emotion information acquired from the text information record to the response DB 700a. The response DB 700a generates a response management record having values of response information and emotion information output from the template generation unit 640. The response DB 700a registers the generated response management record in the response management table 730.

  The message in which the text information value of the text information record having “V002”, “V004”, “V006”, and “V008” as the voice ID values is added to the response management table 730 in the example of FIG. Selected as.

  In the voice automatic response system 1a according to the second embodiment configured as described above, the automatic response device 600a generates a response that can make the user's emotion better in the response of the past operator. It adds to the template of the response message which 600a selects. Therefore, the voice automatic response system 1a can autonomously improve the accuracy of the automatic response function.

<Modification>
In step S201, the automatic response device 600a may be configured to select the response history table 740 based on the degree of change in emotion in the past call. The automatic response device 600a may be configured to perform a predetermined weight on the response message included in the selected response history table 740 and add the response message to the template according to the weight. For example, when the selected response message exists in a plurality of response history tables 740, the automatic response device 600a may weight the number of times the user's emotion has changed due to the response message.

  The automatic response device 600a may be configured to delete, from the response management table 730, a response message that has made the user's emotion worse in the past operator's response. Further, if the automatic response device 600a holds such a response as an NG collection and provides it to the operator, the response quality of the operator can be improved. When a plurality of templates corresponding to the same action ID and emotion information are registered in addition of a template response message by the automatic response device 600a, the automatic response device 600a uses the latest template from the selected templates. It may be configured to respond.

  When the message acquired in step S202 includes personal information such as a name, an address, and a credit card number, the automatic response device 600 excludes the personal information or becomes a response target. A message replaced with personal information of the user may be generated.

  The automatic response devices 600 and 600a in the above-described embodiment may be realized by a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory inside a computer system serving as a server or a client in that case may be included and a program held for a certain period of time. Further, the program may be a program for realizing a part of the above-described functions, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system. It may be realized using a programmable logic device such as an FPGA.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes designs and the like that do not depart from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1, 1a ... Voice automatic response system, 2, 2-1 to 2-3 ... Telephone, 3, 3-1 to 3-3 ... Call center, 4 ... Network, 5, 5-1 to 5-3 ... Call center management apparatus , 100 ... Relay device, 200 ... Emotion recognition device, 300 ... Voice DB, 310 ... Voice information table, 400 ... Voice conversion device, 500 ... Text DB, 510 ... Text information table, 600, 600a ... Automatic response device, 610 ... Communication unit, 620... Response judgment unit, 630, 630a ... Response generation unit, 640 ... Template generation unit, 700, 700a ... Response DB, 710 ... Flow management table, 720 ... Action management table, 730 ... Response management table, 740 ... Response history table

Claims (8)

An emotion recognition unit that acquires user emotion information representing the user's emotion based on voice data transmitted from a call device of a call center user;
A voice conversion unit that converts the voice data into text data;
Based on the text data and the user emotion information, a response message for the user is selected from message candidates, a response unit that converts the selected response message into voice data and transmits it to the call device;
Among the response messages that the call center operator responded to past users, an addition unit that adds or deletes response messages that change the past user's emotions to the message candidates;
With
The adding unit weights the response message that has changed the emotion of the past user according to the number of times used in the past response, and adds the response message to the message candidate according to the weight or Delete,
Response system. The response unit is a case in which the voice data transmitted to the call device is second voice data that the call center operator responds to a past user, and the second voice data is personal. When voice indicating information is included, voice data obtained by excluding voice indicating the personal information from the second voice data, or voice data obtained by replacing the voice indicating the personal information with personal information of the current user Send to call device,
The response system according to claim 1. The addition unit adds a message that can change the emotion of the user to a better emotion among the messages that the call center operator has responded to the user in the past, to the message candidate.
The response system according to claim 1 or 2 . The response unit acquires the response type information based on the text data from a storage unit that stores response type information indicating the type of response to the user and text data in association with each other, and acquires the acquired response type Based on the information and the user emotion information, a message that responds to the user is selected from message candidates.
The response system according to any one of claims 1 to 3 . The response unit stores the response content information based on the response type information from the storage unit that stores the response content information indicating the content of the process to be executed in response to the user and the response type information in association with each other. Acquiring and selecting a message to respond to the user from among the message candidates based on the acquired response content information and the user emotion information;
The response system according to any one of claims 1 to 4 . The response unit provides the user emotion information to a call center operator.
The response system according to any one of claims 1 to 5 . An emotion recognition step of acquiring user emotion information representing the user's emotion based on voice data transmitted from a call device of a call center user;
A voice conversion step of converting the voice data into text data;
A response step of selecting a response message for the user from message candidates based on the text data and the user emotion information, converting the selected response message into voice data and transmitting it to the call device;
An addition step of adding or deleting a response message that has changed the emotion of the past user among response messages that the call center operator responded to the past user;
Have
In the adding step, the response message in which the past user's emotion is changed is weighted according to the number of times used in the past response, and the response message is added to the message candidate according to the weight or Delete,
Response method. The computer program for functioning a computer as a response system as described in any one of Claim 1 to 6 .

Images