WO2014069444A1 - Complaint conversation determination device and complaint conversation determination method - Google Patents

Abstract

Provided is a complaint conversation determination device comprising: a data acquisition unit that acquires the audio data for a conversation and hold segment data indicating the start time and end time of a hold segment within said conversation; and a dissatisfaction degree determination unit that performs predetermined audio analysis processing on the audio data for the conversation participant who is put on hold using a hold segment that is indicated by the hold segment data and that is included in the audio data acquired by the data acquisition unit in order to determine the degree of dissatisfaction occurring in the conversation.

Description

Dissatisfied conversation determination device and dissatisfied conversation determination method

The present invention relates to a conversation analysis technique.

An example of a technology for analyzing conversation is a technology for analyzing call data. For example, data of a call performed in a department called a call center or a contact center is analyzed. Hereinafter, such a department that specializes in the business of responding to customer calls such as inquiries, complaints and orders regarding products and services will be referred to as a contact center.

Customer feedback sent to the contact center often reflects customer needs and satisfaction. Properly answering calls from such customers improves customer satisfaction and helps repeat customers. It is important to increase. Therefore, various methods for extracting customer emotions (anger, irritation, discomfort, etc.) and customer needs by analyzing telephone voices from customers have been proposed.

In Patent Document 1 below, emotion is recognized from the voice of a phone call from a customer, and whether or not the voice content is a complaint is determined by whether or not the emotion represents at least one of “anger” and “excitement”. A technique for notifying an appropriate person in charge according to the determination result has been proposed. In Patent Document 2 below, periodic fluctuations in the amplitude envelope are detected from the input speech signal in order to detect the anger and irritation of the speaker without being affected by individual differences, language differences, and regional differences, and the detection thereof. There has been proposed a method for discriminating whether or not the input voice is a strong voice according to the result. In Patent Document 3 below, in order to provide information that meets customer needs, it is determined whether or not a keyword set in advance during a call between a call center operator and the customer is spoken. There has been proposed a method of grasping potential needs of customers and providing guidance information associated with the keywords in advance to the customers. In the following Patent Document 4, the time limit for each call partner is calculated from the information about the caller or the receiver whose call is put on hold, and information about the caller and the receiver, There has been proposed a method for notifying an operator of progress information with respect to a time limit and a psychological state of a communication partner according to a comparison result with the time limit.

JP 2011-009902 A JP 2009-003162 A JP 2009-182432 A JP 2009-1111829 A

In the proposed methods in Patent Documents 1 to 3, since the entire conversation (call) is analyzed, it takes time to determine whether or not each call is a dissatisfied call. In addition, in those proposed methods, when customer dissatisfaction is expressed only in a part of the call, the dissatisfaction may not be appropriately detected because the dissatisfaction is leveled throughout the entire call. For example, a customer who has been gently answering may be dissatisfied as soon as the call is put on hold. In such a case, the dissatisfaction may not be detected by the proposed method.

In the above-mentioned Patent Document 4, it is proposed to determine whether or not the call partner has lost normality while the call is on hold. However, in this proposed method, dissatisfaction due to the waiting time for call holding is only determined, and no detection of dissatisfaction regarding the call itself is assumed.

The present invention has been made in view of such circumstances, and provides a technique for appropriately detecting the degree of dissatisfaction with a conversation. Here, the degree of dissatisfaction of the conversation means the degree of dissatisfaction that a person who participates in the conversation in the conversation (hereinafter referred to as a conversation participant) will feel. It may indicate only whether or not it is felt.

Each aspect of the present invention employs the following configurations in order to solve the above-described problems.

The first aspect relates to a dissatisfied conversation determination device. The dissatisfied conversation determination device according to the first aspect includes a data acquisition unit that acquires hold section data indicating start time and end time of a conversation hold section and voice data of the conversation, and voice data acquired by the data acquisition unit A dissatisfaction level determination unit that determines the dissatisfaction level of the conversation by performing a predetermined voice analysis process on the voice data of the held-side conversation participant in the hold period indicated by the hold period data included in Have.

The second aspect relates to a dissatisfied conversation determination method executed by at least one computer. The dissatisfied conversation determination method according to the second aspect obtains hold section data indicating start time and end time of a hold section of conversation and voice data of the conversation, and uses the hold section data included in the acquired voice data. It includes determining the degree of dissatisfaction of the conversation by performing a predetermined voice analysis process on the voice data of the held conversation participant in the hold section shown.

Another aspect of the present invention may be a program that causes at least one computer to implement each configuration in the first aspect, or a computer-readable recording medium that records such a program. There may be. This recording medium includes a non-transitory tangible medium.

According to each aspect described above, it is possible to provide a technique for appropriately detecting the degree of dissatisfaction with the conversation.

The above-described object and other objects, features, and advantages will be further clarified by a preferred embodiment described below and the following drawings attached thereto.

It is a conceptual diagram which shows the structural example of the contact center system in 1st Embodiment. It is a figure which shows notionally the process structural example of the call analysis server in 1st Embodiment. It is a figure which shows notionally the reservation area and the speech area in a reservation area. It is a flowchart which shows the operation example of the telephone call analysis server in 1st Embodiment. It is a figure which shows notionally the process structural example of the call analysis server in 2nd Embodiment.

Hereinafter, embodiments of the present invention will be described. In addition, each embodiment given below is an illustration, respectively, and this invention is not limited to the structure of each following embodiment.

The dissatisfied conversation determination device according to the present embodiment includes a data acquisition unit that acquires hold section data indicating start time and end time of a conversation hold section, and voice data of the conversation, and voice data acquired by the data acquisition unit A dissatisfaction level determination unit that determines the dissatisfaction level of the conversation by performing a predetermined voice analysis process on the voice data of the held-side conversation participant in the hold period indicated by the hold period data included in Have.

The dissatisfied conversation determination method according to the present embodiment is executed by at least one computer, acquires hold section data indicating the start time and end time of a hold section of the conversation, and voice data of the conversation, and acquires the acquired voice data. It includes determining the degree of dissatisfaction of the conversation by performing a predetermined voice analysis process on the voice data of the held-side conversation participant in the hold section indicated by the hold section data included.

Here, “conversation” means that two or more speakers speak by expressing their intentions by uttering a language. In some conversations, conversation participants can speak directly, such as at bank counters and cash registers at stores, and in remote conversations such as telephone conversations and video conferencing. There may be a form in which the participants talk.

Also, the on-hold section means that the conversation is on hold. Holding a conversation means that at least one of a plurality of conversation participants having a conversation does not participate in the conversation. The remaining at least one conversation participant is described as a held-side conversation participant. In other words, the held-side conversation participant means a conversation participant on the side waiting for the conversation to be held and the hold state to be released. Here, the state in which the conversation participant does not participate in the conversation means that the conversation participant is consciously separated from the conversation, such as a state away from the conversation seat or a state where the call is put on hold. It is illustrated by. Therefore, as an example of the reserved section, when two customers and a bank person in charge are having a conversation at the bank window, the bank person in charge is out of the seat. In this example, the held-side conversation participant is a customer waiting for a bank representative.

The present inventors have obtained the knowledge that the held-side conversation participant says the true intention only while the conversation is on hold. From this knowledge, the utterance of the held-side held conversation participant is not satisfied with the held-side conversation participant. We found that there is a high probability that For example, even a conversation participant who is responding moderately (a held-side conversation participant) may be dissatisfied during the hold. A situation in which dissatisfaction is likely to appear during a hold can occur in all conversations including calls. For example, there may be a scene in which customers talk about dissatisfaction while the person in charge is out of their seats. Therefore, in the present embodiment, the conversation hold section data and the voice data of the conversation are acquired, and predetermined voice analysis processing is performed on the voice data of the held-side conversation participant in the hold section specified by the hold section data. As a result, the degree of dissatisfaction with the conversation is determined.

The conversation voice data acquired in the present embodiment includes voice data of a held conversation participant who is holding a conversation. The predetermined speech analysis processing is, for example, emotion recognition processing as proposed in Patent Documents 1 and 2, speech recognition processing for extracting character data indicating the contents from speech data, and the like. The present embodiment does not limit the predetermined voice analysis process itself for determining the degree of dissatisfaction. Further, the degree of dissatisfaction of the call determined in the present embodiment may indicate only whether or not the held-side conversation participant expressed dissatisfaction, or the degree of dissatisfaction may be indicated by a numerical value.

Thus, according to the present embodiment, focusing on the hold during which dissatisfaction is likely to be expressed, since the voice data of the held-side held conversation participant is a target for analysis processing for dissatisfaction determination, Compared with the technique of analyzing the entire conversation, the dissatisfaction determination of each conversation can be executed at high speed. Furthermore, according to the present embodiment, since voice data in a section where dissatisfaction is likely to be expressed is targeted for analysis processing for determining dissatisfaction, it is possible to prevent dissatisfaction leveling throughout the entire conversation. As a result, it is possible to appropriately determine the degree of dissatisfaction with the conversation.

Hereinafter, further details of the above-described embodiment will be described. Below, 1st Embodiment and 2nd Embodiment are illustrated as detailed embodiment. Each of the following embodiments is an example when the above-mentioned unsatisfactory conversation determination device and unsatisfactory conversation determination method are applied to a contact center system. Note that the above-mentioned unsatisfactory conversation determination device and unsatisfactory conversation determination method are not limited to application to a contact center system that handles call data, and can be applied to various aspects of handling conversation data. For example, they can also be applied to in-house call management systems other than contact centers, and personal terminals such as PCs (Personal Computers), fixed telephones, mobile phones, tablet terminals, smartphones, etc. . Furthermore, as conversation data, for example, data indicating conversation between a person in charge and a customer at a bank counter or a store cash register can be exemplified. Hereinafter, the term “call” refers to a call from when a caller has a caller to a caller until the call is disconnected. The on-hold section means a section during a call where the call is put on hold.

[First Embodiment]
〔System configuration〕
FIG. 1 is a conceptual diagram showing a configuration example of a contact center system 1 in the first embodiment. The contact center system 1 in the first embodiment includes an exchange (PBX) 5, a plurality of operator telephones 6, a plurality of operator terminals 7, a file server 9, a call analysis server 10, and the like. The call analysis server 10 corresponds to the unsatisfactory conversation determination device in the above-described embodiment. In the first embodiment, the customer corresponds to the above-mentioned held-side conversation participant, and the operator corresponds to the conversation participant on the holding side.

The exchange 5 is communicably connected to a call terminal 3 such as a PC, a fixed phone, a mobile phone, a tablet terminal, or a smartphone that is used by a customer via the communication network 2. The communication network 2 is a public network such as the Internet or a PSTN (Public Switched Telephone Network), a wireless communication network, or the like. Further, the exchange 5 is connected to each operator telephone 6 used by each operator of the contact center. The exchange 5 receives the call from the customer and connects the call to the operator telephone 6 of the operator corresponding to the call.

Each operator uses an operator terminal 7. Each operator terminal 7 is a general-purpose computer such as a PC connected to a communication network 8 (LAN (Local Area Network) or the like) in the contact center system 1. For example, each operator terminal 7 records customer voice data and operator voice data in a call between each operator and the customer. Each operator terminal 7 also records voice data of customers who are on hold. The customer voice data and the operator voice data may be generated by being separated from the mixed state by predetermined voice processing. Note that this embodiment does not limit the recording method and the recording subject of such audio data. Each voice data may be generated by a device (not shown) other than the operator terminal 7.

Furthermore, each operator terminal 7 detects an operator's hold operation and hold release operation on the operator telephone 6. Thereby, each operator terminal 7 generates hold section data indicating the start time and end time of the hold section in each call. In addition, this embodiment does not limit the production | generation method and production | generation main body of such reservation area data. For example, the operator terminal 7 may generate the hold section data by detecting the hold sound included in the voice data of the operator without detecting the hold operation and the hold release operation of the operator. The generation of the reserved section data may be performed by a device (not shown) other than the operator terminal 7.

The file server 9 is realized by a general server computer. The file server 9 stores the call data of each call between the customer and the operator together with the identification information of each call. Each call data includes a pair of customer voice data and operator voice data, and the above-described hold section data. The file server 9 acquires customer voice data and operator voice data from another device (each operator terminal 7 or the like) that records each voice of the customer and the operator. Further, the file server 9 acquires the hold section data from each operator telephone 6, each operator terminal 7, the exchange 5 and the like.

The call analysis server 10 analyzes the customer dissatisfaction state with respect to each call data stored in the file server 9.
As shown in FIG. 1, the call analysis server 10 includes a CPU (Central Processing Unit) 11, a memory 12, an input / output interface (I / F) 13, a communication device 14 and the like as a hardware configuration. The memory 12 is a RAM (Random Access Memory), a ROM (Read Only Memory), a hard disk, a portable storage medium, or the like. The input / output I / F 13 is connected to a device that accepts an input of a user operation such as a keyboard and a mouse, and a device that provides information to the user such as a display device and a printer. The communication device 14 communicates with the file server 9 and the like via the communication network 8. Note that the hardware configuration of the call analysis server 10 is not limited.

[Processing configuration]
FIG. 2 is a diagram conceptually illustrating a processing configuration example of the call analysis server 10 in the first embodiment. The call analysis server 10 according to the first embodiment includes a data acquisition unit 20, an utterance detection unit 21, a holding section analysis unit 23, an object determination unit 24, a dissatisfaction determination unit 25, and the like. Each of these processing units is realized, for example, by executing a program stored in the memory 12 by the CPU 11. Further, the program may be installed from a portable recording medium such as a CD (Compact Disc) or a memory card, or another computer on the network via the input / output I / F 13 and stored in the memory 12. Good.

The data acquisition unit 20 acquires the call data of the call to be analyzed from the file server 9 together with the identification information of the call. As described above, the acquired call data includes a pair of customer voice data and operator voice data, and hold section data. The call data may be acquired by communication between the call analysis server 10 and the file server 9, or may be acquired via a portable recording medium.

The utterance detection unit 21 detects the utterance section of the customer in the holding section from the voice data of the customer in the holding section indicated by the holding section data included in the voice data acquired by the data acquisition section 20. For example, the utterance detection unit 21 detects, as the utterance section, a section in which the amplitude equal to or greater than a predetermined value is continued in the voice waveform indicated by the customer's voice data in the hold section. The detection of the utterance section means detecting a section indicating one utterance of the customer in the voice data, whereby the start time and the end time of the section are acquired.

FIG. 3 is a diagram conceptually showing a reserved section and an utterance section in the reserved section. The holding section data as shown in FIG. 3 is indicated by the holding section data. Further, the first voice among the voices uttered by the customer during the call holding, that is, during the holding section, is detected as the first speaking section as shown in FIG.

The inventors have found that the utterance of a customer who is on hold of a call is likely due to the customer's dissatisfaction, and further, among the dissatisfactions expressed during a call hold, It has been found that there are unsatisfied causes (hereinafter referred to as unsatisfied waiting) and other unsatisfied reasons (hereinafter referred to as non-holding unsatisfactory). Dissatisfaction other than waiting on hold may be dissatisfaction with the reception of the other party in the call, or dissatisfaction with the product or service that attracted the call itself.

Furthermore, the present inventors tend to utter utterances indicating dissatisfaction waiting on hold after a certain amount of time has passed since the beginning of the hold interval or the end of the previous utterance in the hold interval. It was found that utterances showing dissatisfaction other than waiting for waiting tend to be uttered relatively soon when the call is on hold. Therefore, in the present embodiment, by providing the holding section analysis unit 23 and the object determination unit 24, the customer's utterance in the holding section is caused by two types of dissatisfaction (holding wait dissatisfaction and dissatisfaction other than the hold waiting). Classification is performed, and only one of the two classified speech groups can be set as a dissatisfaction determination target. Thereby, according to this embodiment, the dissatisfaction degree about only the cause of the specified dissatisfaction can be determined.

The hold section analysis unit 23 calculates a time difference between the start time of the first utterance section detected by the utterance detection unit 21 and the start time of the hold section. Hereinafter, this time difference calculated by the reserved section analysis unit 23 is referred to as a start time difference. This start time difference is also shown in FIG. Furthermore, the reserved section analysis unit 23 compares the calculated start time difference with a first predetermined time threshold. This first predetermined time threshold is a statistical value of the time from the start of the hold interval until the utterance that expresses the hold waiting dissatisfaction is uttered, and is held in advance by the hold interval analysis unit 23 so as to be adjustable. The

Thereby, according to the comparison result between the start time difference and the first predetermined time threshold, it is possible to specify whether the customer's utterance in the hold section in this call represents only the hold waiting dissatisfaction or other than that. it can. The latter represents only dissatisfaction other than hold waiting, or both dissatisfaction other than hold wait and non-hold wait. As the first predetermined time threshold, the start and end times of the operator's hold section and the customer's utterance section, and about 1000 calls labeled whether the customer's utterance in the hold section is unsatisfied with the hold are used. Is set. For example, a histogram of the difference in start time calculated on the horizontal axis and the frequency that the customer's first utterance is an unsatisfied utterance waiting on the vertical axis and the frequency that is not an unsatisfied utterance pending on the vertical axis is created. The most separated time difference is set as a first predetermined time threshold.

Furthermore, in order to identify whether the customer's utterance in the holding section represents only dissatisfaction other than waiting for waiting, or both dissatisfaction other than holding waiting and dissatisfaction other than holding waiting, Two adjacent utterance sections in which the time width between two adjacent utterance sections is larger than the second predetermined time threshold are specified from the plurality of utterance sections detected by the utterance detection unit 21. This second predetermined time threshold is a statistical value of the time from the end of the previous utterance in the hold interval until the utterance that expresses the hold waiting dissatisfaction is uttered, and the hold interval analysis can be adjusted in advance. Held by the unit 23. As with the first predetermined time threshold, the second predetermined time threshold is labeled with the start and end times of the operator's hold section and the customer's utterance section, and whether the customer's utterance in the hold section is unsatisfied with the hold It is set using about 1000 calls made. For example, the horizontal axis represents the time difference from the end of a certain utterance of a customer in the holding section to the beginning of the next utterance, and the ordinate represents that the certain utterance is not dissatisfied on hold and the next utterance is dissatisfied on hold. A histogram is created that takes a certain frequency and a frequency in which both utterances are not unsatisfactory utterances on hold, and a time difference at which these two frequency distributions are most separated is defined as a second predetermined time threshold.

When the two adjacent utterance sections are specified by the holding section analysis unit 23, the customer's utterance in the holding section may include both an utterance indicating a waiting dissatisfaction and an utterance indicating a dissatisfaction other than the holding wait. Identified. In this case, the utterance section after the utterance section after the adjacent two utterance sections specified by the hold section analysis unit 23 represents dissatisfaction waiting, and the utterance sections before the previous utterance section are other than hold wait Represents dissatisfaction. On the other hand, when it is determined by the holding section analysis unit 23 that the two adjacent utterance sections do not exist, it is specified that the utterance of the customer in the holding section represents only dissatisfaction other than waiting for holding.

The target determination unit 24 determines whether or not the voice data of the customer in the hold section is to be a dissatisfaction determination target based on the comparison result between the start time difference by the hold section analysis unit 23 and the first predetermined time threshold. For example, the target determination unit 24 adjustably holds information indicating that the determination target of the degree of dissatisfaction desired by the user of the call analysis server 10 is a hold wait dissatisfaction or a dissatisfaction other than hold wait, and based on this information The determination of the degree of dissatisfaction is performed. Specifically, if the start time difference is greater than or equal to the first predetermined time threshold, the customer's utterance in the holding section in the call is likely to have been expressed only from the hold waiting dissatisfaction. If it is desired to determine the degree of dissatisfaction other than waiting for a hold, it is determined that the voice data of the customer in the hold section is not subject to the dissatisfaction determination.

The object determination unit 24 obtains the voice data of the customer in the holding section based on the comparison result between the start time difference and the first predetermined time threshold and the specific result of the two adjacent utterance sections by the holding section analysis unit 23. Decide whether or not to make a dissatisfaction level. Specifically, when the start time difference is smaller than the first predetermined time threshold and the two adjacent speech sections are not specified (does not exist) by the holding section analysis unit 23, the customer of the holding section in the call is not identified. The utterance is likely to have been expressed only from dissatisfaction other than waiting on hold. Therefore, at this time, if it is desired to determine the degree of dissatisfaction other than hold waiting, the target determination unit 24 determines the voice data of the customer in the hold section as a dissatisfaction determination target, If the determination of dissatisfaction is desired, it is determined that the voice data of the customer in the reserved section is not subject to dissatisfaction determination.

Further, when the difference between the start times is smaller than the first predetermined time threshold value and the two adjacent speech sections are specified by the holding section analysis unit 23, the target determination unit 24 specifies the specified adjacent 2 Dissatisfaction with either the speech data of the utterance section before the previous utterance section of one utterance section or the speech data of the utterance section after the utterance section after the specified two adjacent utterance sections Determine to be judged. When the difference between the start times is smaller than the first predetermined time threshold and the two adjacent utterance sections are specified by the holding section analysis unit 23, the customer's utterance in the holding section in the call is other than the hold waiting There is a high possibility of including utterances indicating dissatisfaction and utterances indicating dissatisfaction waiting. Therefore, at this time, when it is desired to determine the degree of dissatisfaction other than waiting for waiting, the target determining unit 24 determines the utterance section before the previous utterance section of the two adjacent utterance sections specified. Only the voice data is determined as a dissatisfaction level determination target. On the other hand, in the case where it is desired to determine the degree of dissatisfaction of the waiting waiting dissatisfaction, the object determining unit 24 only selects the audio data of the utterance section after the utterance section after the specified two adjacent utterance sections. Decide on a dissatisfaction level.

The dissatisfaction level determination unit 25 performs a predetermined voice analysis process on the voice data of the customer in the holding section determined as the determination target of the dissatisfaction level by the target determination unit 24, thereby determining the dissatisfaction level of the call. judge. The dissatisfaction determination unit 25 generates, as a determination result, output data including identification information related to the analysis target call and information indicating the dissatisfaction level, and determines the determination to the display unit or other output device via the input / output I / F 13. Output the result. The present embodiment does not limit the specific form of output of the determination result.

[Operation example]
Hereinafter, the dissatisfied call determination method in the first embodiment will be described with reference to FIG. FIG. 4 is a flowchart showing an operation example of the call analysis server 10 in the first embodiment. FIG. 4 shows an example in which information indicating that the dissatisfaction determination target desired by the user of the call analysis server 10 is dissatisfaction other than waiting for waiting is set in the call analysis server 10 (target determination unit 24). Show. According to this example, the call analysis server 10 determines the presence or absence or degree of dissatisfaction other than the hold waiting dissatisfaction for the target call.

The call analysis server 10 acquires call data (S41). In the first embodiment, the call analysis server 10 acquires call data to be analyzed from a plurality of call data stored in the file server 9. The acquired call data includes a pair of customer voice data and operator voice data, and hold section data. The voice data of the customer includes voice uttered by the customer while the call is on hold.

The call analysis server 10 acquires customer voice data and hold section data from the call data acquired in step (S41), and the customer utterance in the hold section from the customer voice data in the hold section indicated by the hold section data. A section is detected (S42). The call analysis server 10 acquires the start time and the end time for each utterance section.

The call analysis server 10 calculates the time difference (start time difference) between the start time of the first utterance section and the start time of the hold section in the utterance section detected in the step (S42) (S43).

Subsequently, the call analysis server 10 compares the start time difference calculated in the step (S43) with the first predetermined time threshold value (S44). When the start time difference is equal to or greater than the first predetermined time threshold (S44; NO), the call analysis server 10 determines that the customer's voice data in the hold section is not a subject of dissatisfaction determination in the call (S45), Finish the process. The determination (S44; NO) is the determination and consent that the utterance of the customer in the holding section in the call represents only the holding wait dissatisfaction. In this case, the call analysis server 10 may output information indicating that the customer did not express dissatisfaction with the call.

On the other hand, when the start time difference is smaller than the first predetermined time threshold (S44; YES), the call analysis server 10 determines the time width between the adjacent utterance intervals from the utterance interval detected in the step (S42). Two adjacent utterance sections whose length is larger than the second predetermined time threshold are specified (S46). When the two adjacent utterance sections are not specified in the step (S46) (S47; NO), the call analysis server 10 determines the voice data of the customer in the hold section in the call as the object of dissatisfaction determination (S48). . The determination (S47; NO) is a determination and consent that the customer's utterance in the holding section in the call represents only dissatisfaction other than waiting for holding.

When the two adjacent utterance sections are specified in the step (S46) (S47; YES), the call analysis server 10 determines the utterance section before the previous utterance section in the two specified adjacent utterance sections. Only the audio data is determined as a dissatisfaction level determination target (S49). Here, the determination (S47; YES) is the determination and consent that the utterance of the customer in the holding section in the call includes an utterance that represents dissatisfaction other than waiting for waiting and an utterance that represents dissatisfaction other than waiting for holding. .

The call analysis server 10 determines the degree of dissatisfaction of the call to be analyzed using the audio data determined as the degree of dissatisfaction determination (S50). In this case, the call analysis server 10 may output information indicating the determined degree of dissatisfaction.

In FIG. 4, the case where the determination target of the degree of dissatisfaction desired by the user is dissatisfaction other than the waiting for hold is illustrated. The unsatisfactory call determination method in one embodiment is applicable. In this case, in the step (S45), the voice data of the customer in the holding section in the call is determined as a dissatisfaction level determination target. In step (S48), it is determined that the customer's voice data in the holding section is not a dissatisfaction determination target in the call, and in step (S49), the utterances behind the two specified adjacent utterance sections. Only the voice data of the utterance section after the section is determined as a dissatisfaction level determination target.

Furthermore, the above-mentioned dissatisfied call determination method further includes a step of determining whether the determination target of the dissatisfaction level desired by the user is dissatisfaction other than waiting for waiting or dissatisfied with waiting for holding, and the determination result includes Accordingly, the processing contents of steps (S45), (S48), and (S49) may be switched. In addition, the degree of dissatisfaction related to dissatisfaction other than waiting for waiting and the degree of dissatisfaction relating to waiting for dissatisfaction may be determined separately.

[Operation and Effect of First Embodiment]
As described above, in the first embodiment, the time difference (start time difference) between the start time of the first utterance section and the start time of the hold section in the customer utterance section in the hold section is calculated. 1 The predetermined time threshold is compared. This comparison is based on the characteristics (trends) of a person (customer) that an utterance that represents dissatisfaction on hold is uttered after a certain amount of time has elapsed since the start of the hold interval. If the threshold is equal to or greater than one predetermined time, it can be specified that the customer's utterance in the hold section represents only the hold wait dissatisfaction.

Furthermore, in the first embodiment, when the start time difference is smaller than the first predetermined time threshold, two adjacent utterance intervals in which the time width between the two adjacent utterance intervals is larger than the second predetermined time threshold are specified. Is done. This specification is based on the characteristics (trends) of a person (customer) that an utterance indicating dissatisfaction waiting is uttered after a certain amount of time has elapsed since the end of the previous utterance in the hold section. Is impossible, it is possible to specify that the customer's utterance in the holding section represents only dissatisfaction other than holding on hold. When the identification is possible, it is possible to specify that the customer's utterance in the holding section includes an utterance indicating dissatisfaction other than the waiting waiting and an utterance indicating the holding waiting dissatisfaction. Furthermore, the two adjacent utterance sections that have been identified can distinguish a plurality of utterances in the holding section into utterances that indicate dissatisfaction other than waiting for waiting and utterances that indicate dissatisfaction waiting.

As described above, according to the first embodiment, two types of dissatisfaction are caused by the customer's utterance in the holding section by the information processing (comparison processing and specific processing) in consideration of the characteristics regarding the customer's utterance during the call holding. It is possible to classify them according to (unsatisfied with waiting and unsatisfied other than waiting). Thereby, according to 1st Embodiment, the dissatisfaction degree of the call can be determined only using the audio | voice data of the speech corresponding to the desired dissatisfaction cause among the classified two types of dissatisfaction causes.

Therefore, according to the first embodiment, it is possible to determine the degree of dissatisfaction with respect to only a desired cause of dissatisfaction with high accuracy. Furthermore, according to the first embodiment, the predetermined speech analysis processing is performed only on the speech data corresponding to the desired cause of dissatisfaction, so that the entire call data or the entire holding section is targeted for determination. Compared to the above, the processing efficiency can be improved, and the processing speed can be increased.

[Second Embodiment]
The second embodiment generates information indicating the cause of dissatisfaction with the degree of dissatisfaction of the call determined as in the first embodiment. Hereinafter, the contact center system 1 in the second embodiment will be described focusing on the content different from the first embodiment. In the following description, the same contents as those in the first embodiment are omitted as appropriate.

[Processing configuration]
FIG. 5 is a diagram conceptually illustrating a processing configuration example of the call analysis server 10 in the second embodiment. The call analysis server 10 in the second embodiment further includes a cause estimating unit 27 in addition to the configuration of the first embodiment. The cause estimation unit 27 is realized, for example, by executing a program stored in the memory 12 by the CPU 11, similarly to the other processing units.

The cause estimation unit 27 holds based on the comparison result between the start time difference and the first predetermined time threshold by the hold segment analysis unit 23 and the identification result of the two adjacent utterance intervals by the hold segment analysis unit 23. Cause information indicating a cause of dissatisfaction corresponding to the customer's utterance in the section is generated. The cause estimating unit 27 may generate output data including the cause information and output the determination result to the display unit or another output device via the input / output I / F 13. However, the specific form of outputting the cause information is not limited.

As the cause of dissatisfaction corresponding to the customer's utterance in the reserved section indicated by the cause information, for example, there are the following three cases. In the first case, only the length of the waiting waiting time is caused as a cause of dissatisfaction. In the second case, only the length of the hold waiting time is caused as a cause of dissatisfaction. The third case is a case where both the first type and the second type are caused by dissatisfaction.

The cause estimating unit 27 generates cause information in the first case when the start time difference is equal to or greater than a first predetermined time threshold. The cause estimating unit 27 does not have two adjacent utterance sections in which the start time difference is smaller than the first predetermined time threshold and the time width between the two adjacent utterance sections is larger than the second predetermined time threshold. In the case, the cause information in the second case is generated. The cause estimating unit 27 has two adjacent utterance sections in which the difference in start time is smaller than the first predetermined time threshold and the time width between the two adjacent utterance sections is larger than the second predetermined time threshold. In the case, the cause information in the third case is generated.

[Operation example]
Hereinafter, the dissatisfied call determination method in the second embodiment will be described. In addition to the operation of the first embodiment (see FIG. 4), the call analysis server 10 further generates cause information as described above. The call analysis server 10 may generate the cause information after determining the degree of dissatisfaction as in the first embodiment, or may generate the cause information before determining the degree of dissatisfaction.

According to the example of FIG. 4, in the case of (S44; NO), the call analysis server 10 generates cause information indicating that only the length of the waiting waiting time is the cause of dissatisfaction. In the case of (S44; YES) and (S47; NO), the call analysis server 10 generates cause information indicating that only the length of the holding waiting time is a cause of dissatisfaction. In the case of (S44; YES) and (S47; YES), the call analysis server 10 generates cause information indicating that the length of the holding waiting time and the rest are the causes of dissatisfaction.

[Operation and Effect of Second Embodiment]
As described above, in the second embodiment, the cause information indicating the cause of dissatisfaction corresponding to the customer's utterance in the holding section is generated together with the determination of the degree of dissatisfaction. For the estimation of the cause of dissatisfaction, the processing results of the reserved section analysis unit 23 and the object determination unit 24 in the first embodiment described above are used. Therefore, according to the second embodiment, the cause of dissatisfaction can be estimated with high accuracy by information processing (comparison processing and specific processing) that takes into account the characteristics relating to the customer's utterance while the call is on hold.

[Modification]
In the first embodiment and the second embodiment described above, an example in which the call data includes a pair of customer voice data and operator voice data, and holding section data is shown. , Customer and operator voice text data, and time information of each voice text may be further included. Here, the voice text data is data in which a voice uttered by a customer or an operator is converted into text. Further, the time information of each voice text is information related to the time when the utterance indicated by each voice text is uttered. This voice text data is generated, for example, by applying voice recognition processing to each voice of the customer and the operator at each operator terminal 7 or file server 9.

In this case, the utterance detection unit 21 of the call analysis server 10 is not necessary. The holding section analysis unit 23 can perform processing using the time information of each voice text and the holding section data. Further, the dissatisfaction level determination unit 25 may perform dissatisfaction level determination using the voice text data. As described above, the customer utterance section in the holding section detected by the utterance detection unit 21 may be generated by a device other than the call analysis server 10.

In each of the above-described embodiments, the holding section analysis unit 23 compares the start time difference with the first predetermined time threshold. However, the holding section analysis unit 23 uses other values (representative) obtained from the start time difference. The threshold value may also be referred to as a first predetermined time threshold value. For example, the representative value may be a value output from a predetermined function having the start time difference as an input. In this case, the predetermined function may be, for example, a function that inputs a start time difference and outputs a value indicating the accuracy that the dissatisfaction corresponding to the utterance in the hold section is only the hold wait dissatisfaction.

Similarly, each embodiment described above shows an example in which the holding section analysis unit 23 specifies two adjacent utterance sections in which the time width between the two adjacent utterance sections is larger than the second predetermined time threshold. However, the holding section analysis unit 23 has two adjacent values whose other values (also referred to as representative values) obtained from the time width between the two adjacent speech sections are larger than the second predetermined time threshold. The utterance section may be specified. In this case, the representative value may be a value output from a predetermined function having the time width as an input. In this case, the predetermined function is, for example, a function that takes a time width as an input and outputs a value indicating the accuracy that the dissatisfaction corresponding to the utterance in the hold section is only due to the length of the hold waiting time. It is good.

According to such a modified example, the above-described cause estimation unit 27 determines that the cause of dissatisfaction corresponding to the customer's utterance in the holding section is based on the length of the holding waiting time. It may be included in the cause information as a value indicating the certainty of being. Similarly, the cause estimating unit 27 uses the above representative value obtained from the time width as a value indicating the accuracy that the cause of dissatisfaction corresponding to the customer's utterance in the holding section is only due to the length of the holding waiting time. May be included in the cause information.

Further, in the first embodiment and the second embodiment described above, the reserved section analysis unit 23 determines that two adjacent speech sections in which the time width between the two adjacent speech sections is larger than the second predetermined time threshold value. Although the example to identify was shown, you may make it the pending | holding area analysis part 23 not perform the specific process. That is, the reserved section analysis unit 23 may perform only the comparison between the start time difference and the first predetermined time threshold. In this case, the object determination unit 24 may determine whether or not the voice data of the customer in the reserved section is to be a dissatisfaction level determination target based on the comparison result. In this form, when utterances indicating unsatisfied waiting and utterances indicating dissatisfaction other than waiting are mixed in the holding section, the degree of dissatisfaction is determined from utterances indicating both causes of dissatisfaction. Such a form is effective when there is no need to limit the cause of dissatisfaction. In this case, the cause estimating unit 27 may generate cause information indicating whether the utterance of the customer on hold is caused only by the cause of dissatisfaction with the length of the hold waiting time.

[Other Embodiments]
In each of the above-described embodiments, the call data is handled. However, the above-mentioned dissatisfied conversation determination device and the dissatisfied conversation determination method may be applied to an apparatus or a system that handles conversation data other than a call. In this case, for example, a recording device for recording a conversation to be analyzed is installed at a place (conference room, bank window, store cash register, etc.) where the conversation is performed. Further, when the conversation data is recorded in a state in which the voices of a plurality of conversation participants are mixed, the conversation data is separated from the mixed state into voice data for each conversation participant by a predetermined voice process.

In the first and second embodiments described above, the hold interval data is generated by detecting the hold operation and hold release operation by the operator, detecting the hold sound included in the operator's voice data, and the like. In the form in which conversation data is handled, the movement of conversation participants who put the conversation on hold is detected, and the standard expression that is uttered when putting the conversation on hold and releasing the hold (for example, “I am a little rude”) ), Etc. may be automatically generated. Of course, the hold section data may be input by a user operation.

In the flowchart used in the above description, a plurality of steps (processes) are described in order, but the execution order of the steps executed in the present embodiment is not limited to the description order. In the present embodiment, the order of the illustrated steps can be changed within a range that does not hinder the contents. Moreover, each above-mentioned embodiment and each modification can be combined in the range with which the content does not conflict.

Some or all of the above-described actual forms and modifications may be specified as in the following supplementary notes. However, each actual form and each modification are not limited to the following description.

(Appendix 1)
A data acquisition unit for acquiring hold section data indicating start time and end time of a hold section of the conversation and voice data of the conversation;
By performing a predetermined voice analysis process on the voice data of the held-side conversation participant in the holding section indicated by the holding section data included in the voice data acquired by the data acquisition unit, A dissatisfaction determination unit that determines the dissatisfaction level of the conversation;
A dissatisfied conversation determination device comprising:

(Appendix 2)
From the voice data of the held-side conversation participant in the holding section indicated by the holding section data, included in the voice data acquired by the data acquisition unit, the first of the held-side conversation participants in the holding section An utterance detection unit for detecting an utterance section of
The time difference between the start time of the first utterance section detected by the utterance detection unit and the start time of the holding section is calculated, and the time difference or a representative value corresponding to the time difference is compared with a first predetermined time threshold value. A holding section analysis unit;
Based on the comparison result by the holding section analysis unit, an object determining unit that determines whether or not the voice data of the held-side conversation participant in the holding section is a determination target of the dissatisfaction degree determination unit;
The unsatisfactory conversation determination device according to supplementary note 1, further comprising:

(Appendix 3)
Based on the comparison result by the holding section analysis unit, cause information indicating whether or not the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the holding section is only due to the length of the holding waiting time. A cause estimator to generate,
The unsatisfactory conversation determination device according to Appendix 2, further comprising:

(Appendix 4)
Based on the time difference calculated by the holding section analysis unit, the probability that the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the holding section is only due to the length of the holding waiting time. A cause estimator that generates cause information to indicate,
The unsatisfactory conversation determination device according to Appendix 2, further comprising:

(Appendix 5)
The utterance detection unit detects a plurality of utterance sections of the held-side conversation participant in the holding section from the voice data of the held-side conversation participant of the holding section,
The holding section analysis unit includes two adjacent utterance sections in which a time width between two adjacent utterance sections or a representative value corresponding to the time width is larger than a second predetermined time threshold from the plurality of utterance sections. Identify
The target determination unit is configured to determine whether the reserved section is based on the comparison result between the time difference or the representative value and the first predetermined time threshold, and the specific result. Determining whether or not the voice data of the conversation participant is to be determined by the dissatisfaction level determination unit;
The unsatisfactory conversation determination device according to any one of supplementary notes 2 to 4.

(Appendix 6)
The target determination unit includes voice data of an utterance section before the previous utterance section of the two adjacent utterance sections specified by the holding section analysis section, and the two adjacent utterance sections specified Any one of the speech data of the utterance section after the utterance section behind is set as the determination target of the dissatisfaction level determination unit.
The unsatisfactory conversation determination device according to appendix 5.

(Appendix 7)
The cause estimation unit is based on the identification result of the two adjacent utterance intervals by the hold interval analysis unit, and the cause of dissatisfaction corresponding to the utterance of the held conversation participant in the hold interval is a hold waiting time. Generating cause information indicating whether it is only due to the length of the waiting time, or both due to the length of the holding waiting time and other than the length of the holding waiting time,
The unsatisfactory conversation determination device according to appendix 5 or 6.

(Appendix 8)
Based on the time width between the two adjacent utterance intervals calculated by the holding interval analysis unit, the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the hold interval is the hold waiting time. A cause estimator that generates cause information indicating the accuracy of being only due to length other than
The unsatisfactory conversation determination device according to appendix 5 or 6, further comprising:

(Appendix 9)
In a dissatisfied conversation determination method executed by at least one computer,
Obtaining the hold section data indicating the start time and the end time of the conversation hold section and the voice data of the conversation;
The speech dissatisfaction level is determined by performing a predetermined voice analysis process on the voice data of the held-side conversation participant in the hold section indicated by the hold section data included in the acquired voice data. judge,
A method for determining dissatisfied conversations.

(Appendix 10)
Detecting the first utterance section of the held-side conversation participant in the holding section from the voice data of the held-side conversation participant in the holding section indicated by the holding section data included in the acquired voice data And
Calculating the time difference between the start time of the detected first utterance interval and the start time of the hold interval;
Comparing the time difference or a representative value corresponding to the time difference with a first predetermined time threshold;
Based on the comparison result, it is determined whether or not the voice data of the held-side conversation participant in the holding section is the object of the dissatisfaction determination.
The unsatisfactory conversation determination method according to supplementary note 9, further including:

(Appendix 11)
Based on the comparison result, generate cause information indicating whether or not the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the hold section is due to the length of the hold waiting time.
The dissatisfied conversation determination method according to supplementary note 10, further including:

(Appendix 12)
Based on the calculated time difference, the cause information indicating the probability that the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the hold section is only due to the length of the hold waiting time is generated. ,
The dissatisfied conversation determination method according to supplementary note 10, further including:

(Appendix 13)
From the voice data of the held-side conversation participant in the holding section, detecting a plurality of utterance sections of the held-side conversation participant in the holding section,
From the plurality of utterance sections, specify two adjacent utterance sections in which the time width between two adjacent utterance sections or a representative value corresponding to the time width is greater than a second predetermined time threshold.
Further including
The determination regarding the object of determination of the degree of dissatisfaction is based on the comparison result between the time difference or the representative value and the first predetermined time threshold, and the specific result, and the held-side conversation participant in the holding section. To determine whether or not the voice data is subject to the dissatisfaction determination,
The dissatisfied conversation determination method according to any one of appendices 10 to 12.

(Appendix 14)
The determination regarding the object of determination of the dissatisfaction level includes voice data of an utterance section before the previous utterance section of the two adjacent utterance sections specified, and of the two adjacent utterance sections specified. Either one of the speech data of the utterance section after the subsequent utterance section is the object of the dissatisfaction determination,
The method for determining a dissatisfied conversation according to appendix 13.

(Appendix 15)
The generation of the cause information is based on the result of specifying the two adjacent utterance intervals, and the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the hold interval is other than the length of the hold waiting time. Generating cause information that indicates whether it is only due to the length of the hold waiting time or not due to the length of the hold waiting time,
15. The method for determining a dissatisfied conversation according to appendix 13 or 14.

(Appendix 16)
Based on the time width between the two adjacent utterance intervals, the cause of dissatisfaction corresponding to the utterance of the held conversation participant in the hold interval is only due to the length of the hold waiting time. Generate cause information indicating the accuracy of
The unsatisfactory conversation determination method according to supplementary note 13 or 14, further including:

(Appendix 17)
A program for causing at least one computer to execute the unsatisfactory conversation determination method according to any one of appendices 9 to 16.

(Appendix 18)
A recording medium for recording the program according to attachment 17 in a computer-readable manner.

This application claims priority based on Japanese Patent Application No. 2012-240759 filed on October 31, 2012, the entire disclosure of which is incorporated herein.

Claims (15)

1. A data acquisition unit for acquiring hold section data indicating start time and end time of a hold section of the conversation and voice data of the conversation;
   By performing a predetermined voice analysis process on the voice data of the held-side conversation participant in the holding section indicated by the holding section data included in the voice data acquired by the data acquisition unit, A dissatisfaction determination unit that determines the dissatisfaction level of the conversation;
   A dissatisfied conversation determination device comprising:
2. From the voice data of the held-side conversation participant in the holding section indicated by the holding section data, included in the voice data acquired by the data acquisition unit, the first of the held-side conversation participants in the holding section An utterance detection unit for detecting an utterance section of
   The time difference between the start time of the first utterance section detected by the utterance detection unit and the start time of the holding section is calculated, and the time difference or a representative value corresponding to the time difference is compared with a first predetermined time threshold value. A holding section analysis unit;
   Based on the comparison result by the holding section analysis unit, an object determining unit that determines whether or not the voice data of the held-side conversation participant in the holding section is a determination target of the dissatisfaction degree determination unit;
   The dissatisfied conversation determination device according to claim 1, further comprising:
3. Based on the comparison result by the holding section analysis unit, cause information indicating whether or not the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the holding section is only due to the length of the holding waiting time. A cause estimator to generate,
   The dissatisfied conversation determination device according to claim 2, further comprising:
4. Based on the time difference calculated by the holding section analysis unit, the probability that the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the holding section is only due to the length of the holding waiting time. A cause estimator that generates cause information to indicate,
   The dissatisfied conversation determination device according to claim 2, further comprising:
5. The utterance detection unit detects a plurality of utterance sections of the held-side conversation participant in the holding section from the voice data of the held-side conversation participant of the holding section,
   The holding section analysis unit includes two adjacent utterance sections in which a time width between two adjacent utterance sections or a representative value corresponding to the time width is larger than a second predetermined time threshold from the plurality of utterance sections. Identify
   The target determination unit is configured to determine whether the reserved section is based on the comparison result between the time difference or the representative value and the first predetermined time threshold, and the specific result. Determining whether or not the voice data of the conversation participant is to be determined by the dissatisfaction level determination unit;
   The unsatisfactory conversation determination apparatus of any one of Claim 2 to 4.
6. The target determination unit includes voice data of an utterance section before the previous utterance section of the two adjacent utterance sections specified by the holding section analysis section, and the two adjacent utterance sections specified Any one of the speech data of the utterance section after the utterance section behind is set as the determination target of the dissatisfaction level determination unit.
   The unsatisfactory conversation determination device according to claim 5.
7. The cause estimation unit is based on the identification result of the two adjacent utterance intervals by the hold interval analysis unit, and the cause of dissatisfaction corresponding to the utterance of the held conversation participant in the hold interval is a hold waiting time. Generating cause information indicating whether it is only due to the length of the waiting time, or both due to the length of the holding waiting time and other than the length of the holding waiting time,
   The unsatisfactory conversation determination apparatus according to claim 5 or 6.
8. Based on the time width between the two adjacent utterance intervals calculated by the holding interval analysis unit, the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the hold interval is the hold waiting time. A cause estimator that generates cause information indicating the accuracy of being only due to length other than
   The dissatisfied conversation determination device according to claim 5 or 6, further comprising:
9. In a dissatisfied conversation determination method executed by at least one computer,
   Obtaining the hold section data indicating the start time and the end time of the conversation hold section and the voice data of the conversation;
   The speech dissatisfaction level is determined by performing a predetermined voice analysis process on the voice data of the held-side conversation participant in the hold section indicated by the hold section data included in the acquired voice data. judge,
   A method for determining dissatisfied conversations.
10. Detecting the first utterance section of the held-side conversation participant in the holding section from the voice data of the held-side conversation participant in the holding section indicated by the holding section data included in the acquired voice data And
    Calculating the time difference between the start time of the detected first utterance interval and the start time of the hold interval;
    Comparing the time difference or a representative value corresponding to the time difference with a first predetermined time threshold;
    Based on the comparison result, it is determined whether or not the voice data of the held-side conversation participant in the holding section is the object of the dissatisfaction determination.
    The dissatisfied conversation determination method according to claim 9, further comprising:
11. Based on the comparison result, generate cause information indicating whether or not the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the hold section is due to the length of the hold waiting time.
    The dissatisfied conversation determination method according to claim 10, further comprising:
12. Based on the calculated time difference, the cause information indicating the probability that the cause of dissatisfaction corresponding to the utterance of the held-side conversation participant in the hold section is only due to the length of the hold waiting time is generated. ,
    The dissatisfied conversation determination method according to claim 10, further comprising:
13. From the voice data of the held-side conversation participant in the holding section, detecting a plurality of utterance sections of the held-side conversation participant in the holding section,
    From the plurality of utterance sections, specify two adjacent utterance sections in which the time width between two adjacent utterance sections or a representative value corresponding to the time width is greater than a second predetermined time threshold.
    Further including
    The determination regarding the object of determination of the degree of dissatisfaction is based on the comparison result between the time difference or the representative value and the first predetermined time threshold, and the specific result, and the held-side conversation participant in the holding section. To determine whether or not the voice data is subject to the dissatisfaction determination,
    The dissatisfied conversation determination method according to any one of claims 10 to 12.
14. The determination regarding the object of determination of the dissatisfaction level includes voice data of an utterance section before the previous utterance section of the two adjacent utterance sections specified, and of the two adjacent utterance sections specified. Either one of the speech data of the utterance section after the subsequent utterance section is the object of the dissatisfaction determination,
    The unsatisfactory conversation determination method according to claim 13.
15. A program for causing at least one computer to execute the unsatisfactory conversation determination method according to any one of claims 9 to 14.

Images