JP5691174B2 - Operator selection device, operator selection program, operator evaluation device, operator evaluation program, and operator evaluation method - Google Patents

Description

The present invention relates to an operator selection device , an operator selection program , an operator evaluation device, an operator evaluation program, and an operator evaluation method .

  In recent years, various products such as health foods and cosmetics have been mail-ordered. In mail order sales, for example, user orders are received by e-mail or telephone. For example, there is a call center as one of the windows for receiving orders from users. Call centers receive calls from a wide range of users, and call center operators respond to a wide range of users.

  Moreover, an operator responds not only to the user who made the call for the purpose of purchasing goods, but also to the user who made the call for the purpose of talking with the operator. A user who makes a phone call for the purpose of talking to an operator gives priority to talking over an order for a product. Hereinafter, the user is also referred to as “customer”. In addition, calling a call center for the purpose of talking with an operator, not for the purpose of purchasing a product, is also described as “calling the call center in half the talk”. In addition, a user who makes a call to the call center in half a story is also referred to as a “half story user” or a “half story customer”.

  Some users who are half-talked are somewhat interested in the product, and some actually order the product. For this reason, even if the operator is a half-user, it is required to provide sufficient customer service. For example, the operator is required to fully explain the product and answer the questions from the customers carefully. Also, call center managers want to operate the call center efficiently while providing sufficient customer service. For example, a call center manager manages the number of calls accepted per operator. There is also an apparatus for judging whether the client's product knowledge and pronunciation are good or bad by analyzing the utterance of the client and the utterance of the operator.

JP 2008-22493 A

  However, the conventional technology has a problem in that the customer service is not sufficiently provided. For example, the number of calls received per person is managed by the operator, and the conversation may be terminated without sufficiently responding to a half-talking user. For this reason, it did not provide enough customer service to half the users.

  Further, the conventional technology has a problem that the business efficiency of the call center has been reduced. For example, it is a mental burden for an operator who is not good at talking to people to talk for a long time with a half-talking user. For this reason, the operator work of the operator who is not good at public talk stagnated, and the work efficiency of the call center was reduced.

  The disclosed technique has been made in view of the above. The disclosed technology can select an operator that is suitable for half-talking users, and as a result, it can provide sufficient customer service to half-speaking users, and operator selection that can operate the call center efficiently. An object is to provide an apparatus and an operator selection program.

  In one aspect, the disclosed operator selection device includes a conversation voice data storage unit that stores conversation voice data between a user using the call center and the operator in association with operator information for identifying the operator of the call center. Is provided. In addition, the operator selection device calculates the value related to the operator's listening skill by analyzing the conversation voice data stored in the conversation voice data storage unit using a predetermined analysis standard for each operator information. A part. The operator selection device includes a selection unit that selects an operator based on the value calculated by the calculation unit. In addition, the operator selection device performs analysis using a first number indicating the number of times the operator's voice and the user's voice overlap as a predetermined analysis criterion, and the first result obtained as an analysis result A value indicating that the listening skill is high is calculated as the number of times is small, and a value indicating that the listening skill is low as the first number of times is large.

  According to one aspect of the disclosed operator selection device, it is possible to select an operator that is suitable for a half-speaker user, and as a result, sufficient customer service can be provided to a half-speaker user, thereby improving the efficiency of the call center. This has the effect of being able to be operated manually.

FIG. 1 is a block diagram illustrating an example of the configuration of the operator selection device according to the first embodiment. FIG. 2 is a diagram illustrating an overall image of the operator selection device according to the second embodiment. FIG. 3 is a block diagram illustrating an example of the configuration of the operator selection device according to the second embodiment. FIG. 4 is a diagram illustrating an example of information stored in the operator table according to the second embodiment. FIG. 5 is a diagram illustrating an example of information stored in the customer table in the second embodiment. FIG. 6 is a diagram illustrating an example of information stored in the conversation voice table according to the second embodiment. FIG. 7 is a diagram illustrating an example of information stored in the keyword table in the second embodiment. FIG. 8 is a diagram illustrating an example of information stored in the coefficient table according to the second embodiment. FIG. 9 is a diagram illustrating an example of information stored in the excluded text table in the second embodiment. FIG. 10 is a schematic diagram illustrating an example of information stored in the identification table according to the second embodiment. FIG. 11 is a diagram illustrating an example of information stored in the duplication count table according to the second embodiment. FIG. 12A is a diagram illustrating a case where the operator's voice and the user's voice do not overlap. FIG. 12B is a diagram illustrating a case where an operator's voice and a user's voice overlap each other. FIG. 13 is a diagram illustrating an example of information stored in the ratio table in the second embodiment. FIG. 14 is a diagram illustrating an example of information stored in the exclusion time ratio table according to the second embodiment. FIG. 15 is a diagram illustrating an example of information stored in the number-of-interests table according to the second embodiment. FIG. 16 is a diagram illustrating an example of information stored in the expansion number table according to the second embodiment. FIG. 17A is a diagram for describing a case where the user time becomes long before and after the operator hits the alarm. FIG. 17B is a diagram illustrating a case where the user time does not become long before and after the operator hits the alarm. FIG. 18 is a diagram illustrating an example of information stored in the multiple decompression number table according to the second embodiment. FIG. 19A is a diagram for explaining a case where the user time continuously increases two or more times before and after the operator hits the alarm. FIG. 19B is a diagram illustrating a case where the user time does not continuously increase two or more times before and after the operator hits the alarm. FIG. 20 is a diagram illustrating the ratio calculation unit according to the second embodiment. FIG. 21 is a diagram illustrating an example of values used for explaining the dedicated operator selection unit according to the second embodiment. FIG. 22 is a flowchart illustrating an example of the flow of the distribution process performed by the distribution process control unit according to the second embodiment. FIG. 23 is a flowchart illustrating an example of the flow of overall processing performed by the selection processing control unit according to the second embodiment. FIG. 24 is a flowchart illustrating an example of a flow of processing performed by the duplication number calculation unit according to the second embodiment. FIG. 25 is a flowchart illustrating an example of a process flow by the ratio calculation unit according to the second embodiment. FIG. 26 is a flowchart for explaining an example of a flow of processing by the excluded ratio calculation unit in the second embodiment. FIG. 27 is a flowchart for describing an example of a flow of processing performed by the identification information calculation unit according to the second embodiment. FIG. 28 is a schematic diagram illustrating an example of a computer that executes an operator selection program according to the second embodiment.

  Hereinafter, embodiments of the disclosed operator selection device and operator selection program will be described in detail with reference to the drawings. Note that the invention disclosed by this embodiment is not limited. Each embodiment can be appropriately combined within a range in which processing contents do not contradict each other.

  An example of the configuration of the operator selection device 100 according to the first embodiment will be described with reference to FIG. FIG. 1 is a block diagram illustrating an example of the configuration of the operator selection device according to the first embodiment. In the example illustrated in FIG. 1, the operator selection device 100 includes a conversation voice data storage unit 101, a calculation unit 102, and a selection unit 103.

  The conversation voice data storage unit 101 stores conversation voice data between a user who uses the call center and the operator in association with operator information for identifying the operator of the call center. Then, the calculation unit 102 analyzes the conversation voice data stored in the conversation voice data storage unit 101 for each operator information using a predetermined analysis standard, thereby calculating a value related to the operator's listening skill for each operator information. . The selection unit 103 selects an operator based on the value calculated by the calculation unit 102.

  As described above, the operator selection device 100 according to the first embodiment analyzes the conversation between the operator of the call center and the user, and selects the operator based on whether the operator is making the user speak for a long time. That is, the operator selection device 100 analyzes the conversation between the operator and the user based on a criterion for evaluating whether the operator is a good listener, and selects a good operator based on the analysis result.

  If it is an operator who is good at listening, it is considered that it is less likely to end the conversation without sufficiently responding to a user who is half a story, and is suitable for a user who is half a story. Based on this, in the first embodiment, an operator who is good at listening can be selected by selecting an operator who is good at listening, and sufficient customer service can be provided to a user who is half a story.

  Moreover, if the operator is a good listener, even if he / she talks with a half-talking user for a long time, it is considered that there is no mental burden, and it is considered suitable for a half-speaking user. Based on this, in the first embodiment, an operator who is good at listening can be selected by selecting an operator who is good at listening, so that the operator's operations are not stagnated and the call center can be operated efficiently. It is.

[Overview of operator selection device according to embodiment 2]
Next, an operator selection device 200 according to the second embodiment will be described. Below, the whole image of the operator selection apparatus 200, a structure, the flow of a process, and an effect are demonstrated in order. In the second embodiment, the operator selection apparatus 200 executes a series of processes for selecting a full-time operator who is an operator corresponding to a call from a half-talked customer, and also performs a distribution process for allocating a call made to a call center to an operator. An example of execution will be described. However, the present invention is not limited to this. For example, the operator selection device 200 may not execute the distribution process.

  The overall image of the operator selection device 200 according to the second embodiment will be described with reference to FIG. FIG. 2 is a diagram illustrating an overall image of the operator selection device according to the second embodiment. In FIG. 2, reference numeral 301 denotes a customer terminal that is a terminal used by a customer. For example, the customer terminal 301 corresponds to a mobile phone or a fixed phone. 302 indicates an operator, and 303 indicates a dedicated operator. Reference numeral 310 denotes a customer table that stores information about call center customers. Reference numeral 320 denotes a conversation voice table that stores conversation voice data between the operator and the customer. Reference numeral 330 denotes a distribution process control unit that executes the distribution process, and reference numeral 340 denotes a selection process control unit that executes a series of processes for selecting a dedicated operator. Note that details of the operator selection device 200 will be described later, and in the description with reference to FIG.

  As illustrated in FIG. 2, the operator selection device 200 includes a customer table 310, a conversation voice table 320, a distribution processing control unit 330, and a selection processing control unit 340. In the operator selection device 200, when a call is made from the customer terminal 301, the distribution processing control unit 330 distributes the call to the operator 302 or the dedicated operator 303. For example, as shown in (1) of FIG. 2, the distribution processing control unit 330 distributes the call to the operator 302 when receiving the first call.

  In addition, for example, as shown in (2) of FIG. 2, when the distribution processing control unit 330 accepts the second and subsequent calls, if the call is from a customer who wants to place an order, As shown in (2a), the telephone is distributed to the operator 302. On the other hand, as shown in (2b) of FIG. 2, the distribution processing control unit 330 distributes the call to the dedicated operator 303 if the call is from a half-talked customer.

  That is, when the first call is accepted, the distribution processing control unit 330 cannot distinguish whether the call is from a half-talked customer before allocating, and distributes the call to the operator 302 for the time being. On the other hand, in the case of the second and subsequent calls, the sorting processing control unit 330 uses the customer table 310 to distinguish whether the customer is a half-speaker. If the customer is a half-talker, the telephone is distributed to the full-time operator 303.

  As shown in FIG. 2, the selection processing control unit 340 selects a dedicated operator. For example, as shown in (3) of FIG. 2, the selection process control unit 340 analyzes the conversation voice table. Then, the selection processing control unit 340 is configured based on the number of times that the customer's voice and the operator's voice are overlapped, the ratio of the time that the customer was talking to the time that the operator was talking, and the like (4) in FIG. A full-time operator is selected as shown in

[Configuration of operator selection device]
An example of the configuration of the operator selection device 200 according to the second embodiment will be described with reference to FIG. FIG. 3 is a block diagram illustrating an example of the configuration of the operator selection device according to the second embodiment. In the example illustrated in FIG. 3, the operator selection device 200 includes a first storage unit 410, a second storage unit 420, a third storage unit 430, a distribution processing control unit 510, and a selection processing control unit 520. And have. Moreover, the operator selection apparatus 200 is connected with the customer terminal 600 in the example shown in FIG.

(Customer terminal)
Customer terminal 600 is connected to operator selection device 200. The customer terminal 600 is a terminal used by a customer. For example, the customer terminal 600 corresponds to a mobile phone or a fixed phone. The customer terminal 600 is used when a customer calls a call center. The customers who use the customer terminal 600 include not only customers who have made phone calls for the purpose of purchasing commodities but also half-talkers.

(Operator selection device)
Returning to the description of the operator selection device 200. Hereinafter, for convenience of explanation, first, each unit related to the distribution process, that is, the first storage unit 410 and the distribution process control unit 510 will be described. Thereafter, each unit related to a series of processes for selecting a dedicated operator, that is, the second storage unit 420, the third storage unit 430, and the selection process control unit 520 will be described.

(Each part related to distribution processing)
Each unit related to the distribution process will be described. The first storage unit 410 is connected to the distribution process control unit 510 and the selection process control unit 520. The first storage unit 410 stores data used for distribution processing by the distribution processing control unit 510. The first storage unit 410 corresponds to, for example, a semiconductor memory device such as a random access memory (RAM), a read only memory (ROM), and a flash memory, or a hard disk or an optical disk. In the example illustrated in FIG. 3, the first storage unit 410 includes an operator table 411 and a customer table 412.

  As shown in FIG. 4, the operator table 411 stores operator information for identifying an operator and information indicating whether or not the operator is a full-time operator in association with each other. FIG. 4 is a diagram illustrating an example of information stored in the operator table according to the second embodiment. In the example shown in FIG. 4, the operator table 411 stores “operator code” indicating a series of alphanumeric characters for identifying the operator by the operator selection device 200 and “operator name” indicating the name of the operator as operator information. To do. Further, the operator table 411 stores a “dedicated operator flag” that is information indicating a dedicated operator in association with an operator code and a name. Hereinafter, a case where the dedicated operator flag “1” indicates that the operator is a dedicated operator and the dedicated operator flag “null” indicates that the operator is not a dedicated operator will be described as an example. Note that null indicates a state in which no data is included.

  In the example illustrated in FIG. 4, the operator table 411 stores a dedicated operator flag “null” in association with the operator code “OP0000001” and the operator name “Yuzu Yamamoto”. The operator table 411 stores a dedicated operator flag “1” in association with the operator code “OP0000003” and the operator name “Masako Araki”. That is, for example, the operator table 411 stores that “Yamamoto Yuzu” is not a full-time operator and “Araki Masako” is a full-time operator. Similarly, the operator table 411 stores dedicated operator flags for other operator information as well.

  Of the information stored in the operator table 411, the operator code and the operator name are input by, for example, a call center administrator. The dedicated operator flag is input by the selection processing control unit 520. Information stored in the operator table 411 is used by the distribution processing control unit 510.

  As shown in FIG. 5, the customer table 412 stores information identifying a call center customer and information indicating whether the customer is a half-talker or not. FIG. 5 is a diagram illustrating an example of information stored in the customer table in the second embodiment. In the example shown in FIG. 5, the customer table 412 includes “customer code” indicating a series of alphanumeric characters used by the operator selection device 200 to identify a customer and “customer” indicating the name of the customer as information for identifying the customer of the call center. "Name" is memorized. In addition, the customer table 412 stores a “story half flag” that is information indicating whether or not the customer is a half story customer in association with the customer code or the customer name. In the following, a case where the story half flag “1” indicates that the customer is a story half customer and the story half flag “null” indicates that the customer is not a story half customer will be described as an example.

  In the example illustrated in FIG. 5, the customer table 412 stores the story half flag “1” in association with the customer code “00000001” and the customer name “Fujitsu Hanako”. Further, the customer table 412 stores a story half flag “null” in association with the customer code “00000003” and the customer name “Fujitsu Hajime”. That is, for example, the customer table 412 stores that “Fujitsu Hanako” is a half-customer and “Fujitsu Hajime” is not a half-talker. Similarly, the customer table 412 stores a half-talk flag for other customer codes and customer names.

  Of the information stored in the customer table 412, the customer code and the customer name are registered by the operator when the customer calls the call center for the first time, for example. The talk half flag is input by the operator based on, for example, the judgment of the operator corresponding to the customer.

  Returning to the description of FIG. The distribution process control unit 510 is connected to the first storage unit 410. The distribution processing control unit 510 corresponds to an electronic circuit such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a central processing unit (CPU), and a micro processing unit (MPU).

  When a call is made to the call center, the distribution processing control unit 510 determines whether or not the call is from a half-talked customer. If the distribution processing control unit 510 determines that the call is from a half-talked customer, the distribution process control unit 510 distributes the call to a dedicated operator. For example, the distribution processing control unit 510 searches the customer table 412 using the customer code or customer name as the call originator as a search key, and based on whether the half talk flag is “1”, the half talk customer Determine if the call is from. If the half-talk flag is “1”, that is, if it is determined that the call is from a customer who is half-talk, the distribution processing control unit 510 distributes the call to a dedicated operator.

  A case where a call is made from “Fujitsu Hanako” will be described as an example. In this case, the distribution process control unit 510 searches the customer table 412 using “Fujitsu Hanako” as a search key, and acquires the talk half flag “1”. Here, since the story half flag of “Fujitsu Hanako” is “1”, the distribution processing control unit 510 determines that the customer is a half story customer. Then, the distribution processing control unit 510 refers to the operator table 411 and distributes a call from “Fujitsu Hanako” to an operator whose dedicated operator flag is “1”. For example, the distribution processing control unit 510 distributes the call from “Fujitsu Hanako” to the operator “Masako Araki”.

  Further, a case where a call is made from “Fujitsu Hajime” will be described as an example. In this case, the distribution processing control unit 510 searches the customer table 412 using “Fujitsu Hajime” as a search key, and acquires the talk half flag “null”. Here, since the story half flag of “Fujitsu Hajime” is “null”, the distribution processing control unit 510 determines that the customer is not a story half customer. Then, the distribution processing control unit 510 refers to the operator table 411 and distributes a call from “Fujitsu Hajime” to an operator whose dedicated operator flag is “null”. In other words, the distribution processing control unit 510 distributes calls to general operators. For example, the distribution processing control unit 510 distributes the call from “Fujitsu Hajime” to the operator “Yuzu Yamamoto”. In the above description, the case where the distribution processing control unit 510 distributes a call to an operator who is not a full-time operator when the customer is not a half-talked customer is described as an example, but the present invention is not limited to this. The telephone may be assigned to a dedicated operator. For example, when a call is received from a customer who is not registered in the customer table 412, the distribution processing control unit 510 performs the same processing as when the talk half flag “null” is acquired.

(Each part related to a series of processes to select a dedicated operator)
Each unit related to a series of processes for selecting a dedicated operator will be described. As will be described below, the operator selection device 200 calculates an analysis process for analyzing conversation voice data for each operator information using a predetermined analysis standard, and a value related to the operator's listening ability based on an analysis result of the analysis process. The calculation process is executed. Further, the operator selection device 200 executes a selection process for selecting a dedicated operator based on the calculation result.

  Specifically, the duplication count calculation unit 521, the ratio calculation unit 522, the excluded ratio calculation unit 523, and the Aichichi information calculation unit 524 of the selection processing control unit 520 each store the second storage unit 420. The analysis process is executed using the third storage unit 430, and the calculation process is executed using the third storage unit 430. Further, the dedicated operator selection unit 525 of the selection process control unit 520 executes the selection process using the processing results of the duplication number calculation unit 521 to the identification information calculation unit 524.

  Returning to the description of FIG. The second storage unit 420 is connected to the selection process control unit 520. The second storage unit 420 stores data used for analysis processing by the selection processing control unit 520. The second storage unit 420 corresponds to, for example, a semiconductor memory device such as a RAM, a ROM, or a flash memory, or a hard disk or an optical disk. In the example illustrated in FIG. 3, the second storage unit 420 includes a conversation voice table 421, a keyword table 422, a coefficient table 423, an excluded sentence table 424, and an Aizuchi table 425. In FIG. 3, for convenience of explanation, a plurality of conversation voice tables 421 and a plurality of keyword tables 422 are shown. However, the plurality of conversation voice tables 421 and the plurality of keyword tables 422 respectively indicate the same table.

  Each table included in the second storage unit 420 is used for calculation processing by the selection processing control unit 520. The details of the calculation process performed by the selection process control unit 520 will be described later, and a description thereof will be omitted here. The information stored in each table of the second storage unit 420 is input by the call center manager or the designer of the operator selection device 200 prior to the calculation process by the selection process control unit 520.

  As shown in FIG. 6, the conversation voice table 421 stores conversation voice data between a customer who uses a call center and an operator in association with operator information. FIG. 6 is a diagram illustrating an example of information stored in the conversation voice table according to the second embodiment. In the example shown in FIG. 6, the conversation voice table 421 is associated with “time” indicating the time when a call is made, and “operator code”, “operator name”, “customer code”, “customer name”, and “conversation”. Audio data "is stored.

  In the example shown in FIG. 6, the conversation voice table 421 is associated with the time “2009/12/08 08:10” in association with the operator code “OP0000003”, the operator name “Masako Araki”, and the customer code “00000001”. The customer name “Fujitsu Hanako” and conversation voice data “A” are stored. That is, for example, the conversation voice table 421 stores conversation voice data between “Masako Araki” and “Hanako Fujitsu”. The conversation voice table 421 also stores conversation voice data for other operator codes and operator names.

  The conversation voice data stored in the conversation voice table 421 is input to the conversation voice table 421 by the operator selection device 200 every time there is a call from a customer, for example. In the example shown in FIG. 6, the case where the customer code and the customer name are stored in association with the operator information has been described as an example. However, the present invention is not limited to this, and the customer code and the customer name are stored. You don't have to.

  As shown in FIG. 7, the keyword table 422 stores keywords used when an operator describes a product in association with a product code or a product name. FIG. 7 is a diagram illustrating an example of information stored in the keyword table in the second embodiment. In the example shown in FIG. 7, the keyword table 422 is associated with the product code “FA1001-01-030” and the product name “product A pack”, and the keywords “product A pack”, “black vinegar”, “black sesame”, and the like. Remember. That is, for example, the keyword table 422 stores “product A pack”, “black vinegar”, and “black sesame” as keywords for the product “product A”. Similarly, the keyword table 422 stores keywords for other products.

  As shown in FIG. 8, the coefficient table 423 stores coefficients that are data used by the selection processing control unit 520. Since how the coefficient is used will be described later, the description is omitted here. FIG. 8 is a diagram illustrating an example of information stored in the coefficient table according to the second embodiment.

  In the example illustrated in FIG. 8, the coefficient table 423 stores a “product code” and a “product name” indicating a series of alphanumeric characters for the operator selection device 200 to identify a product that is mail-ordered at a call center. The coefficient table 423 stores coefficients used by the selection processing control unit 520 in association with “product code” and “product name”. Further, the coefficient table 423 stores “product characteristics” indicating the properties and characteristics of the product and the category of the product in association with “product code” and “product name”. In addition, the coefficient table 423 is associated with “product code” and “product name”, and when the time required for explaining a predetermined product is “1”, the length of time required for explaining the associated product is long. The “speak length multiple” indicating the length is stored. In the example shown in FIG. 8, the value when the time required for the description of the product name “product A pack (drink once a day, 30)” is “1” is shown.

  In the example illustrated in FIG. 8, the coefficient table 423 is associated with the product code “FA1001-01-100” and the product name “product A”, and the product characteristic “food / normal” and the talk length multiple “0. 8 ”and the coefficient“ 125% ”are stored. That is, in the coefficient table 423, the category of “product A” is “food / ordinary”, and the time required for explanation of “product A” is “product A pack (drink once a day / 30)”. Remember that it is 0.8 ”times.

  The “talk length multiple” will be further described. The operator speaks not only about the description of the product but also about other things such as public talk. For this reason, the “speaking length multiple” is calculated by taking an average value or statistics of the time required for explaining the product and the time required for talking about other matters.

  The “coefficient” will be further described. The “coefficient” stored by the coefficient table 423 is the inverse of the value of the talk length multiple. That is, the “coefficient” is a value for correcting the time spoken by the operator in the conversation voice data to the time when the customer and the operator talk about the predetermined product. For example, the coefficient becomes a smaller value as the product takes a longer time for explanation, and becomes a larger value as the product takes a shorter time for explanation. As a result, the value obtained by multiplying the time spoken by the operator by the “coefficient” is the time assumed when the customer and the operator have a conversation about a predetermined product.

  As shown in FIG. 9, the excluded sentence table 424 stores fixed phrases that the operator speaks to the user. FIG. 9 is a diagram illustrating an example of information stored in the excluded text table in the second embodiment. The excluded sentence table 424 is also referred to as a “fixed phrase storage unit”. In the example shown in FIG. 9, the excluded sentence table 424 includes an “exclude sentence” indicating a fixed phrase in association with an “exclude sentence code” indicating a series of alphanumeric characters for the operator selection device 200 to identify the fixed phrase. Remember.

  In the example shown in FIG. 9, the excluded sentence table 424 stores an excluded sentence “Thank you for calling. This is the Xx call center” in association with the excluded sentence code “10”.

  As shown in FIG. 10, the word table 425 stores words and phrases used as words. FIG. 10 is a schematic diagram illustrating an example of information stored in the identification table according to the second embodiment. In the example illustrated in FIG. 10, the Aizuchi table 425 stores “Yes”, “That ’s right”, “I see,” “That ’s right,” “That ’s right”, and the like.

  Returning to the description of FIG. The third storage unit 430 is connected to the selection process control unit 520. The third storage unit 430 stores data used for calculation processing by the selection processing control unit 520. The third storage unit 430 corresponds to, for example, a semiconductor memory device such as a RAM, a ROM, or a flash memory, or a hard disk or an optical disk. In the example illustrated in FIG. 3, the third storage unit 430 includes a duplication count table 431, a ratio table 432, an exclusion time ratio table 433, an identification count table 434, and an expansion count table 435. A plurality of expansion number tables 436. The information stored in each table of the third storage unit 430 is input by the call center manager or the designer of the operator selection device 200 prior to the calculation process by the selection process control unit 520.

  As shown in FIG. 11, the duplication number table 431 stores “duplication number code” indicating a series of alphanumeric characters for the operator selection device 200 to identify the number of times the operator's voice and the customer's voice overlap. . Further, the duplication count table 431 stores “duplication count” indicating the number of times the operator's voice and the customer's voice overlap each other in association with the “duplication count code”. Further, the duplication count table 431 stores “score conversion%” in association with “duplication count code” and “duplication count”. Regarding the details of the “point conversion%”, after having explained all the tables of the third storage unit 430 based on the fact that “point conversion%” is stored in each table of the third storage unit 430, explain. FIG. 11 is a diagram illustrating an example of information stored in the duplication count table according to the second embodiment. The number of times the operator's voice and the customer's voice overlap in the conversation voice data is also referred to as “the number of times of duplication” or “the first number of times”.

  In the example illustrated in FIG. 11, the duplication count table 431 stores the duplication count “0 to 1” and the score conversion percentage “100%” in association with the duplication count code “10”. That is, the duplication count table 431 stores that the point conversion percentage is “100%” when the duplication count is “0 to 1”. Similarly, the number-of-duplications table 431 stores a point conversion% for other number of duplications.

  Here, the number of times of duplication will be described with reference to FIGS. 12A and 12B. FIG. 12A is a diagram illustrating a case where the operator's voice and the customer's voice do not overlap. FIG. 12B is a diagram illustrating a case where an operator's voice and a customer's voice overlap. 12A and 12B, the horizontal axis indicates a time axis. “T1” to “t4” indicate elapsed times when “t0” is the starting point. The double line indicates the time when the customer is speaking. A single line indicates the time the operator is speaking.

  In the example shown in FIG. 12A, there is “t2” that is the time when the operator starts talking after “t1” that is the time when the customer finishes talking. Further, there is “t4” when the customer starts talking after “t3” when the operator finishes talking. That is, in the example shown in FIG. 12A, the time when the customer is speaking and the time when the operator is speaking do not overlap, and the number of overlaps is “0”.

  In the example shown in FIG. 12B, there is “t2” that is the time when the operator starts speaking before “t1” that is the time when the customer finishes speaking. Further, there is “t4” when the customer starts talking before “t3” when the operator finishes talking. That is, in the example shown in FIG. 12B, the time when the customer is speaking and the time when the operator is speaking overlap from “t1” to “t2” and “t3” to “t4”. "become.

  As shown in FIG. 13, the ratio table 432 stores a “conversation ratio number code” indicating a series of alphanumeric characters for the operator selection device 200 to identify a ratio indicating the ratio between the operator time and the customer time. The operator time indicates the length of time spoken by the operator, and the customer time indicates the length of time spoken by the customer. Further, the ratio table 432 stores “ratio” and “point conversion%” in association with “conversation ratio number code”. FIG. 13 is a diagram illustrating an example of information stored by the ratio table in the second embodiment.

  In the example shown in FIG. 13, the ratio table 432 is associated with the conversation ratio number code “10”, the ratio indicating the customer time “61% or more” and the operator time “41% or less”, and the point conversion percentage “100”. % ". The customer time “61% or more” and the operator time “41% or less” are “61% or more” when the total time of the customer time and the operator time is “100%”. The operator time is “41% or less”. That is, the ratio table 432 stores that, in the total time of the customer time and the operator time, when the customer time is “61% or more”, the point conversion percentage is “100%”. Further, the ratio table 432 similarly stores the point conversion% for other ratios.

  Similarly to the ratio table 432, the exclusion ratio table 433 stores “ratio” and “point conversion%” in association with “conversation ratio number code”, as shown in FIG. 14. FIG. 14 is a diagram illustrating an example of information stored in the exclusion time ratio table according to the second embodiment.

  The difference between the ratio table 432 and the exclusion ratio table 433 will be briefly described. As will be described later, the data stored in the ratio table 432 is used by the ratio calculation unit 522 of the selection processing control unit 520, and the data stored in the exclusion ratio table 433 is used by the exclusion ratio calculation unit 523. Used.

  As shown in FIG. 15, the number table 434 for a number of times indicates a series of alphanumeric characters for the operator selection device 200 to identify the number of times for the number of times per predetermined time that the operator hits a number of times. "Number code" is stored. Further, the table for number of letters 434 stores “number of letters” and “point conversion percentage” in association with “number of letters”. FIG. 15 is a diagram for explaining an example of information stored by the identification number table in the second embodiment. The number of matches is also referred to as “second number of times”.

  In the example illustrated in FIG. 15, the number of times table 434 stores the number of times of “one time / minute” and the point conversion percentage “100%” in association with the number of codes “10”. That is, the number-of-earnings table 434 stores that the point conversion percentage is “100%” when the operator hits “one time / minute”. Similarly, the number-of-earnings table 434 stores the point conversion% for the other numbers of times.

  As shown in FIG. 16, the expansion number table 435 is associated with “expansion number code” and “point conversion” in association with an “expansion number code” indicating a series of alphanumeric characters for the operator selection device 200 to identify the expansion number. % ". Here, the number of expansions refers to the number of times that the customer has continuously spoken after the operator hit the phone, rather than the time that the customer has continuously spoken until the time the operator hit the phone. Show. The number of expansions is also referred to as “third number”. FIG. 16 is a diagram illustrating an example of information stored in the expansion number table according to the second embodiment.

  In the example shown in FIG. 16, the expansion number table 435 stores the expansion number “1” and the point conversion percentage “0%” in association with the expansion number code “10”. That is, the expansion number table 435 stores that the point conversion percentage is “0%” when the number of times that the customer time has increased before and after the operator hits the “a-time” is “1 time”. Similarly, the expansion number table 435 stores a point conversion% for other expansion times.

  Here, the number of expansions will be described with reference to FIGS. 17A and 17B. FIG. 17A is a diagram for explaining a case where the customer time becomes long before and after the operator hits a hit. FIG. 17B is a diagram illustrating a case where the customer time does not become long before and after the operator hits the door. In FIG. 17A and FIG. 17B, a double line shows the time when the customer is speaking. A single line indicates the time the operator is speaking. The horizontal axis indicates the time axis. “T1” to “t6” indicate elapsed times when “t0” is the starting point. In FIG. 17A and FIG. 17B, the case where the operator hits “t3” to “t4” will be described as an example.

  In the example shown in FIG. 17A, the customer times “t5” to “t6” after the operator hits are longer than the customer times “t1” to “t2” before the operator hits the hit. In other words, in the example shown in FIG. 17A, the customer time is long before and after the operator hits the alarm, and the number of expansions is “1”.

  In the example shown in FIG. 17B, the customer times “t5” to “t6” after the operator hits are shorter than the customer times “t1” to “t2” before the operator hits the hit. In other words, in the example shown in FIG. 17B, the customer time is not long before and after the operator hits the alarm, and the number of expansions is “0”.

  As shown in FIG. 18, the multiple expansion count table 436 is associated with a “multiple expansion count code” indicating a series of alphanumeric characters used by the operator selection device 200 to identify the multiple expansion counts. And “point conversion%” are stored. Here, the number of times of extension is the number of times that the customer continuously talks after the operator hits the phone more than twice the time the customer talks continuously until the operator hits the phone. The number of times it has become longer. FIG. 18 is a diagram illustrating an example of information stored in the multiple decompression count table 436 according to the second embodiment.

  In the example illustrated in FIG. 18, the multiple decompression count table 436 stores a plurality of decompression counts “1 or less” and a score conversion percentage “0%” in association with the multiple decompression count code “10”. That is, in the multiple expansion number table 436, when the number of times that the customer time is continuously increased by two or more times before and after the operator hits the “one time” is “1 time”, the point conversion percentage is “0%”. I remember that. Similarly, the multiple expansion count table 436 stores the point conversion% for the other multiple expansion counts.

  Here, the multiple expansion times will be described with reference to FIGS. 19A and 19B. FIG. 19A is a diagram for explaining a case where the customer time is continuously increased twice or more before and after the operator hits a hit. FIG. 19B is a diagram for explaining a case where the customer time does not increase continuously twice or more before and after the operator hits the door. In FIG. 19A and FIG. 19B, a double line shows the time when the customer is speaking. A single line indicates the time the operator is speaking. The horizontal axis indicates the time axis. “T1” to “t10” indicate elapsed time with “t0” as a starting point. In FIG. 19A and FIG. 19B, a case where the operator hits “t3” to “t4” and “t7” to “t8” will be described as an example.

  In the example shown in FIG. 19A, the customer times “t5” to “t6” after the operator hits the time are longer than the customer times “t1” to “t2” before the operator hits the name. Further, the customer times “t9” to “t10” after the operator hits the head are longer than the customer times “t5” to “t6” before the operator hits the head. In other words, in the example shown in FIG. 19A, the customer time is continuously longer than two times before and after the operator hits the alarm, and the number of multiple expansions is “1”.

  In the example illustrated in FIG. 19B, the customer times “t5” to “t6” after the operator hits the time are longer than the customer times “t1” to “t2” before the operator hits the name. In addition, customer times “t9” to “t10” after the operator hits the head are shorter than customer times “t5” to “t6” before the operator hits the head. That is, in the example shown in FIG. 19B, the customer time is not continuously increased two or more times before and after the operator hits the alarm, and the number of multiple expansions is “0”.

  The “point conversion%” stored in each table of the third storage unit 430 will be described. “Point conversion%” is used when the selection processing control unit 520 calculates a value related to the operator's listening skill. The selection processing control unit 520 calculates a larger value as the value related to the operator's listening ability as the “point conversion%” is larger, and the smaller the “point conversion%” as the value related to the operator's listening ability. Calculate the value. The larger the value related to the operator's listening ability, the more the operator is listening. The smaller the operator's listening value is, the less the operator is.

  Here, it is considered that the smaller the number of times of duplication, that is, the fewer the number of times the operator speaks when the customer is speaking, the better the operator is. In view of this, the duplication count table 431 stores a larger value as “point conversion%” as the number of duplication is smaller, and stores a smaller value as “point conversion%” as the number of duplication increases. To do.

  In addition, it can be considered that the greater the proportion of customer time in the ratio, that is, the longer the customer is speaking, the better the operator is. Based on this, the ratio table 432 and the exclusion ratio table 433 store a larger value as “point conversion%” as the ratio of the customer time in the ratio is larger. Further, the ratio table 432 and the exclusion ratio table 433 store a smaller value as “point conversion%” as the ratio of the customer time in the ratio is smaller.

  In addition, it is considered that the customer can continue to talk comfortably when hit by the operator. As a result, it is considered that the more the number of letters is, the better the operator is. Based on this fact, the number of matches in the table 434 stores a larger value as the “score conversion%” as the number of matches increases, and the smaller the number of matches, the smaller as the “point conversion%”. Store the value.

  In addition, when the customer has spoken for a longer time than before hitting the operator, the operator has successfully hit the user and the operator is considered to be a good listener. As a result, the greater the number of expansions and the number of expansions, the better the operator is. Based on this, the expansion number table 435 and the multiple expansion number table 436 store a larger value as “point conversion%” as the expansion number and the plurality of expansion times increase. Further, the expansion number table 435 and the multiple expansion number table 436 store smaller values as “point conversion%” as the expansion number and the plurality of expansion times are smaller.

  Returning to the description of FIG. The selection processing control unit 520 is connected to the first storage unit 410, the second storage unit 420, and the third storage unit 430. The selection processing control unit 520 corresponds to an electronic circuit such as ASIC, FPGA, CPU, MPU, for example. In the example shown in FIG. 3, the selection processing control unit 520 includes an overlap number calculation unit 521, a ratio calculation unit 522, an exclusion time ratio calculation unit 523, an aiding information calculation unit 524, and a dedicated operator selection unit 525. Have.

  As will be described below, the duplication count calculation unit 521, the ratio calculation unit 522, the exclusion time ratio calculation unit 523, and the identification information calculation unit 524 execute an analysis process and a calculation process, respectively. That is, each of the duplication number calculation unit 521 to the identification information calculation unit 524 analyzes the conversation voice data stored in the conversation voice table 421 for each operator information using a predetermined analysis standard. Then, each of the duplication count calculation unit 521 to the identification information calculation unit 524 calculates a value related to the operator's listening skill based on the analysis result.

  Below, the case where weighting is performed with respect to the calculation result by the duplication number calculation part 521-the identification information calculation part 524 is demonstrated to an example. Specifically, a case will be described as an example in which a score is set in advance for each type of calculation result in the duplication count calculation unit 521 to the identification information calculation unit 524. Here, the preset score is a value used when a value related to the operator's listening skill is calculated. As a result, as will be described later, the duplication count calculation unit 521 to the identification information calculation unit 524 each of the operator based on the score given in advance and the score conversion percentage read from the third storage unit 430. Calculate a value for listening. In the following, the case of performing weighting will be described, but the present invention is not limited to this. For example, each of the calculation results by the duplication count calculation unit 521 to the identification information calculation unit 524 may be treated equally.

  The duplication number calculation unit 521 analyzes the conversational voice data using the duplication number as a predetermined analysis criterion, and calculates a value related to the operator's listening skill. Below, the analysis process and calculation process by the duplication number calculation part 521 are demonstrated in order.

  The analysis process by the duplication number calculation unit 521 will be described. The duplication number calculation unit 521 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key. Then, the duplication count calculation unit 521 identifies the customer's voice and the operator's voice included in the conversation voice data, thereby identifying the time when the customer is speaking, as shown in FIG. 12A and FIG. Identify the time when is talking. And the duplication number calculation part 521 extracts the duplication location where the time when the customer is speaking and the time when the operator is talking overlaps on the time axis, and counts the number of the duplication portions detected, thereby duplicating Calculate the number of times. For example, the overlapping portion corresponds to “t1” to “t2” and “t3” to “t4” in the example illustrated in FIG. 12B. And the duplication number calculation part 521 calculates duplication number "2" in the example shown to FIG. 12B.

  Note that the duplication count calculation unit 521 separates the customer's voice and the operator's voice by using, for example, the fact that the frequency of the customer's voice and the frequency of the operator's voice are different. The method for separating the customer's voice and the operator's voice is not limited to the method using the difference in voice frequency, and a known method may be used.

  The calculation process by the duplication number calculation unit 521 will be described. The overlapping number calculation unit 521 calculates a larger value as a value indicating the listening ability as the number of overlapping times obtained as an analysis result is smaller, and a smaller value as the value indicating the listening ability as the first number is larger. Calculate the value. For example, the duplication count calculation unit 521 acquires the point conversion% from the duplication count table 431 using the duplication count obtained as an analysis result as a search key. More specifically, when the number of duplication times “0” is obtained as the analysis result, the duplication number calculation unit 521 acquires the point conversion% “100%” corresponding to the duplication number “0”. Here, for example, when “40” is set as the score for the number of times of duplication, the number of times of duplication calculation unit 521 multiplies “40” by “100%”, thereby obtaining a value “ 40 "is calculated.

  The ratio calculation unit 522 analyzes the conversation voice data using the ratio as a predetermined analysis standard, and calculates a value related to the operator's listening skill. Below, the analysis process and calculation process by the ratio calculation part 522 are demonstrated in order.

  An analysis process performed by the ratio calculation unit 522 will be described. The ratio calculation unit 522 acquires conversation voice data from the conversation voice table 421 using the operator information on the operator to be processed as a search key. Then, the ratio calculation unit 522 identifies the customer's voice and the operator's voice included in the conversation voice data, thereby identifying the time when the customer is speaking, as shown in FIG. 12A and FIG. Identify the speaking time. Then, the ratio calculation unit 522 calculates customer time and operator time.

  In addition, when the read product is a product that requires time for explanation as compared with other products, the ratio calculation unit 522 calculates the ratio after correcting the operator time to a smaller time. On the other hand, if the read product is a product that does not require time for explanation as compared with other products, the ratio calculation unit 522 calculates the ratio after correcting the operator time to a larger time.

  Specifically, the ratio calculation unit 522 identifies the product that is the subject of conversation in the conversation voice data, corrects the operator time according to the identified product, and calculates the ratio using the corrected operator time. To do. For example, the ratio calculation unit 522 extracts a keyword from character data obtained by executing character recognition processing on conversation voice data, and reads a product corresponding to the extracted keyword from the keyword table 422. For example, the case where the keyword “black vinegar” is obtained as a result of performing character recognition processing on conversational voice data will be described. The ratio calculation unit 522 reads the product code “FA1001-01-030” product name “product A pack (drink once a day, 30 bottles)” from the keyword table 422 using the keyword “black vinegar” as a search key. That is, the ratio calculation unit 522 identifies that the product that is the subject of conversation is “Product A pack (drink once per day / 30)”.

  Further, for example, the ratio calculating unit 522 uses the read product code “FA1001-01-030” product name “product A pack (drink once a day, 30 bottles)” as a search key, and uses the coefficient “100” from the coefficient table 423. % ". Then, the ratio calculation unit 522 corrects the operator time using the coefficient “100%” acquired from the coefficient table 423. More specifically, the ratio calculation unit 522 multiplies the operator time by a coefficient “100%” to obtain the corrected operator time. That is, the ratio calculation unit 522 corrects the operator time according to the identified “product A pack (drink once a day / 30 bottles)”.

  For example, the ratio calculation unit 522 calculates the ratio using the corrected operator time. That is, the ratio calculation unit 522 calculates the ratio of the customer time and the ratio of the operator time when the total time of the customer time and the operator time is “100”. For example, the ratio calculation unit 522 calculates the customer time “55%” and the operator time “45%”.

  Here, as described above, the coefficient becomes a larger value as the product takes a longer time for explanation, and becomes a smaller value as the product takes a shorter time for explanation. As a result, the corrected operator time is corrected to a smaller time when the product is a product that requires time for explanation compared to other products, and the product is time for explanation compared to other products. If the product does not need to be corrected, it is corrected to a larger time.

  The correction process by the ratio calculation unit 522 will be further described with reference to FIG. FIG. 20 is a diagram illustrating the ratio calculation unit according to the second embodiment. Reference numeral 701 in FIG. 20 indicates a ratio between the customer time and the operator time before correction. Reference numeral 702 denotes a ratio between the corrected operator time and customer time in the case of a product that requires time to explain compared to other products. Reference numeral 703 denotes a ratio between the corrected operator time and the customer time in the case where the product does not require time for explanation as compared with other products. Further, 701a, 702a, and 703a indicate operator time, and 701b, 702b, and 703b indicate customer time, respectively.

  If the ratio calculation unit 522 is a product that requires time to explain compared to other products, the ratio calculation unit 522 calculates the ratio after correcting the operator time to be small. As a result, as indicated by 702 in FIG. 20, the ratio of the operator time in the ratio is smaller than that in 701 in FIG. In addition, when the ratio calculation unit 522 is a product that does not require time for explanation as compared with other products, the ratio calculation unit 522 calculates the ratio after correcting the operator time to increase. . As a result, as indicated by reference numeral 703 in FIG. 20, the ratio of the operator time in the ratio is larger than that in FIG.

  A calculation process by the ratio calculation unit 522 will be described. The ratio calculation unit 522 calculates a larger value as a value indicating the better listening ability as the ratio of the customer time in the ratio obtained as the analysis result is larger. In addition, the ratio calculation unit 522 calculates a smaller value as a value indicating the listening skill as the ratio of the customer time in the ratio is smaller.

  For example, the case where the customer time “51% to 60%” and the operator time “41% to 50%” are calculated as an analysis result will be described. In this case, the ratio calculation unit 522 acquires the point conversion percentage “80%” from the ratio table 432 using the customer time “51% to 60%” and the operator time “41% to 50%” as a search key. Here, for example, when “30” is set as the score for the ratio, the ratio calculation unit 522 multiplies “30” by “80%”, thereby obtaining a value “24” relating to the operator's listening skill. Is calculated.

  The excluded ratio calculation unit 523 analyzes the conversation voice data using the ratio as a predetermined analysis criterion, and calculates a value related to the operator's listening skill. Hereinafter, the analysis process and the calculation process performed by the exclusion ratio calculation unit 523 will be described in order.

  An analysis process performed by the exclusion ratio calculation unit 523 will be described. The exclusion ratio calculation unit 523 executes a process using the excluded sentence table 424 and a process using the keyword table 422, as described below.

  Processing using the excluded text table 424 will be described. The excluded time ratio calculation unit 523 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key. Then, the excluded time ratio calculation unit 523 extracts a portion that matches the fixed phrase stored in the excluded sentence table 424 from the character data obtained by executing the character recognition process on the conversational voice data. Then, the exclusion ratio calculation unit 523 deletes the conversation voice data corresponding to the extracted character data, and calculates the ratio using the deleted conversation voice data.

  As for the cliché, any operator can speak. For this reason, it is considered that the difference between operators becomes difficult to see when the part where the phrase is spoken is taken into consideration. Based on this, the excluded ratio calculation unit 523 calculates the ratio after deleting the time during which the phrase is spoken. In the following, the ratio calculated using the excluded sentence table 424 is also referred to as a “fixed phrase exclusion ratio”.

  For example, explanation will be given using a case where the word data includes the phrase “Payment can be made at post office”. In this case, the excluded ratio calculation unit 523 extracts “payment can be made at the post office for payment” from the character data. Then, the excluded ratio calculation unit 523 deletes the portion corresponding to the extracted “payment can be made at the post office” from the conversation voice data. Then, the exclusion ratio calculation unit 523 calculates the ratio using the conversation voice data after the portion corresponding to the fixed phrase such as “Payment can be made at the post office can be used” is deleted. By calculating the ratio in the same manner as the part 522, the ratio at the time of exclusion of the fixed phrase is calculated.

  Processing using the keyword table 422 will be described. The excluded time ratio calculation unit 523 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key. Then, the excluded time ratio calculation unit 523 extracts a portion that matches the keyword stored in the keyword table 422 from the character data obtained by executing the character recognition process on the conversation voice data. Then, the exclusion ratio calculation unit 523 deletes the conversation voice data corresponding to the extracted character data, and calculates the ratio using the deleted conversation voice data.

  With regard to merchandise keywords, it is considered that any operator can make a continuous call. For this reason, it is considered that the difference between operators becomes difficult to see when the portion where the keyword is spoken is considered. Based on this, the excluded ratio calculation unit 523 calculates the ratio after deleting the time when the keyword is spoken. Hereinafter, the ratio calculated using the keyword table 422 is also referred to as a “keyword exclusion ratio”.

  For example, the case where the keyword “black sesame” is included in the character data will be described. In this case, the excluded ratio calculation unit 523 extracts “black sesame” from the character data. Then, the exclusion ratio calculation unit 523 deletes the data corresponding to the extracted “black sesame” from the conversational voice data. Then, the exclusion ratio calculation unit 523 calculates the ratio in the same manner as the ratio calculation unit 522 by using the conversation voice data after the part corresponding to the keyword such as “black sesame” is deleted, thereby removing the keyword. Calculate the duty ratio.

  The calculation process by the exclusion ratio calculation unit 523 will be described. The exclusion ratio calculation unit 523 calculates a larger value as a value indicating the better listening ability as the ratio of the customer time in the fixed phrase exclusion ratio or the keyword exclusion ratio obtained as the analysis result is larger. In addition, the ratio calculation unit 522 and the exclusion ratio calculation unit 523 calculate a smaller value as a value indicating the better listening ability as the ratio of the customer time in the fixed phrase exclusion ratio and the keyword exclusion ratio is smaller.

  For example, the case where the customer time “65%” and the operator time “35%” are calculated as an analysis result will be described. In this case, the ratio calculation unit 522 acquires the point conversion percentage “100%” from the exclusion ratio table 433 using the customer time “65%” and the operator time “35%” as search keys. Here, for example, when “10” is set as the score for the ratio, the ratio calculation unit 522 multiplies “10” by “100%” to thereby obtain a value “10” regarding the operator's listening skill. Is calculated.

  As will be described below, the AIZU information calculation unit 524 executes processing using the AICH number, the number of expansions, and the number of expansions as predetermined analysis criteria. Hereinafter, processing using the number of hits, the number of expansions, and the number of expansions as a predetermined analysis criterion will be described in order.

  A process using the number of identifications as a predetermined analysis criterion will be described. The nickname information calculation unit 524 analyzes the conversational voice data using the number of nicknames as a predetermined analysis criterion, and calculates a value related to the operator's listening skill. Below, the analysis process and calculation process by the identification information calculation part 524 when the number of identifications is used will be described in order.

  An analysis process when the number of words is used will be described. The nickname information calculation unit 524 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key. Then, the nickname information calculation unit 524 extracts a portion that matches the phrase stored in the nickname table 425 from the character data obtained by executing the character recognition process on the conversation voice data. Then, the identification information calculation unit 524 calculates the number of times the identification is used by counting the number of extracted parts. Then, the nickname information calculation unit 524 calculates the number of nicknames by dividing the number of times nickname is used by the length of the conversational voice data.

  For example, a case will be described as an example where the nickname information calculation unit 524 uses “12 times” the number of times nickname is used and the length of the conversational voice data is “4 minutes”. In this case, the identification information calculation unit 524 calculates the number of identifications “3 times / minute” by dividing “12 times” by “4 minutes”.

  A calculation process when the number of words is used will be described. The Aichichi information calculation unit 524 calculates a larger value as the value indicating the better listening ability as the number of the obtained identification numbers is larger, and as the value indicating the better listening ability as the number of Aichichi numbers is smaller. Calculate a small value.

  For example, the case where “3 times / minute” is calculated as the analysis result will be described. In this case, the AIZU information calculating unit 524 acquires the score conversion percentage “100%” from the AIZUKI number table 434 using “3 times / min” as a search key. Here, for example, when “5” is set as the score for the number of hits, the ratio calculation unit 522 multiplies “5” by “100%” to obtain a value “ 5 "is calculated.

  Processing using the number of expansions as a predetermined analysis criterion will be described. The nickname information calculation unit 524 analyzes the conversation voice data using the number of expansions as a predetermined analysis criterion, and calculates a value related to the operator's listening skill. Hereinafter, the analysis process and the calculation process by the identification information calculation unit 524 when the number of expansions is used will be described in order.

  An analysis process when the number of expansions is used will be described. The nickname information calculation unit 524 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key. Then, the nickname information calculation unit 524 extracts a place where the word / phrase stored in the nickname table 425 is used from the character data obtained by executing the character recognition process on the conversation voice data. In addition, the identification information calculation unit 524 identifies the voice of the customer and the voice of the operator included in the conversational voice data, and as shown in FIG. 17A and FIG. Identify time. Then, the identification information calculation unit 524 compares the customer times before and after the location where the identification is detected, and calculates the number of expansions by counting the number of locations where the customer time has increased after identification.

  To explain using the example shown in FIG. 17B, the Aichi information calculation unit 524 has the customer times “t1” to “t2” and the customer time “t5” before and after “t3” to “t4” where the Aichi is detected. ~ "T6" is calculated. Then, the identification information calculation unit 524 compares the customer times “t1” to “t2” with the customer times “t5” to “t6” to determine whether the customer time has increased after the identification. In the example illustrated in FIG. 17B, the customer times “t5” to “t6” are longer than the customer times “t1” to “t2”, and the nickname information calculation unit 524 determines that the customer times “t5” to “t6” have become longer. Also, the Aizuchi information calculation unit 524 makes the same determination for other locations. Then, the identification information calculation unit 524 calculates the number of expansions by counting the number determined to be longer. For example, the Aizuchi information calculation unit 524 calculates the number of expansions “3 times”.

  A calculation process when the number of expansions is used will be described. The Aichichi information calculation unit 524 calculates a larger value as a value indicating the listening ability as the number of expansions obtained as an analysis result is larger, and a smaller value as a value indicating the listening ability as the number of expansions is smaller. Is calculated.

  For example, the case where the number of expansions “3 times” is calculated as an analysis result will be described. In this case, the identification information calculation unit 524 acquires the point conversion% “100%” from the expansion number table 435 using the expansion number “3 times” as a search key. Here, for example, when “10” is set as the score for the number of hits, the ratio calculation unit 522 multiplies “10” by “100%” to obtain a value “ 10 "is calculated.

  Processing using a plurality of expansion times as a predetermined analysis criterion will be described. The nickname information calculation unit 524 analyzes the conversation voice data using a plurality of expansion times as a predetermined analysis standard, and calculates a value related to the operator's listening skill. Hereinafter, the analysis process and the calculation process performed by the identification information calculation unit 524 when a plurality of expansion times are used will be described in order.

  An analysis process when using a plurality of expansion times will be described. The nickname information calculation unit 524 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key. Then, the nickname information calculation unit 524 extracts a place where the word / phrase stored in the nickname table 425 is used from the character data obtained by executing the character recognition process on the conversation voice data. Further, the nickname information calculation unit 524 identifies the voice of the customer and the voice of the operator included in the conversation voice data, and the operator is talking about the time the customer is talking as shown in FIGS. 19A and 19B. Identify time. Then, the nickname information calculation unit 524 compares the customer time before and after the location where the nickname is detected, so that the customer continuously speaks twice or more after the nickname is hit. To determine if it has become longer. Then, the identification information calculation unit 524 calculates the number of expansions by counting the number of locations determined to be longer.

  To explain with reference to the example shown in FIG. 19A, the Aichi information calculation unit 524 calculates customer times before and after “t3” to “t4” and “t7” to “t8” where the Aichi is detected. For example, the Aizuchi information calculation unit 524 calculates customer times “t1” to “t2” and customer times “t5” to “t6” before and after “t3” to “t4”, and “t7” to “t6”. Customer times “t5” to “t6” and customer times “t9” to “t10” before and after t8 are calculated. Then, the identification information calculation unit 524 compares the customer times “t1” to “t2” with the customer times “t5” to “t6” to determine whether the customer time has increased after the identification. In the example illustrated in FIG. 19A, the customer times “t5” to “t6” are longer than the customer times “t1” to “t2”, and the Azuchi information calculation unit 524 determines that it has become longer. Here, the contact information calculation unit 524 further compares the customer times “t5” to “t6” with the customer times “t9” to “t10”, so that the customer time after two consecutive contact times. Determine if it has become longer. In the example illustrated in FIG. 19A, the customer times “t9” to “t10” are longer than the customer times “t5” to “t6”, and the Azuchi information calculation unit 524 determines that the customer times “t9” to “t10” have been continuously increased twice. . Also, the Aizuchi information calculation unit 524 makes the same determination for other locations. Then, the identification information calculation unit 524 calculates the number of multiple expansion times by counting the number determined to be longer. For example, the Aizuchi information calculation unit 524 calculates a plurality of expansion times “2 times”.

  A calculation process when a plurality of expansion times are used will be described. The Aichichi information calculation unit 524 calculates a larger value as a value indicating the listening ability as the number of multiple expansions obtained as an analysis result is larger, and as a value indicating the listening ability as the number of multiple expansions is smaller. Calculate a small value.

  For example, the case where a plurality of expansion times “2” is calculated as an analysis result will be described. In this case, the matching information calculation unit 524 acquires the point conversion% “80%” from the multiple expansion count table 436 using the multiple expansion count “2” as a search key. Here, for example, when “5” is set as the score for the number of hits, the ratio calculation unit 522 multiplies “5” by “80%” to obtain a value “ 4 "is calculated.

  The dedicated operator selection unit 525 executes selection processing based on the values calculated by the duplication number calculation unit 521 to the identification information calculation unit 524. For example, the dedicated operator selection unit 525 adds each of the values calculated by the duplication number calculation unit 521 to the identification information calculation unit 524. Then, the dedicated operator selection unit 525 compares the value obtained by the addition with a predetermined threshold value, and if the value obtained by the addition is larger than the predetermined threshold value, selects the operator as the predetermined target as the dedicated operator. To do.

  The dedicated operator selection unit 525 will be described using the values shown in FIG. FIG. 21 is a diagram illustrating an example of values used for explaining the dedicated operator selection unit 525 in the second embodiment. A to F in FIG. 21 indicate the number of times of duplication, the ratio, the ratio at the time of exclusion of a fixed phrase, the ratio at the time of keyword exclusion, the number of rounds, the number of expansions, and the number of expansions, respectively. In FIG. 21, “full score” indicates a preset score, and “score” indicates a value calculated by the duplication count calculation unit 521 to the identification information calculation unit 524.

  In the example illustrated in FIG. 21, the dedicated operator selection unit 525 calculates “91” by adding each of the values calculated by the duplication number calculation unit 521 to the identification information calculation unit 524. Here, for example, when “90” is set as the predetermined threshold, the dedicated operator selection unit 525 determines that the value “91” obtained by the addition is larger than the predetermined threshold “100” and The selected operator is selected as a full-time operator. For example, the dedicated operator selection unit 525 sets the dedicated operator flag of the operator selected as the dedicated operator among the operators stored in the operator table 411 to “1”.

[Processing by distribution processing control unit]
With reference to FIG. 22, an example of the flow of distribution processing by the distribution processing control unit 510 according to the second embodiment will be described. FIG. 22 is a flowchart illustrating an example of the flow of the distribution process performed by the distribution process control unit according to the second embodiment.

  As shown in FIG. 22, when a call is made to the call center (Yes at Step S101), the distribution processing control unit 510 determines whether the call is from a half-talked customer (Step S102). For example, the distribution processing control unit 510 searches the customer table 412 using the customer code or customer name as the call originator as a search key, and determines whether the talk half flag is “1”.

  Here, if the distribution processing control unit 510 determines that the call is from a half-speak customer (Yes at step S102), that is, if the half-story flag is “1”, the call is distributed to a dedicated operator ( Step S103). On the other hand, if the distribution processing control unit 510 determines that the call is not from a half-speak customer (No at Step S102), that is, if the half-speak flag is “null”, the call is distributed to a general operator (Step S102). S104).

[Overall processing by selection processing control unit]
With reference to FIG. 23, an example of the flow of overall processing by the selection processing control unit 520 in the second embodiment will be described. FIG. 23 is a flowchart illustrating an example of the flow of overall processing performed by the selection processing control unit according to the second embodiment.

  As shown in FIG. 23, in the selection process control unit 520, when it is time to select a dedicated operator (Yes at Step S201), the process by the duplication number calculation unit 521 is executed (Step S202), and the process by the ratio calculation unit 522 is performed. It is executed (step S203). Then, in the selection process control unit 520, the process by the exclusion time ratio calculation unit 523 is executed (step S204), and the process by the identification information calculation unit 524 is executed (step S205). That is, in the selection process control unit 520, the duplication count calculation unit 521, the ratio calculation unit 522, the exclusion time ratio calculation unit 523, and the Aichichi information calculation unit 524 execute the analysis process and the calculation process, respectively. .

  Then, the selection process control unit 520 executes a selection process (step S206). Note that the above-described processing procedure is not limited to the above-described order, and may be appropriately changed within a range that does not contradict the processing contents. For example, in the example shown in FIG. 23, the case where steps S202 to S204 are executed in order is shown as an example, but the present invention is not limited to this. For example, the order of steps S202 to S204 may be arbitrarily changed or may be executed in parallel.

[Processing by duplication count calculation unit]
With reference to FIG. 24, an example of the flow of processing by the duplication number calculation unit 521 in the second embodiment will be described. FIG. 24 is a flowchart illustrating an example of a flow of processing performed by the duplication number calculation unit according to the second embodiment. Note that a series of processes described with reference to FIG. 24 corresponds to step S202 in FIG.

  As illustrated in FIG. 24, the duplication count calculation unit 521 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key (step S301). Then, the duplication count calculating unit 521 identifies the customer's voice and the operator's voice included in the conversation voice data, thereby identifying the time when the customer is speaking as shown in FIG. 12A and FIG. S302), the time when the operator is speaking is identified (step S303).

  Then, the number-of-duplications calculation unit 521 extracts an overlapping portion where the time when the customer is speaking and the time when the operator is speaking overlap on the time axis (step S304). For example, in the example illustrated in FIG. 12B, “t1” to “t2” and “t3” to “t4” are detected.

  Then, the number of times of duplication is calculated by counting the number of detected duplication points (step S305). For example, the duplication count calculation unit 521 calculates the duplication count “2” in the example illustrated in FIG. 12B.

  Then, the duplication count calculation unit 521 calculates a value related to the operator's listening skill (step S306). For example, when the duplication count “0” is obtained as an analysis result, the duplication count calculation unit 521 converts the score from the duplication count table 431 using the duplication count “0” as a search key. % "100%" is acquired. Here, for example, when “40” is set as the score for the number of times of duplication, the number of times of duplication calculation unit 521 multiplies “40” by “100%”, thereby obtaining a value “ 40 "is calculated.

  Note that the above-described processing procedure is not limited to the above-described order, and may be appropriately changed within a range that does not contradict the processing contents. For example, the order of step S302 and step S303 described above may be interchanged, or may be executed in parallel.

[Processing by ratio calculation unit]
An example of the flow of processing performed by the ratio calculation unit 522 in the second embodiment will be described with reference to FIG. FIG. 25 is a flowchart illustrating an example of a process flow by the ratio calculation unit according to the second embodiment. A series of processing described with reference to FIG. 25 corresponds to step S203 in FIG.

  As shown in FIG. 25, the ratio calculation unit 522 acquires conversation voice data from the conversation voice table 421 using the operator information about the operator to be processed as a search key (step S401). Then, the ratio calculation unit 522 identifies the customer's voice and the operator's voice included in the conversation voice data, thereby identifying the time when the customer is speaking as shown in FIGS. 12A and 12B (step S402). ), The time when the operator is speaking is identified (step S403). Then, the ratio calculation unit 522 calculates customer time and operator time (step S404).

  Then, the ratio calculation unit 522 performs a character recognition process on the conversation voice data (step S405), and extracts a keyword from the character data obtained by executing the character recognition process (step S406). For example, the ratio calculation unit 522 extracts the keyword “black vinegar”.

  Then, the ratio calculation unit 522 reads out a product corresponding to the extracted keyword from the keyword table 422 (step S407). For example, the ratio calculation unit 522 reads the product code “FA1001-01-030” product name “product A pack (drink once a day / 30)” from the keyword table 422 using the keyword “black vinegar” as a search key. .

  Then, the ratio calculation unit 522 corrects the operator time (step S408). That is, for example, the ratio calculation unit 522 uses the read product code “FA1001-01-030” product name “product A pack (drink once a day, 30 bottles)” as a search key, and uses the coefficient “100” from the coefficient table 423. % ". Then, the ratio calculation unit 522 corrects the operator time using the coefficient “100%”. More specifically, the ratio calculation unit 522 multiplies the operator time by a coefficient “100%” to obtain the corrected operator time.

  Then, the ratio calculation unit 522 calculates the ratio using the corrected operator time (step S409). That is, the ratio calculation unit 522 calculates the ratio of the customer time and the ratio of the operator time when the total time of the customer time and the operator time is “100”. For example, the ratio calculation unit 522 calculates the customer time “55%” and the operator time “45%”.

  Then, the ratio calculation unit 522 calculates a value indicating the listening skill using the ratio obtained as the analysis result (step S410). For example, the ratio calculation unit 522 acquires the point conversion percentage “80%” from the ratio table 432 using the customer time “55%” and the operator time “45%” as search keys. Here, for example, when “30” is set as the score for the ratio, the ratio calculation unit 522 multiplies “30” by “80%”, thereby obtaining a value “24” relating to the operator's listening skill. Is calculated.

  Note that the above-described processing procedure is not limited to the above-described order, and may be appropriately changed within a range that does not contradict the processing contents. For example, a series of processes from steps S405 to S407 may be executed in parallel with the series of processes from steps S402 to S404 described above.

[Processing by exclusion ratio calculation unit]
With reference to FIG. 26, an example of the processing flow by the exclusion ratio calculation unit 523 according to the second embodiment will be described. FIG. 26 is a flowchart for explaining an example of a flow of processing by the excluded ratio calculation unit in the second embodiment. Below, the process flow using the excluded sentence table 424 among the processes by the ratio calculation part 523 at the time of exclusion is demonstrated.

  A series of processing described with reference to FIG. 26 corresponds to step S204 in FIG. Step S501 in FIG. 26 corresponds to step S401 in FIG. Step S505 in FIG. 26 corresponds to steps S402 to S404 and S409 in FIG. Also, step S506 in FIG. 26 corresponds to step S409 in FIG.

  As shown in FIG. 26, the excluded-time ratio calculation unit 523 acquires conversation voice data from the conversation voice table 421 (step S501). Then, the excluded ratio calculation unit 523 executes character recognition processing on the conversational voice data (step S502). Then, the excluded time ratio calculation unit 523 extracts a portion that matches the fixed phrase stored in the excluded sentence table 424 from the character data obtained by executing the character recognition process (step S503). For example, explanation will be given using a case where the word data includes the phrase “Payment can be made at post office”. In this case, the excluded ratio calculation unit 523 extracts “payment can be made at the post office for payment” from the character data.

  Then, the exclusion ratio calculation unit 523 deletes the portion of the conversation voice data corresponding to the extracted character data (step S504), and calculates the fixed phrase exclusion ratio using the deleted conversation voice data (step S504). S505). For example, the excluded ratio calculation unit 523 uses the conversation voice data after the part that is talking about the phrase such as “Payment can be made at the post office is used” as the ratio. By calculating the ratio in the same manner as the calculation unit 522, the fixed phrase exclusion ratio is calculated. For example, the excluded time ratio calculation unit 523 calculates the customer time “65%” and the operator time “35%”.

  Then, the exclusion ratio calculation unit 523 calculates a value indicating the listening skill using the fixed phrase exclusion ratio obtained as the analysis result (step S506). For example, the ratio calculation unit 522 acquires the point conversion percentage “100%” from the exclusion ratio table 433 using the customer time “55%” and the operator time “45%” as search keys. Here, for example, when “10” is set as the score for the ratio, the ratio calculation unit 522 multiplies “10” by “100%” to thereby obtain a value “10” regarding the operator's listening skill. Is calculated.

  Note that the above-described processing procedure is not limited to the above-described order, and may be appropriately changed within a range in which the processing contents do not contradict each other as in FIG. An example of the flow of processing when the exclusion ratio calculation unit 523 performs processing using the keyword table 422 will be briefly described. In this case, the exclusion-time ratio calculation unit 523 extracts a portion that matches the keyword stored in the keyword table 422 from the character data in step S503 of FIG.

[Processing by Aizuchi Information Calculation Unit]
With reference to FIG. 27, an example of the flow of processing by the identification information calculation unit 524 in the second embodiment will be described. FIG. 27 is a flowchart for describing an example of a flow of processing performed by the identification information calculation unit according to the second embodiment. A series of processing described with reference to FIG. 27 corresponds to step S205 in FIG. In FIG. 27, the identification information calculation unit 524 uses a process using the number of identifications as a predetermined analysis criterion, a process using the number of expansions as a predetermined analysis criterion, and a plurality of expansion times as a predetermined analysis criterion. An example of the flow of processing when processing is executed continuously will be described.

  As shown in FIG. 27, the identification information calculation unit 524 acquires conversation voice data from the conversation voice table 421 (step S601). Then, the identification information calculation unit 524 performs a character recognition process on the conversation voice data (step S602). Then, the identification information calculation unit 524 extracts a portion that matches the identification from the character data obtained by executing the character recognition process (step S603). In other words, the nickname information calculation unit 524 extracts a portion that matches the word stored in the nickname table 425.

  Then, the identification information calculation unit 524 calculates the number of times that identification is used by counting the number of extracted locations (step S604). Then, the nickname information calculation unit 524 calculates the number of nicknames by dividing the number of times the nickname is used by the length of the conversational voice data (step S605). For example, a case will be described as an example where the nickname information calculation unit 524 uses “12 times” the number of times nickname is used and the length of the conversational voice data is “4 minutes”. In this case, the identification information calculation unit 524 calculates the number of identifications “3 times / minute” by dividing “12 times” by “4 minutes”.

  Further, the identification information calculation unit 524 identifies the customer voice and the operator voice included in the conversation voice data, and the customer speaks as shown in FIGS. 17A, 17B, 19A, and 19B. Is identified (step S606). Also, the identification information calculation unit 524 identifies the time during which the operator is speaking (step S607).

  Then, the contact information calculation unit 524 compares customer times before and after the location where the contact is detected (step S608). In the example illustrated in FIG. 17B, the contact information calculation unit 524 includes customer times “t1” to “t2” and customer times “t5” to “t6” before and after “t3” to “t4” in which the contact is detected. And compare.

  Then, the identification information calculation unit 524 calculates the number of expansions by counting the number of locations where the customer time has increased after identification (step S609). In the example illustrated in FIG. 17B, the customer times “t5” to “t6” are longer than the customer times “t1” to “t2”, and the nickname information calculation unit 524 counts “1” as the number of expansions. Also, the Aizuchi information calculation unit 524 executes the same process for other locations, and calculates the number of expansions “3 times”, for example.

  Also, the nickname information calculation unit 524 calculates the number of multiple expansions by counting the number of locations where the customer continuously talked for two or more times after the operator hit the nickname. (Step S610). In the example shown in FIG. 19A, the customer times “t5” to “t6” are longer than the customer times “t1” to “t2”, and the customer times “t9” to “t6” are longer than the customer times “t5” to “t6”. t10 "is longer. For this reason, the identification information calculation unit 524 counts “1” for the number of multiple expansions. Also, the Aizuchi information calculation unit 524 performs the same process for other locations, and calculates, for example, a plurality of expansion times “2 times”.

  Then, the nickname information calculation unit 524 uses the number of nicknames, the number of expansions, and the number of multiple expansions to calculate a value indicating the listening skill (step S611). For example, the excluded ratio calculation unit 523 acquires the point conversion percentage “100%” from the table for number of matches 434 using “3 times / minute” as a search key. Here, for example, when “5” is set as the score for the number of hits, the ratio calculation unit 522 multiplies “5” by “100%” to obtain a value “ 5 "is calculated. Similarly, the identification information calculation unit 524 obtains point conversion% for the number of expansions and the number of expansions from the expansion number table 435 and the plurality of expansion number table 436, respectively, and indicates the listening skill. Is calculated.

[Effect of Example 2]
As described above, the operator selection device 200 according to the second embodiment includes the conversation voice table 421 that stores conversation voice data in association with operator information. Then, the operator selection device 200 analyzes the conversation voice data for each operator information using a predetermined analysis standard, and calculates a value related to the operator's listening skill based on the analysis result. Then, the operator selection device 200 selects an operator corresponding to a half-talked customer based on the calculated value. Therefore, according to the second embodiment, it is possible to select an operator suitable for a half-talked customer, and as a result, it is possible to provide sufficient customer service to the half-talked customer and to efficiently operate the call center.

  Further, the operator selection device 200 according to the second embodiment analyzes the conversational voice data using the number of times of duplication as a predetermined analysis standard. The operator selection device 200 calculates a value indicating that the listening skill is higher as the number of times of duplication obtained as an analysis result is smaller, and a value indicating that the listening skill is lower as the number of times of duplication is larger. Is calculated.

  That is, the smaller the number of times of duplication, that is, the smaller the number of times the operator speaks when the customer is speaking, the better the operator is. Based on this, since the operator selection device 200 calculates a value related to listening ability, it is possible to select an operator who speaks less when the customer is speaking as a dedicated operator, and it is possible to accurately select a listening operator. It is.

  Further, the operator selection device 200 according to the second embodiment analyzes the conversation voice data using the ratio as a predetermined analysis standard. Then, the operator selection device 200 calculates a value indicating that the higher the percentage of customer time in the ratio obtained as the analysis result, the higher the listening skill. In addition, the operator selection device 200 calculates a value indicating that the smaller the ratio of customer time in the ratio, the lower the listening skill.

  In other words, the greater the proportion of customer time in the ratio, that is, the longer the customer is speaking, the better the operator is. Based on this, the operator selection device 200 calculates a value related to listening ability, so that it is possible to select an operator who makes the customer speak longer than the operator himself as a full-time operator, and it is possible to accurately select a listening operator. is there.

  In addition, the operator selection device 200 according to the second embodiment analyzes conversation voice data using the number of gaps per predetermined time as a predetermined analysis standard. Then, the operator selection device 200 calculates a value indicating that the listening skill is higher as the number of gaps per predetermined time obtained as an analysis result is larger. The value indicating that the listening skill is low is calculated.

  That is, it is considered that the customer can continue to speak comfortably when the operator hits the phone, and the greater the number of phone calls, the better the operator is. Based on this, the operator selection device 200 calculates a value related to listening skill, so that an operator who hits well can be selected as a dedicated operator, and it is possible to accurately select an operator who is listening well.

  Further, the operator selection device 200 according to the second embodiment analyzes the conversation voice data using the number of expansions as a predetermined analysis standard. Then, the operator selection device 200 calculates a value indicating that the listening skill is higher as the number of expansions obtained as an analysis result is larger, and a value indicating that the listening skill is lower as the number of expansions is smaller. Is calculated.

  In other words, if the customer talks longer than before hitting the operator by hitting the operator, it means that the operator hits the phone successfully and the operator is considered to be good at listening. As a result, the greater the number of expansions and the number of expansions, the better the operator is. Based on this, since the operator selection device 200 calculates a value related to listening ability, it is possible to select a dedicated operator as a dedicated operator so that the customer speaks for a long time, and it is possible to accurately select a listening operator. is there.

  Further, the operator selection device 200 according to the second embodiment includes a keyword table 422. Then, operator selection device 200 extracts a keyword from character data obtained by executing character recognition processing on conversation voice data, and reads a product corresponding to the extracted keyword from keyword table 422. Then, when the read product is a product that requires time to explain compared to other products, the operator selection device 200 calculates the ratio after correcting the operator time to a smaller time. In addition, when the read product is a product that does not require time for explanation compared to other products, the operator selection device 200 calculates the ratio after correcting the operator time to a larger time. For this reason, according to the second embodiment, it is possible to select a full-time operator using the same criteria and processing even if the product requires different time for explanation.

  Further, the operator selection device 200 according to the second embodiment includes an excluded text table 424. Then, the operator selection device 200 extracts a portion that matches the fixed phrase stored in the excluded sentence table 424 from the character data obtained by executing the character recognition process on the conversation voice data. Then, operator selection apparatus 200 deletes the portion of conversational voice data corresponding to the extracted character data. Then, operator selection device 200 calculates the ratio using the conversation voice data after deletion.

  That is, it is considered that any operator can speak about the cliché. For this reason, it is considered that the difference between operators becomes difficult to see when the part where the phrase is spoken is taken into consideration. Based on this, the operator selection device 200 calculates the ratio after deleting the time when the phrase is spoken, so that the ratio between the customer time and the operator time can be calculated with high accuracy.

  Further, the operator selection device 200 according to the second embodiment extracts a portion that matches the keyword stored in the keyword table 422 from the character data obtained by executing the character recognition process on the conversation voice data. Then, operator selection apparatus 200 deletes the portion of conversation voice data corresponding to the extracted character data, and calculates the ratio using the deleted conversation voice data.

  That is, it is considered that any operator can make a series of keywords for a product, and it is considered that the difference between operators is difficult to see in consideration of the portion speaking the keyword. Based on this, the operator selection device 200 calculates the ratio after excluding the portion speaking the keyword of the product, so that the ratio between the customer time and the operator time can be calculated with high accuracy.

  Although the embodiments of the present invention have been described so far, the present invention may be implemented in other embodiments besides the above-described embodiments. Therefore, other embodiments will be described below.

[Third storage unit]
For example, in the above-described embodiment, the case where each table of the third storage unit 430 stores the point conversion% has been described. Here, each table of the 3rd memory | storage part 430 may memorize | store score conversion% different for every goods, and may memorize | store score conversion% different for every category of goods. For example, each table of the third storage unit 430 stores a score conversion% for a product that requires a short time for explanation, and for a product that requires a long time for explanation, a score is provided separately from the product that requires a short time for explanation. You may memorize conversion%.

[Selection processing]
For example, in the above-described embodiment, the case where the dedicated operator selecting unit 525 selects a dedicated operator has been described. The dedicated operator selection unit 525 may select a dedicated operator for each product by using a predetermined threshold different for each product, or may select a candidate for a dedicated operator. Here, for example, when a call is received from a half-talked customer, the distribution processing control unit 510 may assign a call to a dedicated operator candidate if all the dedicated operators are answering the call.

  In addition, for example, the dedicated operator selection unit 525 periodically performs a selection process, and logs a value obtained by adding each of the values calculated by the duplication count calculation unit 521 to the identification information calculation unit 524. You may keep it. Then, the dedicated operator selection unit 525 adds the values calculated by the duplication number calculation unit 521 to the identification information calculation unit 524, compares the obtained value with the previous value, and selects the comparison result as the dedicated operator selection. You may use for. For example, the dedicated operator candidate may be promoted to a dedicated operator when the value obtained for the dedicated operator candidate is continuously greater than or equal to a predetermined value for a predetermined number of times or greater than a predetermined value. Further, for example, when the value obtained for the dedicated operator decreases continuously for a predetermined number of times or when it decreases by a predetermined value or more, the dedicated operator may be demoted to a dedicated operator candidate.

[Points]
Further, for example, in the above-described embodiment, the case where the number of times of duplication calculation 521 to the information calculation unit 524 is set in advance has been described as an example. Here, a different score may be set in advance for each product of the duplication count calculation unit 521 to the identification information calculation unit 524. In addition, for example, the second storage unit 420 and the third storage unit 430 may further include a table that stores points used by the duplication count calculation unit 521 to the identification information calculation unit 524.

[Prescribed analysis criteria]
Further, for example, in the above-described embodiment, as the predetermined analysis criteria, the number of times of duplication, the ratio, the ratio at the time of exclusion of a phrase, the ratio at the time of keyword exclusion, the number of gaps, the number of times of expansion, and the number of times of multiple expansion The case of using was described. However, the present invention is not limited to this. For example, it is not necessary to use some of the number of times of duplication, the ratio, the ratio at the time of rule exclusion, the ratio at the time of keyword exclusion, the number of rounds, the number of expansions, and the number of expansions. May be used in combination.

[Distribution processing]
In addition, for example, the distribution processing control unit 510 may distribute operators and full-time operators in consideration of the customer's hometown and the customer's interest. For example, the distribution processing control unit 510 may select a dedicated operator from the same birthplace as the customer's birthplace when the telephone is assigned to the dedicated operator. It is assumed that the customer's or operator's hometown is stored in the customer table or the operator table.

[Keyword table]
Further, for example, in the above-described embodiment, the keyword table 422 is used when identifying a product, and the keyword table 422 is also used when the exclusion ratio calculation unit 523 excludes a keyword from conversation voice data. Explained. However, the present invention is not limited to this. For example, a keyword different from the keyword table 422 may be stored as a keyword used when the exclusion ratio calculation unit 523 excludes a keyword from conversation voice data. That is, the table used when identifying the product and the table used when excluding the keyword may be different.

[Customer time extension]
Further, for example, in the above-described embodiment, the case where the customer time is used before and after the time of hitting is described as an example. Here, it is also possible to further calculate a ratio that the customer time after the match has become longer than the customer time before the match and use the calculated value to calculate a value related to the listening skill. For example, the operator selection device further includes a table that stores the point conversion% in association with the ratio. Then, when the customer time becomes long before and after the hitting time, the operator selection device calculates the ratio of the customer time becoming long, and acquires the point conversion% corresponding to the calculated ratio. And an operator selection apparatus calculates the value regarding a listening skill based on the acquired point conversion%.

[System configuration]
In addition, all or some of the processes described as being automatically performed among the processes described in the present embodiment may be performed manually. For example, customer time and operator time may be calculated manually. In addition, the processing procedures, control procedures, specific names, and information including various data and parameters (FIGS. 1 to 27) shown in the above-mentioned document and drawings are arbitrary unless otherwise specified. Can be changed.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or a part of the distribution / integration may be functionally or physically distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured. For example, in the example illustrated in FIG. 3, the first storage unit 410, the second storage unit 420, and the third storage unit 430 may be integrated. Further, the distribution processing control unit 510 and the first storage unit 410 may be connected to the operator selection device 200 via a network as an external device of the operator selection device 200.

[Computer]
The various processes described in the above embodiments can be realized by executing a program prepared in advance on a computer such as a personal computer or a workstation. Therefore, in the following, an example of a computer that executes an operator selection program having the same function as in the above-described embodiment will be described with reference to FIG. FIG. 28 is a schematic diagram illustrating an example of a computer that executes an operator selection program according to the second embodiment.

  As shown in FIG. 28, the computer 3000 according to the second embodiment includes an operation unit 3001, a microphone 3002, a speaker 3003, a display 3004, a communication unit 3006, a CPU 3010, a ROM 3011, and an HDD (Hard Disk Drive) 3012. The computer 3000 has a RAM (Random Access Memory) 3013. The computer 3000 is configured by connecting operations 3001 to RAM 3013 via a bus 3009 or the like.

  The ROM 3011 includes a distribution process control unit 510, a duplication count calculation unit 521, a ratio calculation unit 522, an exclusion time ratio calculation unit 523, an identification information calculation unit 524, and a dedicated operator shown in the second embodiment. A control program that exhibits the same function as the selection unit 525 is stored. That is, as shown in FIG. 28, the ROM 3011 stores in advance a distribution process control program 3011a, an overlap count calculation program 3011b, and a ratio calculation program 3011c. In addition, the ROM 3011 stores in advance an exclusion ratio calculation program 3011d, an identification information calculation program 3011e, and a dedicated operator selection program 3011f. Note that these programs 3011a to 3011f may be integrated or separated as appropriate, similarly to each component of the operator selection device 200 shown in FIG.

  Then, the CPU 3010 reads these programs 3011a to 3011f from the ROM 3011 and executes them. As a result, as shown in FIG. 28, each of the programs 3011a to 3011f functions as a distribution process control process 3010a, a duplication count calculation process 3010b, a ratio calculation process 3010c, and an exclusion time ratio calculation process 3010d. Each of the processes 3010a to 3010f includes the distribution processing control unit 510, the duplication number calculation unit 521, the ratio calculation unit 522, the excluded time ratio calculation unit 523, and the identification information calculation unit 524 illustrated in FIG. And a dedicated operator selection unit 525.

  The HDD 3012 is provided with a first table 3012a, a second table 3012b, and a third table 3012c. Each of the tables 3012a to 3012c corresponds to the first storage unit 410, the second storage unit 420, and the third storage unit 430 illustrated in FIG.

  Then, the CPU 3010 reads the first table 3012a, the second table 3012b, and the third table 3012c and stores them in the RAM 3013. Then, the CPU 3010 executes an operator selection program using the first data 3013a, the second data 3013b, and the third data 3013c stored in the RAM 3013.

[Others]
Note that the operator selection program described in this embodiment can be distributed via a network such as the Internet. The operator selection program can also be executed by being recorded on a computer-readable recording medium such as a hard disk, a flexible disk (FD), a CD-ROM, an MO, and a DVD, and being read from the recording medium by the computer.

  The following additional notes are further disclosed with respect to the embodiments including the above examples.

(Supplementary Note 1) A conversation voice data storage unit that stores conversation voice data between a user using the call center and the operator in association with operator information for identifying the operator of the call center;
By calculating the conversation voice data stored by the conversation voice data storage unit using a predetermined analysis standard for each of the operator information, a calculation unit that calculates a value related to the operator's listening skill;
An operator selection device comprising: a selection unit that selects an operator based on the value calculated by the calculation unit.

(Additional remark 2) The said calculation part is analyzed using the 1st frequency | count which shows the frequency | count that the said operator's voice and the said user's voice overlapped as a predetermined | prescribed analysis reference | standard, The said 1st obtained as an analysis result A value indicating that the listening skill is higher as the number of 1s is smaller, and a value indicating that the listening skill is lower as the first number is higher. The operator selection device described in 1.

(Additional remark 3) The said calculation part analyzes using the ratio of the operator time which shows the time which the said operator talked, and the user time which shows the time which the said user spoke as a predetermined | prescribed analysis reference | standard, In the obtained ratio, the larger the proportion of the user time, the higher the hearing ability is calculated. The smaller the proportion of the user time in the ratio, the better the hearing ability. The operator selection device according to appendix 2, wherein a value indicating low is calculated.

(Additional remark 4) The said calculation part analyzes using the 2nd frequency | count which shows the frequency | count per predetermined time that the said operator hits as a predetermined | prescribed analysis reference | standard, The said 2nd frequency | count obtained as an analysis result The value indicating that the better the listening is calculated as the number of is higher, and the value indicating that the listening is lower as the second number of times is smaller is calculated. Operator selection device.

(Additional remark 5) The said calculation part said the said user continuously as the predetermined | prescribed analysis reference | standard after the said time from the time when the said user spoke continuously until the time when the said operator hits Analysis is performed using the third number indicating the number of times that the time has been increased, and the greater the third number obtained as an analysis result, the higher the listening ability is calculated. The operator selecting device according to supplementary note 3, wherein a value indicating that the listening skill is low is calculated as the number of times is less.

(Supplementary Note 6) A keyword storage unit that stores a keyword related to the product for each product,
The calculation unit extracts a keyword from character data obtained by executing a character recognition process on the conversation voice data, reads a product corresponding to the extracted keyword from the keyword storage unit, and the read product is another If it is a product that takes time to explain compared to other products, the ratio is calculated after correcting the operator time to a smaller time, and the read product is time to explain compared to other products. The operator selection device according to appendix 3, wherein the ratio is calculated after the operator time is corrected to a larger time when the product does not require an item.

(Additional remark 7) It further has a fixed phrase memory | storage part which memorize | stores the fixed phrase which the said operator speaks with respect to a user,
The calculation unit extracts a portion of the character data obtained by executing a character recognition process on the conversational voice data, which matches the fixed phrase stored in the fixed phrase storage unit, and is extracted The operator selection device according to appendix 3, wherein the conversation voice data corresponding to the character data is deleted, and the ratio is calculated using the deleted conversation voice data.

(Supplementary Note 8) A keyword storage unit that stores a keyword related to the product for each product is further provided,
The calculation unit extracts a portion that matches a keyword stored in the keyword storage unit from character data obtained by executing a character recognition process on the conversation voice data, and extracts the extracted character data. The operator selection device according to appendix 3, wherein the conversation voice data of the corresponding part is deleted, and the ratio is calculated using the conversation voice data after the deletion.

(Appendix 9)
A reading procedure for reading out the conversation voice data from a conversation voice data storage unit that stores conversation voice data between a user who uses the call center and the operator in association with operator information for identifying a call center operator;
A calculation procedure for calculating a value related to the operator's listening ability by analyzing the conversation voice data read by the reading procedure using a predetermined analysis standard for each operator information,
And a selection procedure for selecting an operator based on the value calculated by the calculation procedure.

DESCRIPTION OF SYMBOLS 100 Operator selection apparatus 101 Conversation voice data storage part 102 Calculation part 103 Selection part 200 Operator selection apparatus 301 Customer terminal 302 Operator 303 Dedicated operator 310 Customer table 320 Conversation voice table 330 Distribution process control part 340 Selection process control part 410 1st Storage unit 411 Operator table 412 Customer table 420 Second storage unit 421 Conversation voice table 422 Keyword table 423 Coefficient table 424 Exclusion sentence table 425 Aichi table 430 Third storage unit 431 Duplication count table 432 Ratio table 433 Exclusion Ratio table 434 Matching number table 435 Expansion number table 436 Multiple expansion number table 510 Distribution processing control unit 520 Selection processing control unit 521 Duplication number calculation unit 522 Ratio calculation unit 523 Exclusion ratio calculation unit 524 Aichichi information calculation unit 525 Dedicated operator selection unit 600 Customer terminal

Claims (9)

A sender information storage unit that stores information that identifies a customer who is a sender of a telephone call and a type of a customer of the sender in advance;
For each of a plurality of call center operators, a calculation unit that calculates a value related to the operator's listening ability based on conversation voice data when the operator has responded in the past;
Electrostatic accepts the story, the originating on the basis of the source information storage unit to identify the type of the calling customer of the received the phone, if the type of the customer identified a type of the given customer, the good listener is A control unit that distributes the received telephone call to a predetermined operator specified by a value related to
An apparatus for selecting an operator, comprising: As the calculation standard of the calculation unit, the value related to the listening skill is increased as the first number, which is the number of times the voice of the operator and the voice of the user using the call center, overlap in the conversation voice data. The operator selection device according to claim 1, further comprising: calculating a value related to the listening skill .   As a calculation standard of the calculation unit, analysis is performed using a ratio of an operator time indicating the time spoken by the operator and a user time indicating a time spoken by a user using the call center, and the analysis result is obtained. Calculate a value indicating that the greater the percentage of the user time in the ratio, the higher the listening ability. The smaller the percentage of the user time in the ratio, the lower the listening ability. The operator selection device according to claim 1, further comprising calculating a value.   As the calculation standard of the calculation unit, the second number indicating the number of times per predetermined time that the operator hits the head is analyzed, and the more the second number obtained as a result of the analysis, the better the listener The operator selection device according to claim 3, further comprising: calculating a value indicating that the sound is high, and calculating a value indicating that the listening skill is low as the second number of times is smaller. .   As a calculation criterion of the calculation unit, the number of times that the user continuously talked after the point of time is longer than the time when the user continuously talks until the time when the operator hits The more the third number obtained as an analysis result is calculated, the higher the hearing ability is calculated. The smaller the third number is, the more the third number obtained as an analysis result is calculated. The operator selection device according to claim 3, further comprising calculating a value indicating that the listening skill is low. For each of a plurality of call center operators, a calculation procedure for calculating a value related to the operator's listening ability based on conversation voice data when the operator has responded in the past,
When receiving a call, and information identifying the calling customer telephone, based on the source information storage unit for previously associating and storing the type of calling customer, the originating type of customer telephone accepted identify, when the type of the customer was especially boss is the type of a given customer, the predetermined operator specified by the value related to the good listener is, a control procedure for distributing received the telephone,
Selection program for causing a computer to execute Means for detecting that a predetermined word or phrase has been issued in the voice data of the telephone received by the operator;
Means for evaluating the skill of the operator based on a result of analyzing a change in the length of time of the voice data corresponding to the caller before and after the predetermined word is detected in the voice data; ,
An operator evaluation device comprising: On the computer,
Detect that a certain word or phrase has been issued in the voice data of the telephone that the operator answered,
Based on the result of analyzing the change in the length of time of the voice data corresponding to the caller before and after the predetermined word is detected in the voice data, the operator's skill is evaluated.
An operator evaluation program characterized by causing Computer
Detect that a certain word or phrase has been issued in the voice data of the telephone that the operator answered,
Based on the result of analyzing the change in the length of time of the voice data corresponding to the caller before and after the predetermined word is detected in the voice data, the operator's skill is evaluated.
An operator evaluation method characterized in that:

Images