JP2022105834A - System for managing incoming call handling operation and method for managing incoming call handling operation - Google Patents

Abstract

To provide systems and methods for managing incoming call handling operations, capable of efficiently and appropriately performing incoming call handling operations by accurately predicting a tendency of incoming calls.SOLUTION: A system for managing incoming call handling operations is configured to: store an incoming call handling state prediction model which is a machine learning model that outputs, as objective variables, the number of incoming callas and the number of responses in a prediction target time zone by inputting, as explanatory variables, attribute information of a person who may call, a prediction target date, a prediction target time zone on the prediction target date, and the number of operators scheduled to be deployed in the prediction target time zone; learn the incoming call handling state prediction model by learning data including attribute information of a person who called in a predetermined time zone on a past predetermined date, the number of operators who were deployed in the predetermined time zone on the predetermined date, and the number of incoming calls and the number of responses in the predetermined time zone on the predetermined date; and determine the objective variables by inputting the explanatory variables into the incoming call handling state prediction model.SELECTED DRAWING: Figure 1

Description

The present invention relates to an incoming call response business management system and an incoming call response business management method.

In the operation of a call center that handles calls from customers, it is necessary to appropriately manage the allocation of operators every hour and day in order to improve customer satisfaction and reduce operating costs such as labor costs. ..

Patent Document 1 describes a personnel planning system configured for the purpose of calculating the optimum allocation of the number of operators for each hour in a call center system. Based on the history of past incoming calls, the personnel planning system calculates the number of corrected incoming calls for each past time unit, excluding incoming calls due to the operator not being able to answer, and based on the above-mentioned number of corrected incoming calls. The number of incoming calls for each time unit is predicted, and the number of operators required for each time unit is calculated based on the predicted number of incoming calls for each time unit.

Patent Document 2 describes a staffing planning system configured for the purpose of appropriately planning staffing at a help desk in consideration of regional characteristics and timing. The staffing planning system acquires the data of the actual number of responses for each business at each help desk retroactively from the specified time by the reference period, calculates the number of increase / decrease in each time period, and corresponds to the calculated increase / decrease number. The increase / decrease tendency value is calculated by multiplying the weighting values related to the above, adding them up, and dividing by the total value of the weighting values. Calculate the required number of personnel for each job in.

Patent Document 3 describes a staffing processing system configured for the purpose of making the staffing of workers appropriate in consideration of the variation in ability. The staffing processing system predicts the amount of work based on past actual data and generates forecast work amount data, calculates the amount of work data based on the processing capacity of the worker and the working hours including attendance, and predicts the amount of work data. And the processing amount data are used to calculate the data on the excess / deficiency state of the labor force.

Japanese Unexamined Patent Publication No. 2006-254094 Japanese Unexamined Patent Publication No. 2011-107902 Japanese Unexamined Patent Publication No. 2002-149931

In an electric power company, for example, a person whose payment of electricity charges is delayed (hereinafter referred to as "delinquent person") is treated as delinquent when the payment deadline (for example, 30 days from the meter reading date) has passed. If payment cannot be confirmed within the specified number of days (for example, 20 days) from the payment deadline, a reminder letter or notification of power transmission cancellation will be sent to the delinquent, and if the specified number of days (for example, 2 months) have passed from the payment deadline, the delinquent payment will be made. The power transmission to the person's electrical equipment (hereinafter referred to as "delinquent's house") is stopped. In addition, the electric power company, for example, takes measures to extend the deadline for stopping power transmission when there is an offer (contact) from the delinquent person about when to pay the fee before stopping power transmission to the delinquent's house. Therefore, a telephone counter (hereinafter referred to as "customer center") is provided to respond to the telephone contact (hereinafter, also referred to as "incoming call") of the above offer.

At the customer center, the number of incoming calls (number of incoming calls), the number of responses (number of responses by the operator), the work load (personal role (ninyaku), artificial (ninku), etc.) for each day and hour, The number of call losses (the number of incoming calls that the operator could not handle) and other past achievements, as well as the experience and intuition of on-site staff, are used to make a human-based judgment, and the required operator placement (number of operators) is determined.

However, it may not be possible to respond to sudden changes in work load due to reasons such as "sudden absence on the day" or "unexpected incoming call (incoming call) on the day". In such cases, for example, support other departments. We are responding by requesting. Further, in that case, the number of support personnel is determined by human beings, and it is not uncommon for the number of support personnel to be insufficient or excessively assigned.

The present invention has been made in view of such a background, and is an incoming call response business management system and an incoming call response business management method capable of accurately predicting an incoming call tendency and efficiently and appropriately performing an incoming call response business. The purpose is to provide.

One of the present inventions for achieving the above object is an information processing system that manages operations for responding to an incoming call, and includes attribute information of a person who may receive an incoming call, a prediction target date, and the above-mentioned prediction. By inputting the prediction target time zone of the target day and the number of operators scheduled to be assigned in the prediction target time zone as explanatory variables, the number of incoming calls and the number of responses in the prediction target time zone are output as objective variables. The incoming call response status prediction model storage unit that stores the incoming call response status prediction model, which is a machine learning model, and the attribute information of the person who received the call in the predetermined time zone of the past predetermined day, are arranged in the predetermined time zone of the predetermined day. The incoming call response status prediction model learning unit that learns the incoming call response status prediction model from the learning data including the number of operators and the number of incoming calls and responses in the predetermined time zone on the predetermined day, and the explanatory variables. It is provided with an incoming call response status prediction unit for inputting into the incoming call response status prediction model and obtaining the objective variable.

As described above, the incoming call response business management system of the present invention accurately obtains the number of incoming calls and the number of responses in the prediction target time zone of the prediction target day by the incoming call response status prediction model learned based on the past actual information. Therefore, the user can grasp in advance the number of incoming calls and the number of responses in the predicted target time zone of the predicted target day, and can allocate an appropriate number of operators to efficiently and appropriately perform the incoming call response work.

The other one of the present invention is the incoming call correspondence business management system, and the incoming call correspondence situation prediction model learning unit has attribute information of a person who has already received an incoming call when the prediction target date is the same day. The incoming call response status prediction model is learned from the learning data including the number of operators assigned in the time zone when the incoming call was received, and the number of incoming calls and the number of responses in the time zone, and the incoming call response status prediction is performed. The unit inputs the explanatory variables into the incoming call response status prediction model after learning, and obtains the objective variable in a predetermined time zone in the future of the prediction target date.

As described above, in the incoming call response business management system of the present invention, the attribute information of the person who received the incoming call on the day of the prediction target day, the number of operators assigned in the time zone when the incoming call was received, and the number of incoming calls in the above time zone and Incoming call response status prediction model is learned from learning data including the number of responses, and sudden changes in work load due to reasons such as "sudden absenteeism on the day" or "unexpected incoming call (incoming call) on the day" are predicted. Since it is reflected in the model, the number of incoming calls and the number of responses in the future time zone of the day to be predicted can be accurately obtained. Therefore, the user can secure an appropriate number of operators on the day and efficiently perform the incoming call response work.

Another one of the present invention is the incoming call response business management system, in which the incoming call response status prediction unit is based on the obtained objective variable of the operator in the prediction target time zone of the prediction target date. Find the excess or deficient number.

In this way, the incoming call response business management system of the present invention obtains the excess / deficient number of operators in the prediction target time zone of the prediction target day. Can be grasped.

Another one of the present invention is the incoming call response business management system, in which the incoming call response status prediction unit is an operator in the future prediction time zone of the prediction target date based on the obtained objective variable. Find the excess or deficiency number of.

In this way, the incoming call response business management system of the present invention obtains the excess / deficient number of operators in the future prediction target time zone of the prediction target date when the prediction target date is the current day, so that the user can predict the future of the day. It is possible to grasp in advance the number of excess and deficiency of operators in the target time zone, and promptly take necessary measures according to the situation of the day such as "sudden absenteeism on the day" and "unexpected incoming call (incoming call) on the day". Can be taken to.

Another one of the present invention is the incoming call response business management system, in which the number of call losses in the prediction target time zone is preset based on the objective variable obtained by the incoming call response status prediction unit. Further, a warning output unit is provided, which determines whether or not the threshold value is exceeded, and outputs information indicating a warning immediately when the determination is exceeded.

As described above, when the number of call losses in the predicted time zone exceeds the preset threshold value, the incoming call response business management system of the present invention immediately presents information indicating a warning to the user, so that many call losses occur to the user. Can quickly inform you that there is a possibility of doing so. Therefore, the user can immediately detect that a large number of call losses may occur and take necessary measures in advance.

In addition, the problems disclosed in the present application and the solutions thereof will be clarified by the column of the form for carrying out the invention and the drawings.

According to the present invention, it is possible to accurately predict the tendency of incoming call and perform the incoming call response work efficiently and appropriately.

It is a figure which shows the schematic structure of the incoming call correspondence business management system. This is an example of the hardware of the information processing device used to realize the business management system for incoming calls. This is an example of delinquent information. This is an example of the incoming call response status history. This is an example of operator placement schedule information. It is a flowchart explaining the prediction model learning process. It is a flowchart explaining the incoming call correspondence situation prediction processing. This is an example of the prediction result presentation screen.

Hereinafter, embodiments for carrying out the invention will be described. In the following description, the same or similar configurations may be designated by a common reference numeral and the description thereof may be omitted.

FIG. 1 shows the main functions of the information processing system (hereinafter referred to as “incoming call response business management system 1”) described as an embodiment of the present invention. The incoming call response business management system 1 is a business related to incoming call response for an "offer" regarding when to pay the electric power charge from the delinquent, which is performed by the electric power company (hereinafter referred to as "incoming call response business"). To support. The electric power company has set up a number of lines necessary for responding to the incoming call for the above offer and a booth for the operator (communicator) corresponding to the incoming call to carry out the work.

The incoming call response business management system 1 uses a machine learning model (incoming call response status prediction model 116 described later) to predict a day to be predicted (hereinafter referred to as a “prediction target day”) (hereinafter referred to as a “prediction target day”). Hereinafter, the number of incoming calls (number of incoming calls), the number of actual incoming calls by the operator (hereinafter referred to as "number of responses"), and the number of call losses (= number of incoming calls-) in the "prediction target time zone"). The number of responses), the response rate (= the number of responses / the number of incoming calls), etc. are obtained, and information on the excess or deficiency of the operator in each prediction target time zone is generated. The user of the incoming call response business management system 1 uses the generated information to detect in advance, for example, a situation in which a large number of call losses occur and a timing in which a large number of call losses occur, and efficiently and appropriately performs the incoming call response business. It will be possible to do.

As shown in the figure, the incoming call response business management system 1 includes a storage unit 110, a contractor information management unit 120, a delinquent information management unit 125, an incoming call response status history management unit 130, an operator placement schedule management unit 140, and learning data. It has functions of a generation unit 150, an incoming call response status prediction model learning unit 160, an incoming call response status prediction unit 170, a prediction result presentation unit 180, and an alarm output unit 190.

The storage unit 110 includes contractor information 111, delinquent information 112, incoming call response status history 113, operator placement schedule information 115, learning data 114, incoming call response status prediction model 116, additional performance information 117 on the day, prediction result 118, and the like. Store information (data).

The contractor information 111 includes information on a person (customer (customer) or consumer; hereinafter referred to as a “contractor”) who has an electricity usage contract with an electric power company. The contractor information 111 is provided, for example, from a master database of customer management information of an electric power company.

The delinquent information 112 is information extracted from the contractor information 111, and includes information on a contractor who is currently delinquent in electricity charges (hereinafter referred to as “delinquent person”). The delinquent information 112 includes information for identifying the delinquent, as well as attribute information of the delinquent (for example, a delinquent ID, a region ID, the number of months of delinquency, and a mode of payment of money received by the delinquent (money such as salary). (Payment date, etc.)), information on the history of power transmission cancellations made to delinquents in the past (date and time when power transmission was stopped, etc.), actions taken by delinquents in the past (offers, payment of electricity charges, etc.) Includes information about the history of.

The incoming call response status history 113 includes information such as the number of incoming calls, the number of responses, the number of call losses, the response rate, the number of arranged operators, and the operator ID of the arranged operators in the predetermined time zone of the past predetermined day. ..

The learning data 114 is data used for learning (machine learning) of the incoming call response status prediction model 116, and is generated based on the delinquent person information 112 and the incoming call response status history 113. The learning data 114 includes attribute information of a person who has received a call in the predetermined time zone of the past predetermined day, the number of operators arranged in the predetermined time zone of the predetermined day, and an incoming call in the predetermined time zone of the predetermined day. Includes number and number of responses.

The operator placement schedule information 115 includes information indicating the number of operators scheduled to be assigned for each day and time zone. The operator allocation schedule information 115 is set, for example, by a user who manages the allocation of operators in an electric power company.

The learning data 114 is data used for learning (machine learning) of the incoming call response status prediction model 116, and is generated based on the delinquent person information 112 and the incoming call response status history 113. The learning data 114 includes attribute information of a person who has received a call in the predetermined time zone of the past predetermined day, the number of operators arranged in the predetermined time zone of the predetermined day, and an incoming call in the predetermined time zone of the predetermined day. Includes number and number of responses.

The incoming call response status prediction model 116 arranges the attribute information of a person who has a possibility of incoming call (delinquent person in this embodiment), the prediction target date, the prediction target time zone of the prediction target date, and the prediction target time zone. This is a machine learning model that outputs the number of incoming calls and the number of responses in the prediction target time zone as objective variables by inputting the number of planned operators as explanatory variables. The incoming call response status prediction model 116 is realized by, for example, gradient boosting (GBDT (Gradient Boosting Decision Tree)) or deep learning (DNN (Deep Neural Network)). However, the method of realizing the incoming call response status prediction model 116 is not necessarily limited.

The additional performance information 117 on the day includes the attribute information of the person who has already received the call on the day of the prediction target, the number of operators assigned in the time zone when the call was received, and the number of calls and responses in the time zone. Etc. are included. The additional performance information 117 on the day is used as additional learning data 114.

The prediction result 118 includes information indicating the incoming call response status in the predicted time zone of the prediction target day, which is predicted by the incoming call response status prediction unit 170 based on the incoming call response status prediction model 116.

Among the functions shown in FIG. 1, the contractor information management unit 120 manages the contractor information 111. Further, the delinquent information management unit 125 manages the delinquent information 112. The delinquent information management unit 125 generates the delinquent information 112 based on the contractor information 111. The delinquent information management unit 125 can also accept registration and editing of the delinquent information 112 from the user.

The incoming call response status history management unit 130 manages the incoming call response status history 113. The incoming call response status history management unit 130 acquires the contents of the incoming call response status history 113 from, for example, another information processing system that manages the incoming call response business. Further, the incoming call response status history management unit 130 receives, for example, an input of the incoming call response status history 113 from the user.

The operator allocation schedule management unit 140 manages the operator allocation schedule for each day and time zone in the past, present, and future in the operator allocation schedule information 115. The operator allocation schedule management unit 140 receives, for example, an operator allocation schedule for each day and time zone from the user via the user interface. As a general rule, future operator placement schedules are set before a predetermined period such as the middle of the previous month.

The learning data generation unit 150 generates learning data 114 (or verification data used for verifying the accuracy of the incoming call response status prediction model 116) based on the delinquent person information 112 and the incoming call response status history 113. Further, when the learning data generation unit 150 generates the additional performance information 117 on the day, the learning data 114 is generated based on the additional performance information 117 on the day. The training data 114 is an explanatory variable (attribute information of a person who may receive an incoming call, a prediction target date, a prediction target time zone of the prediction target date, and the number of operators scheduled to be placed in the prediction target time zone). And the information corresponding to the objective variable (the number of incoming calls and the number of responses in the prediction target time zone) which is the correct answer label is included.

The incoming call response status prediction model learning unit 160 performs processing related to learning and verification of the incoming call response status prediction model 116 using the learning data 114. Further, when the incoming call response status prediction model learning unit 160 generates the additional actual result information 117 on the day, the learning data 114 based on the additional actual result information 117 on the current day is used to perform processing related to learning and verification of the incoming call response status prediction model 116. .. By learning from the above learning data 114, the incoming call response status prediction model 116 can be set to, for example, the time required from when the transmission was stopped until the incoming call was received when the power transmission was stopped in the past for the delinquent, or in the morning. The ratio of the number of incoming calls in the afternoon is also reflected.

The incoming call response status prediction unit 170 arranges the attribute information of a person who may receive an incoming call, the prediction target date, the prediction target time zone of the prediction target date, and the prediction target time zone in the incoming call response status prediction model 116. By inputting the number of planned operators as explanatory variables, the number of incoming calls and the number of responses in the above-mentioned prediction target time zone are obtained as objective variables, and based on the obtained objective variable, the operators in the above-mentioned time zone of the above-mentioned prediction target date. Find the excess or deficiency number of. The incoming call response status prediction unit 170 stores the obtained number of incoming calls, the number of responses, and the number of excess / deficiency of the operator as the prediction result 118.

The prediction result presentation unit 180 presents the contents of the prediction result 118 to the user. For example, the prediction result presenting unit 180 receives a designation of a prediction target date and a prediction target time zone from a user via a user interface, and the number of incoming calls, the number of responses, the call loss rate, and the response rate in the prediction target time zone of the received prediction target day. Information such as is presented to the user.

The alarm output unit 190 determines whether or not the number of call losses in the prediction target time zone exceeds a preset threshold value based on the objective variable obtained by the incoming call response status prediction unit 170, and warns when it is determined. The information indicating that is output.

FIG. 2 shows an example of the hardware of the information processing device used to realize the incoming call support business management system 1. The information processing device 10 exemplified includes a processor 11, a main storage device 12, an auxiliary storage device 13, an input device 14, an output device 15, and a communication device 16. It should be noted that the information processing apparatus 10 illustrated is virtual information in which all or part of the information processing apparatus 10 is provided by using virtualization technology, process space separation technology, or the like, for example, like a virtual server provided by a cloud system. It may be realized by using processing resources. Further, all or a part of the functions provided by the information processing apparatus 10 may be realized by, for example, a service provided by a cloud system via an API (Application Programming Interface) or the like. Further, all or a part of the functions provided by the information processing apparatus 10 are realized by using, for example, SaaS (Software as a Service), PaaS (Platform as a Service), IaaS (Infrastructure as a Service), and the like. It may be one. Further, the incoming call support business management system 1 may be configured by using a plurality of information processing devices 10 connected so as to be communicable.

In the figure, the processor 11 is, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), an FPGA (Field Programmable Gate Array), an ASIC (Application Specific Integrated Circuit), and an AI (Artificial). Intelligence) It is configured by using a chip or the like.

The main storage device 12 is a device for storing programs and data, and is, for example, a ROM (Read Only Memory), a RAM (Random Access Memory), a non-volatile memory (NVRAM (Non Volatile RAM)), and the like.

The auxiliary storage device 13 includes, for example, an SSD (Solid State Drive), a hard disk drive, an optical storage device (CD (Compact Disc), DVD (Digital Versatile Disc), etc.), a storage system, an IC card, an SD card, or an optical recording device. A reading / writing device for a recording medium such as a medium, a storage area for a cloud server, and the like. The auxiliary storage device 13 can read programs and data via a reading device of a recording medium and a communication device 16. Programs and data stored (stored) in the auxiliary storage device 13 are read into the main storage device 12 at any time.

The input device 14 is an interface that accepts input from the outside, and is, for example, a keyboard, a mouse, a touch panel, a card reader, a pen input type tablet, a voice input device, and the like.

The output device 15 is an interface that outputs various information such as processing progress and processing results. The output device 15 is, for example, a display device (liquid crystal monitor, LCD (Liquid Crystal Display), graphic card, etc.) that visualizes the above-mentioned various information, and a device (voice output device (speaker, etc.)) that visualizes the above-mentioned various information. , A device (printing device, etc.) that converts the above various information into characters. In addition, for example, the information processing device 10 may be configured to input and output information to and from another device via the communication device 16.

The input device 14 and the output device 15 form a user interface for receiving and presenting information with the user.

The communication device 16 is a device that realizes communication with other devices. The communication device 16 is a wired or wireless communication interface that realizes communication with other devices via a communication network such as the Internet, and is, for example, a NIC (Network Interface Card), a wireless communication module, or a USB. Modules, etc. Further, the communication device 16 communicates with an information processing system such as a power distribution control system that controls a smart meter installed in a consumer's house.

For example, an operating system, a file system, a DBMS (DataBase Management System) (relational database, NoSQL, etc.), a KVS (Key-Value Store), or the like may be introduced in the information processing apparatus 10.

Each of the above-mentioned functions included in the incoming call corresponding business management system 1 is performed by the processor 11 of the information processing device 10 reading and executing the program stored in the main storage device 12, or by executing the incoming call corresponding business management system 1. It is realized by the constituent hardware (FPGA, ASIC, AI chip, etc.). The incoming call response business management system 1 stores the various information (data) described above as, for example, a database table or a file managed by the file system.

FIG. 3 shows an example of the delinquent information 112. The exemplary delinquent information 112 is one or more records (entries) having the respective items of contractor ID 1120, area ID 1121, delinquent months 1122, power transmission status 1123, salary payment date 1126, delinquency history 1127, and incoming call history 1128. ). One of the records of delinquent information 112 corresponds to one delinquent.

Among the above items, the contractor ID 1120 is set to the contractor ID, which is an identifier of the delinquent (also a contractor).

In the area ID 1121, an area ID which is an identifier of the area where the delinquent's house (customer's house) to which the delinquent is supplied with electric power exists is set.

Information indicating the period during which the delinquent person is currently continuing to delinquent payment of the electric power charge is set in the delinquent months 1122. The electric power company, for example, stops power transmission when 50 days (including a grace period of 20 days) have elapsed from the day after the meter reading date.

Information indicating the current power transmission status of the delinquent to the delinquent's house is set in the power transmission status 1123 (“power transmission is in progress” if the power is currently being transmitted, and “stopped” if the power transmission is currently stopped. Is set).

Information indicating the date on which the delinquent person receives salary is set in the salary payment date 1126.

In the delinquency history 1127, whether or not the delinquent has delinquently paid the electricity bill in the past (for example, within the past year), various information regarding the delinquency (the period and date and time during which the delinquency was continued, and the offer) are provided. Historical information about the past delinquency of the delinquent, such as the date and time of payment and the date and time of payment), is set.

The incoming call history (one or more) of the delinquent person is set in the incoming call history 1128.

FIG. 4 shows an example of the incoming call response status history 113. The figure shows the response history of the incoming call status in each time zone of one day "October 1, 2020" in the incoming call response status history 113. As shown in the figure, the example incoming call correspondence status history 113 includes items of date 1130, time zone 1131, incoming call number 1132, response number 1133, call loss number 1134, response rate 1135, operator number 1136, and operator ID 1137. Contains multiple records with.

The date 1130 is set to a date indicating which day the incoming call history belongs to.

In the time zone 1131, information (hereinafter, referred to as “the time zone”) indicating each time zone of the day of the above date (hereinafter, referred to as “the incoming call date”) is set.

The number of incoming calls 1132 is set to the number of incoming calls in the time zone of the incoming call date.

The number of responses in the time zone of the incoming call date is set in the number of responses 1132.

The number of call losses 1134 is set to the number of call losses (= number of incoming calls-number of responses) in the relevant time zone on the relevant incoming call date.

In the response rate 1135, the response rate (= number of responses / number of incoming calls) in the time zone of the incoming call date is set.

In the number of operators 1136, the number of operators arranged in the time zone of the incoming call date is set.

The operator ID 1137 is set to the identifier of the operator (hereinafter, referred to as "operator ID") arranged in the time zone of the incoming call date.

FIG. 5 shows an example of the operator placement schedule information 115. The operator allocation schedule information 115 manages the number of operators scheduled to be assigned on each business day of the forecast month. The figure illustrates the contents of the business day 1150 with the operator placement schedule information 115. As shown in the figure, the operator placement schedule information 115 manages the number of operators to be placed 1152 and the identifier of the operator to be placed (operator ID 1153) for each time zone 1151 of the business day.

Next, the main processing performed by the incoming call response business management system 1 will be described.

FIG. 6 is a flowchart illustrating a process (hereinafter, referred to as “prediction model learning process S600”) performed by the incoming call response business management system 1 when learning the incoming call response status prediction model 116. Hereinafter, the prediction model learning process S600 will be described with reference to the figure.

The delinquent information management unit 125 generates delinquent information 112 at any time (regularly, timing when the contractor information 111 is updated, etc.) based on the contractor information 111 (S611).

When a predetermined timing (regularly, when an instruction is received from a user, etc.) arrives, the learning data generation process S612 generates learning data 114 based on the delinquent information 112 and the incoming call response status history 113 (S612).

When a predetermined timing (regularly, when new learning data is generated, etc.) arrives, the incoming call response status prediction model learning unit 160 learns the incoming call response status prediction model 116 based on the learning data 114.

The process of verifying the accuracy of the incoming call response status prediction model 116 is also performed in the same flow as the learning process S613. In that case, verification data having the same data structure as the training data 114 is generated, and processing is performed using the data.

FIG. 7 shows a process in which the incoming call response business management system 1 inputs an explanatory variable into the incoming call response status prediction model 116, obtains an objective variable, and predicts the incoming call response status in the predicted time zone of the prediction target day (hereinafter, “incoming call”. It is a flowchart explaining "correspondence situation prediction processing S700".

First, the incoming call response status prediction unit 170 receives the attribute information 71 of the delinquent, the prediction target date, and the prediction time zone 72 from the user. Further, the incoming call response status prediction unit 170 acquires the number of operators assigned 73 in the predicted time zone of the received prediction target day from the operator allocation schedule information 115 (S711).

Subsequently, the incoming call response status prediction unit 170 determines whether or not the received prediction target date is today (S712). If the received prediction target date is today (S712: YES), the process proceeds to S713, and if the received prediction target date is not today (S712: NO), the process proceeds to S716.

In S713, the incoming call response status prediction unit 170 determines whether or not there is today's incoming call response status history (additional actual information 117 on the day) that has already been acquired at the present time. If the incoming call correspondence status history exists (S713: YES), the process proceeds to S714, and if the incoming call correspondence status history does not exist (S713: NO), the process proceeds to S716.

In S714, the learning data generation unit 150 generates the learning data 114 based on the additional performance information 117 on the day. Subsequently, the incoming call response status prediction model learning unit 160 learns the incoming call response status prediction model 116 based on the generated learning data 114. As a result, the incoming call response status prediction model 116 reflects information on sudden changes in work load due to reasons such as "sudden absenteeism on the day" and "unexpected incoming call (incoming call) on the day", and the prediction accuracy is improved. It can be improved. After that, the process proceeds to S716.

In S716, the incoming call response status prediction unit 170 inputs the explanatory variables received in S711 into the incoming call response status prediction model 116, predicts the incoming call response status in the predicted time zone of the prediction target day, and uses the result as the prediction result 118. Generate.

Subsequently, the alarm output unit 190 generates a screen (hereinafter referred to as “prediction result presentation screen 800”) describing the contents of the generated prediction result 118, and immediately displays the generated prediction result presentation screen 800 to the user. Present (S716). Further, the alarm output unit 190 determines whether or not it is necessary to output an alarm based on the prediction result 118, and if necessary, displays the alarm on the prediction result presentation screen 800 to alert the user. The alarm output unit 190 issues an alarm when, for example, a large number of call losses are expected to occur based on the predicted incoming call response status (for example, when the predicted call loss exceeds a preset threshold value). It is displayed on the prediction result presentation screen 800.

FIG. 8 shows an example of the prediction result presentation screen 800. As shown in the figure, the exemplary prediction result presentation screen 800 has a display column 811 for the prediction result 118 and an alarm display column 812.

As described above, the incoming call response business management system 1 of the present embodiment accurately obtains the number of incoming calls and the number of responses in the prediction target time zone of the prediction target day by the incoming call response status prediction model 116 and presents them to the user. .. Therefore, the user can grasp the number of incoming calls and the number of responses in the predicted target time zone of the predicted target day, and can allocate an appropriate number of operators to efficiently and appropriately perform the incoming call response work.

In addition, the incoming call response business management system 1 obtains the attribute information of the person who received the incoming call on the day of the prediction target, the number of operators assigned in the time zone when the incoming call was received, and the number of incoming calls and the number of responses in the above time zone. The incoming call response status prediction model is learned from the included learning data, and sudden changes in work load due to reasons such as "sudden absence on the day" or "unexpected incoming call (incoming call) on the day" are reflected in the incoming call response status prediction model. Therefore, the number of incoming calls and the number of responses in the future time zone of the day to be predicted can be accurately obtained, and the user can secure an appropriate number of operators and efficiently perform the incoming call response work.

Further, when the number of call losses in the prediction target time zone exceeds the preset threshold value, the incoming call response business management system 1 immediately presents the information indicating the warning to the user, so that there is a possibility that many call losses occur to the user. Can quickly inform you that there is. Therefore, the user can promptly detect that a large number of call losses may occur and take necessary measures in advance.

Although the embodiments of the present invention have been described in detail above, the above description is for facilitating the understanding of the present invention and does not limit the present invention. The present invention can be modified and improved without departing from the spirit thereof, and it goes without saying that the present invention includes an equivalent thereof. For example, the above-described embodiment has been described in detail in order to explain the present invention in an easy-to-understand manner, and is not necessarily limited to the one including all the described configurations. Further, it is possible to add, delete, or replace a part of the configuration of the above embodiment with another configuration.

For example, the application target of the incoming call response business management system 1 described above is not necessarily limited to the electric power equipment managed by the electric power company, and the incoming call response business management system 1 of the present embodiment is, for example, a gas company, a telephone company, or the like, and is in arrears. It can be applied to various incoming call response operations in organizations that provide social infrastructure that require the management of people.

1 Incoming call response business management system, 110 storage unit, 111 subscriber information, 112 delinquent information, 113 incoming call response status history, 114 learning data, 115 operator placement schedule information, 116 incoming call response status prediction model, 117 additional actual information on the day, 118 Prediction result, 120 Contractor information management department, 125 Delinquent information management department, 130 Incoming call correspondence status history management department, 140 Operator placement schedule management department, 150 Learning data generation department, 160 Incoming call correspondence status prediction model learning department, 170 Incoming call Correspondence status prediction unit, 180 prediction result presentation unit, 190 alarm output unit, S600 prediction model learning process, S700 incoming call response status prediction process, 800 prediction result presentation screen

Claims (10)

It is an information processing system that manages operations for responding to incoming calls.
By inputting the attribute information of the person who may receive a call, the prediction target date, the prediction target time zone of the prediction target date, and the number of operators scheduled to be assigned in the prediction target time zone as explanatory variables. An incoming call response status prediction model storage unit that stores an incoming call response status prediction model, which is a machine learning model that outputs the number of incoming calls and the number of responses in the prediction target time zone as objective variables.
Attribute information of a person who has received a call in a predetermined time zone on a predetermined day in the past, the number of operators assigned in the predetermined time zone on the predetermined day, and the number of incoming calls and responses in the predetermined time zone on the predetermined day. The incoming call correspondence situation prediction model learning unit that learns the incoming call correspondence situation prediction model by the learning data including
An incoming call response status prediction unit that inputs the explanatory variables into the incoming call response status prediction model to obtain the objective variable, and
A business management system that supports incoming calls. The incoming call response business management system according to claim 1.
When the prediction target date is the current day, the incoming call response status prediction model learning unit has the attribute information of the person who has already received the incoming call, the number of operators arranged in the time zone when the incoming call was received, and the time. The incoming call response status prediction model is learned from the learning data including the number of incoming calls and the number of responses in the band.
The incoming call response status prediction unit inputs the explanatory variables into the incoming call response status prediction model after learning, and obtains the objective variable in a predetermined time zone in the future of the prediction target date.
Incoming call support business management system. The incoming call response business management system according to claim 1.
The incoming call response status prediction unit obtains the excess / deficient number of operators in the prediction target time zone of the prediction target day based on the obtained objective variable.
Incoming call support business management system. The incoming call response business management system according to claim 2.
The incoming call response status prediction unit obtains the excess / deficient number of operators in the future prediction time zone of the prediction target day based on the obtained objective variable.
Incoming call support business management system. The incoming call response business management system according to claim 1 or 2.
Based on the objective variable obtained by the incoming call response status prediction unit, it is determined whether or not the number of call losses in the prediction target time zone exceeds a preset threshold value, and if it is determined, a warning is given. Warning output section, which outputs immediately
A business management system that supports incoming calls. It is a method of managing operations for responding to incoming calls.
Information processing equipment
By inputting the attribute information of the person who may receive a call, the prediction target date, the prediction target time zone of the prediction target date, and the number of operators scheduled to be assigned in the prediction target time zone as explanatory variables. A step of storing an incoming call response status prediction model, which is a machine learning model that outputs the number of incoming calls and the number of responses in the prediction target time zone as objective variables, and
Attribute information of a person who has received a call in a predetermined time zone on a predetermined day in the past, the number of operators assigned in the predetermined time zone on the predetermined day, and the number of incoming calls and responses in the predetermined time zone on the predetermined day. The step of learning the incoming call response status prediction model using the training data including
The step of inputting the explanatory variable into the incoming call response status prediction model to obtain the objective variable, and
How to manage incoming calls. The incoming call response business management method according to claim 6.
The information processing device
If the forecast target date is the current day,
The incoming call response status prediction model is obtained from the learning data including the attribute information of the person who has already received an incoming call, the number of operators arranged in the time zone when the incoming call was received, and the number of incoming calls and the number of responses in the time zone. Steps to learn and
A step of inputting the explanatory variables into the incoming call response status prediction model after learning to obtain the objective variable in a predetermined time zone in the future of the prediction target date, and
A business management method for incoming calls that further executes. The incoming call response business management method according to claim 6.
A step of obtaining the excess / deficient number of operators in the prediction target time zone of the prediction target day based on the obtained objective variable by the information processing apparatus.
A business management method for incoming calls that further executes. The incoming call response business management method according to claim 7.
A step in which the information processing apparatus obtains the excess / deficient number of operators in the future prediction time zone of the prediction target date based on the obtained objective variable.
A business management method for incoming calls that further executes. The incoming call response business management method according to claim 6 or 7.
Based on the obtained objective variable, the information processing apparatus determines whether or not the number of call losses in the prediction target time zone exceeds a preset threshold value, and if it determines that the number exceeds, immediately provides information indicating a warning. Step to output,
A business management method for incoming calls that further executes.

Images