JP6035404B1 - Visit preparation system - Google Patents

Abstract

A visit preparation system for supporting life insurance sales is provided. SOLUTION: Customer attributes, customer contract statuses, customer preferences, salesperson activity history of salespersons and messages for salespersons are stored in a database for a contracted customer and a new customer, and a probability model is created. And calculate the predicted contract probability of multiple product patterns for each contracted customer and new customer, priorities determined based on customer preferences, customer attributes and customer contract status, Based on the predicted contract probability, a plurality of messages for sales representatives are selected from the message table. [Selection] Figure 8

Description

  The present invention relates to a visit preparation system, and more specifically, to a visit preparation system for supporting life insurance sales.

  With the recent increase in capacity of storage media such as databases accompanying the development of science and technology, the information obtained from a large amount of accumulated data is enormous, diverse and complicated. For this reason, data that is difficult to handle with the conventional statistical analysis method is mixed.

  Here, the sales activities of life insurance products were collected in the first process of collecting information necessary for contracts with existing customers and new customers. The third process for making a rough proposal for insurance products based on information, the fourth process for making proposals for more specific and detailed products by grasping customer needs according to the proposal, and the conclusion of a final contract ( And a fifth process for performing (closing).

  In general, in marketing activities such as marketing, it is necessary to draw up useful information from these diverse and complex data in various types of databases, but it is necessary to formulate a plan. In insurance product sales activities, we depended on the skills and experience of individual sales staff for individual and specific insurance sales for individual customers, so there are enough matters to be proposed and to be heard. In some cases, it could not be confirmed and sales activities became inefficient. Furthermore, entrusting the contents of activities to individual sales staff has also had the negative effect of making it difficult for management to provide efficient guidance.

  In view of this situation, a marketing concept such as Customer Relationship Management (CRM) has emerged. In CRM, it is necessary to identify the characteristics of each individual customer, and data mining is attracting attention. Yes. Data mining is a technology that analyzes a large amount of data using data analysis techniques such as statistics, pattern recognition, and artificial intelligence, and finds hidden correlations and meanings. Is an analysis method that extracts data from data.

  For example, Patent Document 1 teaches a method and system for calculating the expected degree of contract that each salesperson has determined based on his / her own experience based on an objective measure. In the sales support system of Patent Document 1, the contact amount acquisition unit 211 acquires the contact amount, which is the number of times the sales person contacts the customer (including uncontracted and already contracted), and the amount of information For each customer, the acquisition unit 212 acquires the information amount of information related to the customer collected by the sales person from the customer, and the expectation degree determination unit 213 expects the contract to be concluded for the customer according to the contact amount and the information amount. Is determined. The system and method disclosed in Patent Document 1 uses the calculated expected degree of contract as a guideline for subsequent sales activities for each customer.

JP 2013-200655 A

  However, in the system and method of Patent Document 1, each salesperson decides a sales policy based on his / her own experience after obtaining the expected degree of contract establishment, and the matters and interviews that must be proposed must be conducted. Insufficient confirmation of matters that must be done may result in inefficient sales activities. Furthermore, since the contents of activities are entrusted to individual sales representatives, it is difficult to provide efficient guidance by management.

  On the other hand, in CRM, it is necessary to maintain a good relationship with customers and to create a mechanism that can respond to each customer's request, and it is necessary to identify the differences between individual customers. Therefore, it is meaningless to determine a sales policy in a uniform manner regardless of the difference between each customer as in Patent Document 1, and when determining a specific sales policy for each customer, the customer attributes and It is necessary to customize matters to be addressed and proposals based on the insurance examination stage.

  Therefore, in order to solve the above problems, by analyzing and aggregating the information necessary for determining the sales policy and presenting it to the sales staff, it is possible to avoid relying on the skills and experience of individual sales staff, and It is necessary to build a system that can make sales policy decisions that match the characteristics of each customer.

According to a first aspect of the present invention, the present invention is a sales support system that is connected to at least one client terminal via a communication network and generates a sales strategy for a plurality of product patterns,
For each of a contracted customer who has contracted at least one of the plurality of product patterns and a new customer to whom none of the plurality of product patterns is contracted, a customer attribute, A customer account that stores customer contract status and customer preferences, and each sales representative uniquely associated with each of at least one salesperson assigned an identifier uniquely associated with the device identifier of the client terminal A database that stores a person account that stores the person person activity history of the person, and a message table that stores messages for the sales person,
Based on the customer attributes of the customer account and the customer contract status, a probability model is generated for each of the plurality of product patterns to calculate a predicted contract probability of the product pattern by the existing contract customer and the new customer, and the probability A model generation unit that calculates the predicted contract probability of each product pattern for each of the contracted customer and the new customer using a model;
The priority determined based on the customer preference of the customer account, the customer attribute, and the customer contract status, the person activity history of the person account, and the at least one based on the predicted contract probability A sales support system including a message determination unit that selects a plurality of messages from the message table for each person in charge of sales.

According to the second aspect of the present invention, the present invention relates to an existing contract customer connected to at least one client terminal via a communication network and making a contract with at least one of a plurality of product patterns, and the plurality For each new customer that has no contract with any of the product patterns, a customer account that uniquely associates with each customer, stores customer attributes, customer contract status, and customer preferences, and a unique device identifier for the client terminal A person account that uniquely associates with each of the sales representatives assigned the identifiers associated with the sales representative, and stores a person activity history of the sales representative, and a message table that stores messages for the sales representative A plurality of product patterns using a sales support system including a database that stores data, a model generation unit, and a message determination unit A sales support method for generating a marketing strategy for,
A probability model in which the model generation unit calculates a predicted contract probability of the product pattern by the existing contract customer and the new customer for each of the plurality of product patterns based on a customer attribute and a customer contract status of the customer account Generating the predicted contract probability of each product pattern for each of the contracted customer and the new customer using the probability model;
Based on the priority determined based on the customer preference of the customer account, the customer attribute, and the customer contract status, the person activity history of the person account, and the predicted contract probability And a step of selecting a plurality of messages from the message table for each of the at least one sales representative.

According to a third aspect of the present invention, the present invention is executed by a computer, connected to at least one client terminal via a communication network, and has a contract with at least one of a plurality of product patterns. A customer account that stores a customer attribute, a customer contract status, and a customer preference uniquely associated with each customer for each of the contracted customer and each of the new customers that are not contracted with the plurality of product patterns, and the client A person account that uniquely associates with each salesperson assigned an identifier that is uniquely associated with the device identifier of the terminal, and stores the person person activity history of the person in charge; A sales support system including a database for storing a message table for storing messages, a model generation unit, and a message determination unit In-time, a software program to be executed by a sales support method for generating a marketing strategy for the commodity pattern of the plurality, the method comprising,
A probability model in which the model generation unit calculates a predicted contract probability of the product pattern by the existing contract customer and the new customer for each of the plurality of product patterns based on a customer attribute and a customer contract status of the customer account Generating the predicted contract probability of each product pattern for each of the contracted customer and the new customer using the probability model;
Based on the priority determined based on the customer preference of the customer account, the customer attribute, and the customer contract status, the person activity history of the person account, and the predicted contract probability And selecting a plurality of messages from the message table for each of the at least one sales representative.

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

  According to the present invention, by analyzing and aggregating information necessary for determining a sales policy and presenting it to each sales representative (hereinafter referred to as “agent”), without depending on the skills and experience of individual agents, In addition, it is possible to determine a business policy unique to each customer. This makes it possible to flexibly customize matters to be dealt with and proposal contents specific to each customer.

  For each customer, the probability that the customer will subscribe based on the customer's attributes, the subscription probability calculated based on the subscription trend of the contracted customer, and the message priority for the product with the highest probability that the customer will subscribe By displaying the products with the highest price, the items to be dealt with and the proposal method (message), and the subscription probability on one screen, it is possible to clarify the points that each agent should act on, and to build a trust relationship with each customer Is possible.

  Furthermore, by clarifying the business policy specific to each customer, guidance by the management layer becomes easy.

It is a figure which shows the network structure of the insurance sales support system according to this invention. It is a figure which shows the hardware constitutions of a client terminal. 2 is a diagram illustrating a hardware configuration of a firewall 104 and a Web server 106. FIG. It is a figure which shows the hardware constitutions of the application server. 2 is a diagram illustrating a hardware configuration of a database server 110. FIG. It is a conceptual diagram of the generation of probabilistic model that outputs the prediction contract probability about the product pattern y _in. It is a figure which shows a decision tree model. It is a flowchart which shows the sales support method according to this invention. It is a flowchart which shows the learning algorithm of the decision tree model performed by step S812 of FIG. It is a diagram showing a state of calculation of the predicted contract probabilities for a specific product pattern y _in which are using the decision tree model generated in accordance with the present invention. It is a figure which shows the visit preparation screen before and behind hearing of four security needs (customer preference) of the customer i by the person in charge j. It is a figure which shows the general view of message selection. It is a figure which shows the visit preparation screen of client terminal 114-j. It is a figure which shows switching of a visit preparation screen and a visit preparation sheet. It is a figure which shows the detail of the visit preparation sheet 1400. FIG. It is a figure which shows the flow of a process of the database server 110 at the time of producing | generating the visit preparation screen 1300. It is a figure which shows the flow of a process of the database server 110 at the time of producing | generating the visit preparation screen 1300. It is a figure which shows the flow of a process of the database server 110 at the time of producing | generating the visit preparation screen 1300. It is a figure which shows the flow of a process of the database server 110 at the time of producing | generating the visit preparation screen 1300. It is a figure which shows the "appoint of appointment and maintenance" message table stored in the message database 514. FIG. It is a figure which shows the "item to be heard" message table stored in the message database 514. FIG. 5 is a diagram showing a “proposed point of cut” message table stored in a message database 514;

  While the invention is susceptible to various modifications and alternative forms, exemplary embodiments of the invention are shown by way of example in the drawings and are herein described in detail. However, the description herein of exemplary embodiments is not intended to limit the invention to the particular forms disclosed, but is intended to be construed as limiting the invention as defined by the claims. Includes all modifications, equivalents, and alternatives falling within the spirit and scope.

  In the following, as an example, a sales support system according to an embodiment of the present invention will be described using an example of application to a system that supports customer relationship management (CRM) in a life insurance company. However, the sales support system of the present invention is not limited to customer relationship management in a life insurance company, but also applies to customer relationship management in a securities company, customer relationship management in a bank, and customer relationship management in a medical institution received by a user. Is possible. The customer (user) does not currently have any insurance products with existing customers who already have insurance products of the life insurance company and other life insurance companies, but in the future Includes potential new customers who may sign insurance products.

(1) Overall Network Configuration FIG. 1 is a diagram showing a network configuration of an insurance sales support system according to the present invention.

The insurance sales support system 100 includes a plurality of client terminals 114-j (j = 1,..., N _ag ), a communication network 112, and a central management center 102. The central management center 102 includes a firewall 104, a web server 106, an application server 108, and a database server 110. The plurality of client terminals 114-j are owned by a sales person and are used for receiving visit preparation data indicating a sales policy from the application server 108 and the database server 110. Here, the index j (j = 1,..., _Nag ) of the client terminal corresponds to the person-in-charge ID of the salesperson.

  The communication network 112 is a circuit switched network such as a public switched telephone network (PSTN) or a packet based network such as the Internet. Communication network 112 includes any network that provides data communication via wireless signals, at least in part. Optional networks include, for example, wireless local area networks (WLANs) that use WiFi / WiMax communications. The internal networking of the communication network 112 may use a technology such as the IMS architecture described in 3GPP. An exemplary implementation of any network may be as a corporate or campus corporate network or as a “hot spot” in a common public place such as an airport, coffee shop, etc. The communication network 112 is normally used for data communication, but may be a network that can transfer voice communication such as a VoIP call. The communication network 112 further includes one or more access points that provide a wireless connection to the plurality of client terminals 114-j. Access points include base stations, base station routers, and access networks.

  The client terminals 114-j are smartphones, tablet terminals, cellular phones, personal digital assistants, text messaging devices, pagers, network interface cards, notebook computers, desktop computers, and personal digital assistants (PDAs). May be. Client terminals 114-j can communicate by exchanging information over an air interface (or wireless communication link) that includes several channels, such as a traffic channel, a signaling channel, a paging channel, and the like.

  The air interface channel is defined according to one or more wireless communication protocols used by the wireless communication system 112. For example, an air interface channel operating according to code division multiple access (CDMA) is defined by an orthogonal code that modulates a radio signal used to transmit information over the air interface. The air interface channel may also be determined by the frequency of the carrier used to transmit information over the air interface. For example, in orthogonal frequency division multiple access (OFDMA), one or more client terminals 114-j may share multiple orthogonal frequencies or tones.

  The central management center 102 may be a server room of a life insurance company. The firewall 104 operates in the transport layer, which is an upper protocol of IP, and is used for communications that require reliability to perform window control, retransmission control, congestion control for Web browsing, mail transmission / reception, file transfer, etc. TCP (Transmission Control Protocol), which is a connection-type protocol used, operates in the same way as the transport layer, which is an upper protocol of IP, and is used for real-time operations such as voice calls, video streaming, multicast communication, broadcast communication, and small amounts of data transfer. It is configured to determine whether communication is permitted or not based on the condition of UDP (User Datagram Protocol), which is a connectionless protocol used for the communication. The Web server 106 receives an HTTP request from the Web browser of the client terminal 114-j and returns a response. The application server 108 executes business logic in response to a processing request from the Web server 106 and makes a data reference or update request to the database server 110 if necessary. The database server 110 mainly manages data, executes a data reference or update processing request from the application server 108, and returns a result.

  When the application server first receives a request from the browser of the client terminal 114-j, the application server generates an identifier for uniquely identifying the browser and passes the identifier to the Web application on the Web server 106. After that, the application server holds information on the Web browser in an external resource that is sharable with the database server 110 and is permanent, using the identifier as a key. The Web server 106 passes the identifier received from the application server 108 to the browser of the client terminal 114-j. For example, when cookie is used, an identifier can be set for the cookie by the application server 108 by passing the cookie as an argument when the Web server 106 and the application server 108 are called. When a request is made again from the browser of the client terminal 114-j, information can be acquired by referring to the external resource based on the identifier set by the application server 108.

(2) Client Terminal FIG. 2 is a diagram illustrating a hardware configuration of the client terminal.

  The client terminal 200 includes a CPU 202, a memory 204, a storage device 206, an input unit 208, a display unit 210, and a transmission / reception interface 212. The CPU 202 implements various functions by reading the program stored in the storage device 206 into the memory 204 and executing it. The memory 204 includes, for example, a random access memory (RAM) and / or a read only memory (ROM). Storage device 206 includes a tape drive, floppy drive, hard disk drive, or compact disk drive. The input unit 208 receives data input, and includes, for example, a keyboard, a mouse, a trackball, a touch panel, a microphone, and the like. The display unit 210 includes, for example, a display and a speaker that output data. The client terminal can print out visit preparation data indicating a business policy from the application server 108 and the database server 110 as a visit preparation sheet by connecting the transmission / reception interface 212 to an external printer or the like.

  The transmission / reception interface 212 accesses the Web server 106 via the communication network 112 and transmits an HTTP request to the Web server 106. In addition, the transmission / reception interface 212 transmits to the Web server 106 the person-in-charge information such as the person-in-charge ID and password input on the Web browser displayed on the display unit 210 in response to the response to the HTTP request from the Web server 106. Send. This completes the login by the person in charge to the sales support system.

(3) Firewall and Web Server FIG. 3 is a diagram illustrating a hardware configuration of the firewall 104 and the Web server 106.

  The firewall 104 and the Web server 106 include a CPU 302, a memory 304, a storage device 306, an input unit 308, an output unit 310, and a transmission / reception interface 312. The storage device 306 may be a hard disk drive, a solid state drive, a flash memory, or the like that stores various data and programs. The CPU 302 implements various functions by reading the program stored in the storage device 306 into the memory 304 and executing it. The transmission / reception interface 312 is an interface for connecting to the communication network 114, and is an adapter for connecting to Ethernet (registered trademark), a modem for connecting to a public telephone line network, or the like. The input unit 308 may be a keyboard, a mouse, a trackball, a touch panel, a microphone, or the like that receives data input. The output unit 310 may be a display, a printer, a speaker, or the like that outputs data.

(4) Application Server FIG. 4 is a diagram illustrating a hardware configuration of the application server 108.

  The application server 108 includes a CPU 402, a memory 404, a storage device 406, an input unit 408, an output unit 410, a transmission / reception interface 412, a model generation unit 414, and a message determination unit 416. The CPU 302, the memory 304, the storage device 306, the input unit 308, the output unit 310, and the transmission / reception interface 312 are the same as those of the firewall 104 and the Web server 106 described above.

The model generation unit 414 stores customer attributes in customer accounts stored for each “customer ID” i (i = 1,..., N _c ) in a contract customer database 502 of the database server 110 described later, and Model capable of calculating the predicted contract probability of each product pattern y _in for each customer i based on the customer contract status in the customer account stored for each “customer ID” i in the contract status database 504 Is generated. The customer attribute F _i includes the age, sex, address, marital status, occupation, annual income, place of work, place of work, household composition, and the like of the customer. Customer contract status CS _i is the product-specific subscription status (life insurance, cancer insurance, one-time payment, deferred pension, non-life insurance, car insurance, life insurance of other companies), insurance (daily, monthly, annual), insurance maturity period, payment period Including insurance premiums. The product pattern y _in may consist of one insurance product or a plurality of insurance products. For example, there may be cases where special provisions such as three major diseases are attached to death protection.

The predicted contract probability is obtained by acquiring the customer attribute F _i and the customer contract status CS _i of the contracted customer database 502 and the contract status database 504 as a learning sample S, and using the learning sample S, a prediction model is obtained by a statistical algorithm. It is calculated | required by producing | generating. Here, the customer attribute F _i and the customer contract status CS _i are collectively expressed as x _i = [F _i , CS _i ]. Statistical algorithms include principal component analysis, regression analysis, clustering, decision tree algorithms, and the like. The probability model that outputs the predicted contract probability p (y _in , x _i ) = p (y _in | x _i ) · p (x _i ) includes a decision tree, a neural network, a hidden Markov model, and the like. The probability model generated by learning using the learning sample S is applied to both the contracted customer and the new customer, and predicts the product pattern y _in having the highest contract probability.

The message determination unit 416 selects a plurality of messages that are output as part of the visit preparation data and displayed on the visit preparation screen and the visit preparation sheet. The plurality of messages include the predicted contract probability p (y _in , x _i ) predicted by the above probability model, the customer attribute F _i , the customer contract status CS _i , the customer preference (preference), and the person-in-charge activity history H _j . Selected based on. An approximate idea of message selection is shown in FIG.

  If the customer preference is “basic” for a new customer, a message in accordance with a typical model plan is displayed. For a new customer, if the customer preference is “serious illness and care needs” or “medical needs”, select the message to appeal for the new product with the highest priority, and select 2 Select to display the second highest priority, and select the third highest priority to display a message that also appeals the death protection and savings parts required for the segment. For new customers, if the customer preference is “death needs”, select the message that appeals for the security product with the highest priority, and the message that appeals for pre-life benefits parts with the second highest priority. Select to display and select to display a message appealing for the new product with the third highest priority. If there is no security product to appeal, select to display a message to promote savings products such as life, pension, and meiste. For new customers, if the customer preference is “Assets and Retirement Needs”, select the message to appeal for savings products with the highest priority and the second message to appeal for other savings products To display a message that appeals for a new product to be displayed with the third highest priority. In the case of contracted customers, products that have already been subscribed regardless of whether the customer preference is “basic”, “serious illness and care needs” or “medical needs”, “death needs”, “assets and retirement needs” No message will be displayed for those who are not covered by the target insurance company or other companies, and a message to appeal for other products will be selected to be displayed as the new message with the highest priority. If the major illness protection is less than 9 million, choose to display the shortage appeal message with the second highest priority, and if the contracted 3 major illness protection is more than 9 million, Choose to display a message appealing for a security review, with the third highest priority, appealing to over 25 million guarantees for customers with one child, considering education costs Message, the child is 2 or more customers to choose in order to display a message that appeals to more than 30 million guarantee.

At this time, for a plurality of selected messages, the person account of the corresponding “person ID” j (j = 1,..., N _ag ) stored in the activity history database 512 of the database server 110 is referred to. Here, it should be noted that a message that has already been executed by the person in charge j is excluded from the display target, and instead, a message with a priority lower by one is displayed. On the other hand, among the messages displayed in a certain time period t−1, the fact that the person in charge j has executed the message executed after that by the person in charge j is reflected in the activity history database 512 of the database server 110. And reflected in the customer account of customer i in the contract status database 504. Then, when the visit preparation data is updated in the time period t, the executed message is deleted, and instead, a message having a lower priority is moved up and displayed.

(5) Database Server FIG. 5 is a diagram illustrating a hardware configuration of the database server 110.

The database server 110 includes a contracted customer database 502, a contract state database 504, a new customer database 506, a household information database 508, a person-in-charge database 510, an activity history database 512, a message database 514, and a prediction model database 516. The contracted customer database 502, the contract status database 504, the new customer database 506, and the household information database 508 form a customer information area and are associated with “customer ID” i (i = 1,..., N _c ). Stores accounts. The person-in-charge database 510 and the activity history database 512 form a sales person-in-charge area, and store an account of each person in charge in association with “person-in-charge ID” j (j = 1,..., N _ag ).

The account of the customer i in the contracted customer database 502 includes, for example, a customer ID for identifying the customer, a household number indicating the customer's household, the date on which the account description is recorded in the contracted customer database 502, the customer i Including the presence or absence of contact information. The account of the customer i can also include information related to search information, such as whether or not the search information has been registered via an online service such as a web page. Further, the presence / absence of a contact can include the presence / absence of a home phone number, the presence / absence of a mobile phone number, the presence / absence of a work phone number, the presence / absence of a mail address, and the like. In addition, the account of the customer i includes the age, sex, address, marital status, occupation, annual income, place of work, place of work, household composition, and the like of the customer i. These are called customer attributes and are represented by variables F _i .

The account of the customer i in the contract status database 504 includes, for example, information related to insurance products that the customer i has already contracted. The account of customer i includes, for example, whether there is an insurance contract in which customer i is a contracting party, whether there is an insurance contract in which customer i is an insured person, endowment insurance or annuity in which customer i is a contractor Whether there is a contract for type products, whether there is a contract for endowment insurance or pension type products for which customer i is set as an insured, and whether there is a contract for other insurance products for which customer i is a contractor or insured Can be included. In addition, the account of customer i is, for example, whether there is a contract for which the prospective customer is a contractor or insured for variable-value products, deferred products, annuities after the start of payment, etc. It is also possible to include information on the attributes of already contracted products, such as whether there is a contract for which the customer is an insured person. These are called customer contract status and are represented by the variable CS _i .

  The account of customer i in the contract status database 504 can include, for example, information regarding after-sales service provided to the contracted customer i. The account of the customer i in the contract status database 504 is, for example, the points and miles (use) that are issued according to the date of application for the present from the customer i (present application date), the contract period, the insurance content, etc. Mileage), points and the start date of the valid period of the mile (mileage product effective start date).

  The account of customer i in the contract status database 504 can also store information indicating events related to customer i. The account of the customer i in the contract status database 504 includes, for example, the date of birth for calculating the age of the customer i, the date of renewal of the non-life insurance of the other company with which the customer i is contracted (the non-life insurance renewal date) ), The date of renewal of the company's non-life insurance for which the customer i is contracted (in-house non-life insurance renewal date), and the like.

  The account of customer i in the new customer database 506 stores information on a potential customer i who has not yet reached a contract. The account of the customer i in the new customer database 504 is “customer ID” i, the household number indicating the household of the customer i, the date when the information is recorded in the account of the customer i in the new customer database 504, and the customer i. Stores the presence or absence of contact information. In addition, the account of customer i in the new customer database 504 is associated with “customer ID” i for identifying the customer, and whether or not the information is registered via an online service such as a web page, etc. Information about search information can be included. In addition, the presence / absence of a contact can include the presence / absence of a home phone number, the presence / absence of a mobile phone number, the presence / absence of a work phone number, the presence / absence of a mail address, and the like.

  The household information database 508 stores “customer ID” i of the customer and information about the household of the customer for both the contracted customer and the new customer. Household information includes a household number that identifies the household, the name of the householder, date of birth, gender, telephone number, and the like. In addition to these, for example, various information about the household such as family name and date of birth may be included.

  The account of the person in charge j in the person-in-charge database 510 stores the person-in-charge ID, a code for identifying the area in charge of the person in charge j, the device identifier of the terminal carried by the person in charge.

The account of the person in charge j in the activity history database 512 stores information related to the action from the sales person in charge j to the customer i (i = 1,..., N _c ). The account of the person in charge j in the activity history database 512 is, for example, the date of the contact that the sales person j has made to the customer i most recently, and the date on which the sales person j had the most recent interview with the customer i ( Date of latest interview 2), date of previous interview with date of latest interview 2 (latest interview date 1), year of confirmation activity most recently conducted by sales person for prospective customer It is possible to include the date (the date of the most recent personal identification activity). In addition, the account of the person in charge j in the activity history database 512 can include various kinds of information related to the approach from the sales person in charge j to the customer i.

  The account of the person in charge j in the activity history database 512 can also store information related to information collection by the person in charge j. The account of the person in charge j in the activity history database 512 may include, for example, the date (date of latest information collection activity) when the sales person j recently collected information. In addition, the account of the person in charge j in the activity history database 512 is whether the customer i is an individual or a family member, is single, has other companies 'life insurance, or has other companies' non-life insurance. It is possible to include information regarding the customer i. Furthermore, the account of the person in charge j in the activity history database 512 can include information indicating whether or not points, miles, or the like have been given as consideration for collecting information.

  The account of the person in charge j in the activity history database 512 is the date when the sales person j plans to bring the proposal to the customer i (the next visit date), and the person in charge of the sales person j is closest to the customer i. The date of the proposal for the insurance product (the date of the latest proposal activity), the date of the most recent proposal by the sales representative j (the date of proposal creation 2), the date of the proposal creation The date when the proposal was created one day before day 2 (proposal creation date 1), and the response of customer i to the latest proposal (customer response life insurance at the time of the latest proposal) can be included. Customer Response Life Insurance at the time of the most recent proposal is a “contract” indicating that it wants to take out insurance with the proposal content, a “visit” indicating that a proposal for another product or a change in coverage, a change in insurance premium, or other requests are required In addition, it is possible to include either “consideration” indicating that the examination is to be performed again or “refusal” indicating that the insurance is not taken out.

Predictive model database 516 is generated by the model generation unit 414 of the application server 108, _x summarizes customer attribute _{F i} and customer contract status _{CS i i = [F i,} CS i] about the product pattern _{y in} based on It stores a probability model that outputs the predicted contract probability p (y _in , x _i ) = p (y _in | x _i ) · p (x _i ). The probability model may be a decision tree, a neural network, a hidden Markov model, or the like. For example, in the case of a decision tree model, the number of decision trees k (k = 1,..., N _T ), the tree depth D _k , the branch function f (v) for the branch node (Split Node), and the feature quantity The threshold value A _{TH of} F _{i and} the probability p (c _u | L) that the data d _{s of the} terminal node (Leaf Node) L belong to each class c _u are stored as parameters. In the case of a neural network, the number of units in the input layer, the number of units in the output layer, the number of intermediate layers, the number of units in each intermediate layer, the coupling coefficient between units in each layer, and the like are stored as parameters. In the case of a hidden Markov model with continuous distribution, the number of states, the average and variance of output symbol probabilities for each state, the average and variance of state transition probabilities for each state, and the like are stored as parameters.

The message database 514 assigns priorities by associating each message with the predicted contract probability p (y _in , x _i ) predicted by the probability model, customer attributes F _i , customer contract status CS _i , and customer preferences. Stored. Messages are categorized into the categories “Appoint and Maintenance”, “Matters to Interview”, and “Proposal”. For example, FIG. 20 shows a message classified as “appoint of appointment and maintenance”, FIG. 21 shows a message classified as “matter to be heard”, and FIG. Indicates a message to be classified.

(6) Prediction Model Generation and Visit Preparation Data Generation FIG. 6 is a conceptual diagram of generation of a probability model that outputs a prediction contract probability for the product pattern y _in .

The probability model is generated by acquiring a sample S of teacher data and learning using the sample S. The sample S of the teacher data targets the customer attribute F _i and the customer contract status CS _i of the customer i (i = 1,..., N _c ) in the contracted customer database 502 and the contract status database 504. First, the customer attribute F _i and the customer contract status CS _i of the contracted customer database 502 and the contract status database 504 of a certain time period t−1 are acquired as a learning sample S, and statistical data is obtained using the learning sample S. Learning is performed by an algorithm, and a prediction model is generated. Here, the period of t may be one hour, one day, or one year. Next, in the time period t, the customer attributes of the accounts of all customers including the contracted customer and the new customer stored in the contracted customer database 502, the contract status database 504, the new customer database 506, and the household information database 508. as the target of the F _i and customer contract status CS _i, using the generated probability model, customer attributes _{F i} and _x summarizes the customer contract status _{CS i i = [F i,} CS i] for, commodity pattern _{y in} The predicted contract probability p (y _in , x _i ) = p (y _in | x _i ) · p (x _i ) is calculated. The probability model that outputs the predicted contract probability p (y _in , x _i ) = p (y _in | x _i ) · p (x _i ) includes a decision tree, a neural network, a hidden Markov model, and the like. That is, the probability model generated by learning using the learning sample S is applied to both the existing contract customer and the new customer, and predicts the product pattern y _in having the highest contract probability. The customer attribute F _i includes the age, sex, address, marital status, occupation, annual income, place of work, place of work, household composition, and the like of the customer. Customer contract status CS _i is the product-specific subscription status (life insurance, cancer insurance, one-time payment, deferred pension, non-life insurance, car insurance, life insurance of other companies), insurance (daily, monthly, annual), insurance maturity period, payment period Including insurance premiums. The product pattern y _in may consist of one insurance product or a plurality of insurance products.

  In the present embodiment, learning and generation of a probability model will be described using a decision tree algorithm as an example.

  FIG. 7 is a diagram illustrating a decision tree model.

The decision tree includes a branch node (Split node) that divides the sample by the feature amount F _i of the sample S and the threshold value A _TH and a terminal node (Leaf node) in which the appearance probability of each class is stored. N _T decision trees are repeatedly created for each k (k = 1,..., N _T ) th decision tree, and a terminal node is created when branching becomes impossible according to a certain criterion. It is built by doing. The branch node branches left and right depending on the magnitude relationship between a certain one-dimensional feature value F _i and the threshold value A _TH . Further, the terminal node L, the data _{d s} end node L probability p belonging to each class _{c u} | Save _(c u L).

  FIG. 8 is a flowchart showing a sales support method according to the present invention.

Processing starts in step S802. First, the model generation unit 414 of the application server 108 acquires the “customer ID” i (i = 1,..., N _c ) of the contracted customer from the contracted customer database 502, and the “person in charge” from the person-in-charge database 510. ID “j” (j = 1,..., N _ag ) is acquired (step S804). The model generation unit 414 obtains the set x _i = [F _i , CS _i ] of the customer attribute F _i and the customer contract status CS _i as the learning sample S from the contracted customer database 502 and the contract status database 504 (step S806). .
The model generation unit 414 constructs a decision tree by using these subsets. The decision tree is constructed by repeatedly creating branch nodes until branching becomes impossible. The branch node determines a branch to the left or right using a certain feature amount F _i and a threshold value A _TH . Therefore, it is necessary to carefully select the feature value F _i and the threshold value A _TH from various candidates. For example, the division into subsets in this case is performed by using the principal component analysis technique, among the feature amounts included in the set x _i = [F _i , CS _i ] of the customer attribute F _i and the customer contract status CS _i . This may be performed by selecting a part from the one having a larger variance given to the distribution of data and then dividing the data into a plurality of subsets by clustering using the selected feature amount. In this case, the division into a plurality of clusters may be performed by at least one or a combination of the k-means method, the Ward method, and the Vision method. The model generation unit 414 sets the subset number to k = 1 (step S810). The model generation unit 414 performs learning for generating N _T decision trees in step S812, step S814, and step S816.

  FIG. 9 is a flowchart showing a decision tree model learning algorithm performed in step S812 of FIG.

  Processing starts in step S902.

The model generation unit 414 of the application server 108 acquires input data (step S904). In this case, the input data includes the learning sample S acquired as the set x _i = [F _i , CS _i ] of the customer attribute F _i and the customer contract status CS _i from the contract customer database 502 and the contract status database 504, and the decision tree. , K (k = 1,..., N _T ), a subset S = [S ₁ , S ₂ ,..., S _NT ] obtained by dividing the learning sample, a tree depth D _k, and a feature selection number l (L = 1,..., N _F ) and threshold selection count m (m = 1,..., N _TH ). The model generation unit 414 sets the decision tree number to k = 1 (step S906). The model generation unit 414 sets the feature quantity selection count to 1 = 1 (step S908). At this time, the model generation unit 414 sets the value of the information gain ΔE to the initial value ΔE ₀ (step S908). The model generation unit 414 randomly selects the branch function f _{(k, l)} (step S910). The model generation unit 414 sets the threshold selection count to m = 1 (step S912). The model generation unit 414 randomly selects the threshold A _TH (step S914).

Model generating unit 414 uses branch function _{f (k, l)} and the threshold value _{A TH,} the data _{S n,} is branched to _{S L} and _{S R} (step S916). Here, a set of data assigned a S _n to the branch node n, the data of the size of the set | S |, n the number of data belonging to (u) to the u-th class (u = 1, ..., NC ) When, entropy E for recognizing the data _{S n (S n)} is
(Step S916). The model generation unit 414 uses the teacher signal to calculate the information gain ΔE,
Model generating unit 414, the threshold value selection frequency m is determined whether less than _{N TH} (step S922), the process, if the threshold number of selections m is less than _{N TH} is the processing of m ← m + 1 Perform (step S924), and the process returns to step S914 again. If the threshold number of selections m is greater than or equal to _{N TH,} the process proceeds to step S926, the model generating unit 414 determines whether the feature quantity selection number l is less than _{N F} (step S926), when the feature quantity selection number l is less than _{N F} performs processing of l ← l + 1 (step S928), processing returns to step S910. When the feature quantity selection number l is greater than or equal to N _F, the process proceeds to step S930, the model generating unit 414, information or gain ΔE is 0, or if the ΔE reaches a predetermined threshold value D Creates a terminal node (step S930). The model generation unit 414 determines whether or not the number k of decision trees is smaller than _NT (step S932). If the number k of decision trees is smaller than _NT , the process of k ← k + 1 is performed. (Step S934), the process returns to step S908 again. If the number k of decision trees is greater than or equal to _NT , the model generation unit 414 determines that all decision trees k (k = 1,..., _NT ) have been generated, and ends learning (step S936). ).

Returning to FIG. 8 again, the model generation unit 414 of the application server 108 sets “customer ID” to i = 1 (step S818). The model generation unit 414 calculates the predicted contract probability p (y _in , x _i ) = p (y _in | x _i ) · p (x _i ) for each product pattern y _in by the generated _NT decision trees. (Step S820). Here, referring to FIG. 10, a state _in which the predicted contract probability p (y _in , x _i ) is calculated for a specific product pattern y _in using the decision tree model generated according to the present invention is shown. . First, the model generation unit 414 uses, as the feature amount F _i , the number of days until the cancellation of the insurance product with which the customer i has contracted, and uses a predetermined branch function f _{(k, l)} and a threshold A _TH. The data element of node 0 is branched into node 1, node 2, and node 3. Next, the model generation unit 414 uses the number of possessed products of the three major disease guarantees as the feature amount F _i , and uses the predetermined branch function f _{(k, l)} and the threshold value A _TH, and the data element of the node 1 Is branched into a node 12 and a node 13. The model generation unit 414 uses the remaining premium payment period as the feature amount F _i , and uses the predetermined branch function f _{(k, l)} and the threshold A _TH to _convert the data element of the node 12 to the end node 46. And branch to the end node 47. On the other hand, the model generation unit 414 uses the household head age as the feature amount F _i , and uses the predetermined branch function f _{(k, l)} and the threshold A _TH to _convert the data element of the node 13 into the end node 48, Branch to the end node 49. In the terminal node 46, the terminal node 47, the terminal node 48, and the terminal node 49, the predicted contract probabilities are calculated as 1.5%, 5.7%, 25.1%, and 14.4%, respectively. In the present embodiment, there are 14 product patterns y _in in total, but the model generation unit 414 generates decision trees for all 14 patterns. That is, in this embodiment, N _T = 14.

The message determination unit 416 of the application server 108 includes the predicted contract probability p (y _in , x _i ) predicted by the probability model, the customer attribute F _i , the customer contract status CS _i , the customer preference, and the person-in-charge activity history H _j . Based on this, a message displayed in the column “Appoint and maintenance” is selected for each “customer ID” i (step S822). The messages displayed in the “Appoint and maintenance section” column are stored in the message database 514 classified into the “Appoint and maintenance section” category and output as part of the visit preparation data. It is displayed on the visit preparation screen and the visit preparation sheet. The message determination unit 416 refers to the “Appoint and maintenance cut” message table of the message database 514 as shown in FIG. 20 and performs the priority shown in the “Appoint and maintenance cut” message table. A plurality of messages are selected in descending order of priority according to the order. At this time, the message determination unit 416 is responsible for the corresponding “person ID” j (j = 1,..., N _ag ) stored in the activity history database 512 of the database server 110 for the selected plurality of messages. Refer to the administrator account. The message determination unit 416 excludes messages that have already been executed by the person in charge j from the display target, and instead displays the messages with a lower priority in order. On the other hand, among the messages displayed during a certain time period, the fact that the person in charge j executed after that is reflected in the activity history database 512 of the database server 110, and the contract state This is reflected in the customer account of customer i in the database 504. Then, when the visit preparation data is updated in the subsequent time period, the message determination unit 416 deletes the executed message, and instead displays the message having a lower priority in order. .

The message determination unit 416 of the application server 108 includes the predicted contract probability p (y _in , x _i ) predicted by the probability model, the customer attribute F _i , the customer contract status CS _i , the customer preference, and the person-in-charge activity history H _j . Based on this, a plurality of messages classified into the “items to be heard” category are selected for each “customer ID” i (step S824). The messages displayed in the “Matters to be heard” column are classified and stored in the “Matters to be heard” category in the message database 514, and are output as part of the visit preparation data to be output as a visit preparation screen and Displayed on the visit preparation sheet. The message determination unit 416 refers to the “matter to be heard” message table of the message database 514 as shown in FIG. 21 and prioritizes according to the priority order shown in the “matter to be heard” message table. Select multiple messages in descending order. Similarly to the above, at this time, the message determination unit 416 refers to the person-in-charge account of the corresponding “person-in-charge ID” j (j = 1,..., N _ag ) for the selected plurality of messages, and the person-in-charge j Is already excluded from the display target, and instead, the message with the lowest priority is displayed in an advanced order. Of the messages displayed during a certain period of time, the fact that the person in charge j executed after that is reflected in the activity history database 512 of the database server 110, and the contract status database It is reflected in the customer account of customer i in 504. Then, when the visit preparation data is updated in the subsequent time period, the message determination unit 416 deletes the executed message, and instead displays the message having a lower priority in order. .

The message determination unit 416 of the application server 108 includes the predicted contract probability p (y _in , x _i ) predicted by the probability model, the customer attribute F _i , the customer contract status CS _i , the customer preference, and the person-in-charge activity history H _j . On the basis of this, a plurality of messages classified into the “suggestion cut” category are selected for each “customer ID” i (step S826). The messages displayed in the “Proposal section” column are classified and stored in the “Proposal section” category in the message database 514, and are output as part of the visit preparation data to be output as a visit preparation screen and a visit preparation. Displayed on the sheet. The message determination unit 416 refers to the “Proposal Cut” message table of the message database 514 as shown in FIG. 22, and determines the priority according to the priority shown in the “Proposal Cut” message table. Select multiple messages in descending order. Similarly to the above, at this time, the message determination unit 416 refers to the person-in-charge account of the corresponding “person-in-charge ID” j (j = 1,..., N _ag ) for the selected plurality of messages, and the person-in-charge j Is already excluded from the display target, and instead, the message with the lowest priority is displayed in an advanced order. On the other hand, among the messages displayed during a certain time period, the fact that the person in charge j executed after that is reflected in the activity history database 512 of the database server 110, and the contract state This is reflected in the customer account of customer i in the database 504. Then, when the visit preparation data is updated in the subsequent time period, the message determination unit 416 deletes the executed message, and instead displays the message having a lower priority in order. .

The message determination unit 416 of the application server 108 displays a visit preparation screen for each “person in charge ID” j using the message group selected above (step S834). In other words, the message determination unit 416 (1) predicts the contract probability p (y _in , x _i ) = p (y _in | x _i ) · p (x _i ) calculated above as the product pattern y _in , ( 2) Determined based on customer attributes F _i , customer contract status CS _i , customer preferences, and person-in-charge activity history H _j , and selected according to the priority shown in the “Appoint and Maintenance” message table “Appoint and maintenance start” message, (3) Determined based on customer attribute F _i , customer contract status CS _i , customer preference, and person-in-charge activity history H _j , and shown in “Matters to be heard” message table and selected the "matters to be hearing," according to the priority is the message, and (4) customer attribute F _i, customer contract status CS _i, customer preferences, and the person in charge activity Is determined on the basis of gravel H _j, the four selected "Proposal of cut" message according to the priority shown in "Proposal of cut" message table, the corresponding "person ID" j (j = 1 ,..., N _ag ) are reflected in the person-in-charge account (step S828).

The message determination unit 416 of the application server 108 determines whether “customer ID” i is smaller than N _c (step S830). If “customer ID” i is smaller than _Nc , message determination unit 416 executes the process i ← i + 1 (step S832), and the process returns to step S820. If “customer ID” i is equal to or greater than N _c , the process proceeds to step S834, and the message determination unit 416 provides visit preparation data to the client terminal 114-j for each corresponding “contact person ID” j. The client terminal 114-j displays the visit preparation screen. In step S836, the process ends.

(7) Display of Visit Preparation Screen FIG. 11 is a diagram showing a visit preparation screen before and after hearing the four security needs (customer preferences) of the customer i by the person in charge j.

  In the screen before listening to the four customer preferences, in the “Proposal section” message field, “Appeal new product”, “Appeal death security for remaining family living expenses, educational expenses, etc.” In addition, three messages are displayed in the order of “Appeal for pension for retirement and future home purchase funds”. On the screen after it turns out that the customer's preference is in the order of old-age assets, serious illness care, death, and medical care, “Appealing pension for retirement and future home purchase funds”, “Lifetime protection” Three messages are displayed in the order of “Appropriate Life, Appealing Mystery” and “Applying New Products (Applying Medical Care)”.

FIG. 12 is a diagram showing an approximate concept of message selection, that is, an approximate concept of how to prioritize messages in each message group. The plurality of messages include a predicted contract probability p (y _in , x _i ) predicted by the above probability model, customer attribute F _i , customer contract status CS _i , customer preference (customer needs), and person-in-charge activity history H _j. Selected based on The priority order is determined based on the customer attribute F _i , the customer contract status CS _i , and the customer preference.

  FIG. 13 is a diagram showing a visit preparation screen of the client terminal 114-j.

FIG. 13 shows the customer “customer ID” i = 4498-1875-120 displayed on the display unit 210 of the client terminal 114-j. In FIG. 13, a message customized for the person in charge j entitled “Development viewpoint” is displayed in an area 1302 in the screen 1300. In this message, it is displayed that the product pattern y _in having the highest predicted contract probability that the customer i contracts is related to nursing care. The message content in this area 1302 includes the predicted contract probability p (y _in , x _i ) predicted by the probability model, the customer attribute F _i of the account of the customer _i , the customer contract status CS _i , the customer preference, the person in charge j It is determined based on the person in charge activity history H _j of the account. Further, the predicted contract probability p (y _in , x _i ) is calculated based on the customer attribute F _i of the account of the customer _i and the customer contract status CS _i . Further, the message in the area 1302 is selected and displayed according to the priority order of the message table, and the priority order is determined based on the customer attribute F _i , the customer contract status CS _i , and the customer preference. Accordingly, if there is a change or update in at least one of the customer i account and the contact j account, the change or update is immediately reflected in the message in area 1302 (ie, in real time).

  FIG. 14 is a diagram illustrating switching between the visit preparation screen 1300 and the visit preparation sheet 1400.

An area 1304 in the visit preparation screen 1300 indicates that the customer is “customer ID” i = 4498-1875-120. The area 1306 includes the name, address, telephone number, e-mail address, business office, business office address, and business office telephone number of the customer i. This information has been retrieved from the contracted customer database 502. It is. An area 1308 is an input area. For example, the items proposed by the person in charge j, the appointment appointment for the next appointment between the person in charge j and the customer i, the information on the establishment of the product contract, the history of sales activities of the person in charge j, and the like are input. The input contents are reflected in the account of the customer i in the contracted customer database 502 or the new customer database 506. An area 1310 indicates the preference of the customer i. For example, in the illustrated example, customer i has a preference for “aged assets” at the first priority, has a preference for “medical” at the second priority, and has a preference for “medical” at the third priority. It has a preference for “serious illness care” and a preference for “death” in the fourth priority. An area 1312 shows household information of the customer i. This information is information retrieved from the household information database 508. Specify another family with the button at the left end of the area 1312 and press the “Prepare in the backboard” button at the lower end of the area 1312 to display an “appointment and maintenance cut” message about the family, “ "Matter to be heard" message, "Proposal point" message, etc. are displayed. An area 1314 indicates existing contract information not only for the customer i but also for other companies. This information has been searched from the contract status database 504, and the subscription status by product (life insurance, cancer insurance, one-time payment, deferred pension, non-life insurance, car insurance, life insurance of other companies), insurance (daily, monthly, Customer contract status CS _i including annual amount), insurance maturity period, payment period, insurance premium, etc.

  The features of this screen are (1) the fact that various customer information is centralized, and (2) the items to be communicated to the customer can be identified at a glance, and the “point of appointment and maintenance”, From the viewpoints of “Matters to be interviewed” and “Proposal point of view”, multiple messages are displayed according to a predetermined priority, and (3) the screen is switched between each family in the household for the purpose of cultivating the household. It is possible. When the “print” button 1318 in the visit preparation screen 1300 is clicked, the visit preparation screen 1300 is switched to the visit preparation sheet 1400.

  FIG. 15 is a diagram showing details of the visit preparation sheet 1500.

The illustrated visit preparation sheet 1400 is for “person in charge ID” j = 1234567 and the person in charge name “Haruko Inoue”. An area 1402 displays customer attributes and customer preferences. An area 1404 displays the company's existing contracted product for the customer i, which is lifetime insurance in the illustrated example. An area 1406 shows a contracted product of another company for the customer i, and in the illustrated example, it is car insurance. An area 1408 shows the household information of the customer i and items to be proposed to the individual families constituting the household. For example, the customer i himself / herself is insured with automobile accident insurance whose renewal period is close to one to three months. In the case of such renewals, insurance premiums often increase. Therefore, the customer i may change to another company's automobile accident insurance with a lower price and the same guarantee. The contract probability in this case can be predicted by the decision tree model according to the present invention, and since the probability is high, it is displayed in the area 1408 as a message having the first priority. The second priority message indicates that the customer i has a life insurance of another company within two years before the renewal. This is the same as above. In addition, the message with the third priority indicates that the child is entering elementary school. An area 1410 displays an “Appoint and maintenance point” message. The message having the first priority indicates that some product should be proposed because there is a mileage that disappears due to the expiration of the period during the year. An area 1412 displays an “item to be heard” message. An area 1412 displays matters that have not been heard from the customer i according to the priority order. An area 1414 is an area for displaying a “suggestion cut” message. The “proposition cut” message in the area 1414 includes the predicted contract probability p (y _in , x _i ) predicted by the probability model, customer attributes F _i such as household information, customer contract status CS _i , customer preference, person in charge j _Are displayed based on the person in charge activity history _Hj .

  FIG. 16 to FIG. 20 are diagrams showing a processing flow of the database server 110 when the visit preparation screen 1300 is generated.

  First, the message determination unit 416 of the application server 108 searches the person-in-charge area of the database server 110 for the “person-in-charge ID” j and the office to which the person-in-charge j belongs, and from the customer information area of the database server 110, The customer “customer ID” i, occupation, occupation, and workplace are searched. Next, from the contract status database 504 in the customer area, the basic information on the existing life insurance policy and the detailed information on the existing life insurance policy, the deferred pension information, the life insurance contract information of the other company, the existing insurance policy information, the existing insurance policy information, the non-life insurance of the other company Search for and display existing contract information. Thereafter, the message determination unit 416 reads from the activity history database 512 in the person-in-charge area, the date and time of the previous appointment between the person in charge j and the customer i, and the proposal contents, and the next time between the person in charge j and the customer i. Search and display appointment appointments. Thereafter, the message determination unit 416 reads customer information, basic information on existing life insurance policies, detailed information on existing life insurance policies, existing life insurance policy information, non-life insurance policy information, non-life insurance policies from the customer area household information database 508. The existing contract information and the like are repeatedly searched and displayed for the customer i and each family constituting the household. After that, the message determination unit 416 searches the message database 514, and, as a pioneering point of view, displays an “appoint of appointment and maintenance” message, a “matter to be heard” message, a “proposal cut” message, etc. Display according to the priority.

  Each step of the method according to the present invention is performed by executing a software program stored in the storage device of the administrator terminal 116 on the CPU of the administrator terminal 116, so that the firewall 104, the web server 106, and the application server 108. , And by controlling the database server 110.

    Each step of the above method according to the present invention may be performed by an operator terminal (not shown) having a CPU and a memory. In an embodiment of the present invention, the software program used to execute the above method may be stored in a program storage device such as a digital data storage medium, and these program storage devices Are machine-readable or computer-readable, and these computers or machines encode program instructions as machine-executable programs or computer-executable programs, and the encoded program instructions are the method of the present invention. Some or all of these steps are executed. The program storage device can be, for example, a digital storage, a magnetic storage medium such as a magnetic disk and magnetic tape, a hard drive, or an optically readable digital data storage medium.

  The present invention may be software and / or a combination of software and hardware, for example, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), It may be implemented using a general purpose computer, or any other hardware equivalent.

  The above description merely provides a disclosure of specific embodiments of the invention and is not intended to limit the invention to only the above embodiments. Accordingly, it is recognized that the present invention is not limited to the embodiments described above, but rather, those skilled in the art can devise alternative embodiments that fall within the scope of the present invention.

100 Insurance Sales Support System 102 Central Management Center 104 Firewall 106 Web Server 108 Application Server 110 Database Server 112 Communication Network 114 Client Terminal

Claims (7)

A sales support system that is connected to at least one client terminal via a communication network and generates a sales strategy for a plurality of product patterns,
For each of a contracted customer who has contracted at least one of the plurality of product patterns and a new customer to whom none of the plurality of product patterns is contracted, a customer attribute, A customer account that stores customer contract status and customer preferences, and each sales representative uniquely associated with each of at least one salesperson assigned an identifier uniquely associated with the device identifier of the client terminal A database that stores a person account that stores the person person activity history of the person, and a message table that stores messages for the sales person,
Based on the customer attributes of the customer account and the customer contract status, a probability model is generated for each of the plurality of product patterns to calculate a predicted contract probability of the product pattern by the existing contract customer and the new customer, and the probability A model generation unit that calculates the predicted contract probability of each product pattern for each of the contracted customer and the new customer using a model;
The priority determined based on the customer preference of the customer account, the customer attribute, and the customer contract status, the person activity history of the person account, and the at least one based on the predicted contract probability A sales support system including a message determination unit that selects a plurality of messages from the message table for each person in charge of sales. In the sales support system according to claim 1,
The sales support system in which the selected message for the at least one sales representative is transmitted to the client terminal associated with the sales representative and displayed on the display unit of the client terminal in the descending order of priority. In the sales support system according to claim 2,
It is recorded in the contact person activity history of the contact person account that the contact person has already executed at least one of the selected messages for the at least one contact person. In this case, the executed message is excluded from the display target, and at least one message having a lower priority is displayed in advance.
Among the messages displayed on the client terminal of the person in charge for a certain period of time, the message executed by the person in charge is reflected so that a record that the person in charge has executed the message is reflected. When the person activity history of the person account is updated, the executed message is deleted and at least one message having a lower priority is displayed in advance in the subsequent time period. Support system. Each of a contracted customer who is connected to at least one client terminal via a communication network and has a contract with at least one of a plurality of product patterns, and a new customer who has not contracted any of the plurality of product patterns A customer account that is uniquely associated with each customer, stores customer attributes, customer contract status, and customer preferences, and each salesperson assigned an identifier that is uniquely associated with the device identifier of the client terminal A database that stores a person-in-charge account that stores the person-in-charge activity history of the salesperson, a message table that stores messages for the salesperson, a model generation unit, and a message determination unit A sales support system that generates a sales strategy for the plurality of product patterns A law,
A probability model in which the model generation unit calculates a predicted contract probability of the product pattern by the existing contract customer and the new customer for each of the plurality of product patterns based on a customer attribute and a customer contract status of the customer account Generating the predicted contract probability of each product pattern for each of the contracted customer and the new customer using the probability model;
Based on the priority determined based on the customer preference of the customer account, the customer attribute, and the customer contract status, the person activity history of the person account, and the predicted contract probability And selecting a plurality of messages from the message table for each of the at least one sales representative. The sales support method according to claim 4,
The sales support method, wherein the selected message for the at least one sales representative is transmitted to the client terminal associated with the sales representative and displayed on the display unit of the client terminal in descending order of priority. The sales support method according to claim 5,
It is recorded in the contact person activity history of the contact person account that the contact person has already executed at least one of the selected messages for the at least one contact person. In this case, the executed message is excluded from the display target, and at least one message having a lower priority is displayed in advance.
Among the messages displayed on the client terminal of the person in charge for a certain period of time, the message executed by the person in charge is reflected so that a record that the person in charge has executed the message is reflected. When the person activity history of the person account is updated, the executed message is deleted and at least one message having a lower priority is displayed in advance in the subsequent time period. Support method. A contracted customer who is executed by a computer and connected to at least one client terminal via a communication network and has a contract with at least one of a plurality of product patterns, and both of the plurality of product patterns are contracted. Each new customer is uniquely associated with each customer and assigned a customer account storing customer attributes, customer contract status, and customer preferences, and an identifier uniquely associated with the device identifier of the client terminal. A database for storing a person account for storing a person person activity history of the person in charge, a message table for storing a message for the person in charge, and a model generation. A sales support system including a message determination unit and a message determination unit. A software program to be executed by a sales support method of generating a business strategy, the method comprising,
A probability model in which the model generation unit calculates a predicted contract probability of the product pattern by the existing contract customer and the new customer for each of the plurality of product patterns based on a customer attribute and a customer contract status of the customer account Generating the predicted contract probability of each product pattern for each of the contracted customer and the new customer using the probability model;
Based on the priority determined based on the customer preference of the customer account, the customer attribute, and the customer contract status, the person activity history of the person account, and the predicted contract probability Selecting a plurality of messages from the message table for each of the at least one sales representative.