KR102226939B1 - Method, apparatus and computer program for matching users using artificial intelligence model with improved effective data extraction performance - Google Patents

Abstract

Provided are a user matching method, device, and computer program using an artificial intelligence model with improved effective data extraction performance. The user matching method using an artificial intelligence model performed by a computing device according to various embodiments of the present invention comprises the steps of: obtaining a matching request between a first user and a second user; extracting effective data for the matching request using a pre-learned artificial intelligence model using pre-stored user information and additional information inputted from the outside as learning data; and matching the first user and the second user based on the extracted effective data.

Description

User matching method, device and computer program using artificial intelligence model with improved effective data extraction performance {METHOD, APPARATUS AND COMPUTER PROGRAM FOR MATCHING USERS USING ARTIFICIAL INTELLIGENCE MODEL WITH IMPROVED EFFECTIVE DATA EXTRACTION PERFORMANCE}

Various embodiments of the present invention relate to a user matching method, apparatus, and computer program using an artificial intelligence model with improved effective data extraction performance.

As wired/wireless information and communication networks are rapidly spreading, a culture of providing and enjoying information and services through information and communication networks is rapidly becoming established, and accordingly, various marketing services (e.g., shopping, promotion of insurance products) through information and communication networks And sales, surveys, telemarketing, etc.) are also being transformed into various forms.

In particular, as customer demands for marketing services increase, the introduction of technologies reflecting various customer services and trends for various marketing is accelerating.

In the conventional Internet, in providing a service that guides advertisements or information related to shopping, insurance product promotion and sales, surveys, telemarketing, etc., technology that considers the characteristics and propensity of users and provides customized services according to them It is constantly being developed. For example, conventional shopping marketing services analyze various information related to customers such as gender, age, income, place of residence, personality, purchase tendency, purchase pattern, and types of cards used to provide products that customers need, It selects products suitable for sale to customers, provides advertisements and information provision services, or provides services that provide a purchase process for the product.

However, for such customer information, a vast amount of information is provided online, but the effectiveness of filtering available data through data mining and marketing application of the filtered data is low. For example, in the telemarketing field, in the case of securing customers through telephone using sales data through conventional data mining, there is a problem that the effectiveness of data on marketing performance actually connected with customers is small.

Korean Patent Publication No. 10-2003-0076509 (2003.09.26)

The problem to be solved by the present invention is an artificial intelligence model learned in consideration of not only various information related to customers, but also additional information including various information related to sales (eg, seller information, sales situation information, information related to sales business, etc.) User matching method using an artificial intelligence model with improved effective data extraction performance that can perform optimized user matching by extracting valid data for matching between customers and sellers using and matching customers and sellers using this, It is to provide devices and computer programs.

Another problem to be solved by the present invention is by retraining the artificial intelligence model using the actual matching result data derived by matching customers and sellers according to the valid data extracted using the artificial intelligence model and the valid data as learning data. It is to provide a user matching method, apparatus, and computer program using an artificial intelligence model with improved effective data extraction performance that enables more accurate extraction of valid data by improving the performance of the intelligent model.

Another problem to be solved by the present invention is to determine the propensity and strength of the seller by considering various information related to the seller (eg, personal information of the seller, personality, sales performance by product, etc.) from the perspective of the seller, and accordingly It is to provide a user matching method, apparatus, and computer program using an artificial intelligence model with improved effective data extraction performance that can increase the seller's success rate of product sales by matching a product suitable for sale by a seller and a customer who will be selling it.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems that are not mentioned will be clearly understood by those skilled in the art from the following description.

In the method of matching a customer and a seller using an artificial intelligence model according to an embodiment of the present invention for solving the above-described problem, in the method performed by a computing device, obtaining a matching request between a first user and a second user , Extracting valid data for the matching request through a pre-learned artificial intelligence model using pre-stored user information and additional information input from the outside as training data, and the first user and the first user based on the extracted valid data It may include the step of matching the second user.

In various embodiments, the user information includes information on a customer as a first user, and the additional information includes at least one of information on a seller as a second user, information on a product, or information related to a sales activity, and the valid The step of extracting data includes at least one of the customer's information, the seller's information, and the product's information as input of the artificial intelligence model, and information on a product to be sold to the customer, and a seller to sell the product. Extracting valid data including at least one of the information of the customer and the information of the customer who is to sell the product, by using the information of the first seller as an input of the artificial intelligence model, a first suitable for sale by the first seller Extracting valid data about product information and information of a first customer who is a target for selling the first product, a second suitable for selling to the second customer by using the information of the second customer as the input of the artificial intelligence model 2 Extracting valid data on product information and information on a second seller who will sell the second product to the second customer, and using the information on the third product as an input of the artificial intelligence model, the third product It may include the step of extracting valid data on the information of the third seller to be sold and the information of the third customer who is to sell the third product.

In various embodiments, the extracting of the valid data comprises a first to perform each of the plurality of consultation procedures when a sales process including a plurality of consultation procedures is to be performed in order to sell a product to the first user. 2 determining a user and extracting valid data related to the determined information of the second user, and matching the first user and the second user comprises: based on the extracted valid data, the And matching each of the determined second users with the first user each time a plurality of counseling procedures are performed.

In various embodiments, the step of extracting the valid data includes a sales procedure for selling the product and a consultation procedure for performing consultation at preset periods for a predetermined period in order to sell the product to the first user. When a sales process is to be performed, extracting valid data on information of a second user to perform the sales procedure and information of a second user to perform the consultation procedure, wherein the first user and the second user 2 The step of matching the user may include performing the sales procedure by matching the first user with a second user who will perform the sales procedure based on the extracted valid data, and the first user through the sales procedure. When the product is purchased, automatically matching a second user to perform the consultation procedure with the first user at each preset period.

In various embodiments, the extracting of the valid data comprises analyzing the consultation data generated by the first user and the second user matched according to the extracted valid data by performing a consultation procedure, and the first user Determining whether a change of the second user matched with is necessary, and when it is determined that the change of the second user is necessary, re-determining a second user to be matched with the first user, and the re-determined second user And extracting valid data related to the information of the second user, wherein the matching of the first user and the second user comprises, in response to extracting valid data regarding information of the re-determined second user, the second The second user matched with the first user may be changed to the re-determined second user and matched again.

In various embodiments, in the extracting of the valid data, when the number of products to be sold to the first user is plural, determining a second user to sell each of the plurality of products to the first user, Determining one or more second users such that the number of second users matched with one user has a minimum value, and extracting valid data on information of the determined one or more second users, wherein the first user and the second user The matching of the second user includes matching the first user and the determined one or more second users based on the extracted valid data, but according to a preset product sales priority and whether a preset condition is satisfied. It may include the step of sequentially matching the determined one or more second users and the first user.

In various embodiments, the matching of the first user and the second user comprises: when the first user and the plurality of second users are matched based on the extracted valid data, the plurality of second users are Matching the first user by grouping into one group, and matching result data generated by matching any one second user included in the one group and the first user to another included in the one group. It may include providing to second users.

In various embodiments, the method may further include retraining the artificial intelligence model by using the extracted valid data and actual matching result data according to the extracted valid data as training data.

A customer and seller matching apparatus using an artificial intelligence model according to another embodiment of the present invention for solving the above-described problem includes a memory for storing one or more instructions and a processor for executing the one or more instructions stored in the memory, , By executing the one or more instructions, the processor may perform a method for matching a customer and a seller using an artificial intelligence model according to an embodiment of the present invention.

A computer program according to another embodiment of the present invention for solving the above-described problems is combined with a computer that is hardware, so that a customer and seller matching method using an artificial intelligence model according to an embodiment of the present invention can be performed. It may be stored in a computer-readable recording medium.

Other specific details of the present invention are included in the detailed description and drawings.

According to various embodiments of the present invention, artificial intelligence learned in consideration of not only various information related to customers, but also additional information including various information related to sales (eg, seller information, sales situation information, information related to sales business, etc.) By using the model to extract valid data for matching the customer and the seller, and using this to match the customer and the seller, optimized user matching can be performed.

In addition, the performance of the artificial intelligence model is improved by retraining the artificial intelligence model using the actual matching result data derived by matching customers and sellers according to the valid data extracted using the artificial intelligence model and the valid data as training data. It enables more accurate extraction of valid data.

In addition, from the seller's point of view, the seller's propensity and strengths are determined by considering various information related to the seller (e.g., personal information of the seller, personality, sales performance by product, etc.). By matching customers who are targets for sale, it is possible to increase a seller's success rate in product sales.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a diagram illustrating a user matching system using an artificial intelligence model according to an embodiment of the present invention.
2 is a hardware configuration diagram of a user matching device using an artificial intelligence model according to another embodiment of the present invention.
3 is a flowchart of a user matching method using an artificial intelligence model according to another embodiment of the present invention.
4 is a diagram illustrating a telemarketing system to which a user matching method using an artificial intelligence model can be applied in various embodiments.
5 is a flowchart of a method of extracting effective data for optimized matching between users in various embodiments.
6 is a diagram illustrating an artificial intelligence model applicable to various embodiments and the form of input/output data of the artificial intelligence model in various embodiments.
7 is a flowchart of a method of extracting valid data by filtering information according to a correlation between user information and additional information, according to various embodiments.
8 is a flowchart of a method of extracting valid data by assigning a weight to information according to a correlation between user information and additional information, according to various embodiments.
9 is a flowchart of a method of determining a learning strength of an artificial intelligence model according to a degree of similarity between a plurality of valid data in various embodiments.
10 is a flowchart of a method of determining a learning strength of an artificial intelligence model according to an actual matching result according to valid data, in various embodiments.

Advantages and features of the present invention, and a method of achieving them will become apparent with reference to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in a variety of different forms. It is provided to fully inform the skilled person of the scope of the present invention, and the present invention is only defined by the scope of the claims.

The terms used in the present specification are for describing exemplary embodiments and are not intended to limit the present invention. In this specification, the singular form also includes the plural form unless specifically stated in the phrase. As used herein, “comprises” and/or “comprising” do not exclude the presence or addition of one or more other elements other than the mentioned elements. Throughout the specification, the same reference numerals refer to the same elements, and "and/or" includes each and all combinations of one or more of the mentioned elements. Although "first", "second", and the like are used to describe various elements, it goes without saying that these elements are not limited by these terms. These terms are only used to distinguish one component from another component. Therefore, it goes without saying that the first component mentioned below may be the second component within the technical idea of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used in the present specification may be used with meanings that can be commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not interpreted ideally or excessively unless explicitly defined specifically.

The term "unit" or "module" used in the specification refers to a hardware component such as software, FPGA or ASIC, and the "unit" or "module" performs certain roles. However, "unit" or "module" is not meant to be limited to software or hardware. The “unit” or “module” may be configured to be in an addressable storage medium, or may be configured to reproduce one or more processors. Thus, as an example, "sub" or "module" refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, Includes procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays and variables. Components and functions provided within "sub" or "modules" may be combined into a smaller number of components and "sub" or "modules" or into additional components and "sub" or "modules". Can be further separated.

Spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe the correlation between a component and other components. Spatially relative terms should be understood as terms including different directions of components during use or operation in addition to the directions shown in the drawings. For example, if a component shown in a drawing is turned over, a component described as "below" or "beneath" of another component will be placed "above" the other component. I can. Accordingly, the exemplary term “below” may include both directions below and above. Components may be oriented in other directions, and thus spatially relative terms may be interpreted according to orientation.

In the present specification, a computer means all kinds of hardware devices including at least one processor, and may be understood as encompassing a software configuration operating in the corresponding hardware device according to embodiments. For example, the computer may be understood as including all of a smartphone, a tablet PC, a desktop, a laptop, and a user client and application running on each device, and is not limited thereto.

In addition, in this specification, the computer applies a method of extracting effective data for optimized matching between users to a telemarketing system (eg, FIG. 4) to extract valid data for matching a telemarketer and a customer, and use the method to extract effective data for matching the telemarketer and the customer. The description is based on the user matching method using the matching artificial intelligence model, but this is only an example for explaining the user matching method using the artificial intelligence model, and is not limited thereto, and specific customers such as door-to-door sales and online sales It can be applied to all fields including technical matters for matching and connecting people to people, as well as product sales fields that sell specific products by connecting them with specific sellers.

In addition, in this specification, the first user refers to a customer who is the target of purchasing a specific product or receiving specific content (eg, advertisements, various events, promotions, surveys, etc.), and the second user refers to the first user. It may mean a seller who sells a specific product or provides a specific content, but this is only an example for explaining a method of matching a customer and a seller for sale, and is not limited thereto.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Each of the steps described herein is described as being performed by a computer, but the subject of each step is not limited thereto, and at least some of the steps may be performed by different devices according to embodiments.

1 is a diagram illustrating a user matching system using an artificial intelligence model according to an embodiment of the present invention.

Referring to FIG. 1, a user matching system using an artificial intelligence model according to an embodiment of the present invention may include a user matching device 100, a user terminal 200, and an external server 300.

Here, the user matching system using the artificial intelligence model shown in FIG. 1 is according to an embodiment, and its components are not limited to the embodiment shown in FIG. 1, and may be added, changed, or deleted as necessary. have.

In one embodiment, the user matching device 100 is for matching a first user (eg, a customer) and a second user (eg, a seller, a counselor, a surveyor, a guide, a telemarketer, an insurance consultant, a communication seller, etc.) Valid data can be extracted, and the first user and the second user can be matched using this. For example, the user matching device 100 uses the user information and additional information (for example, an external factor collected (or input) from the outside) as training data, and uses a pre-learned artificial intelligence model. Valid data for matching between the and the second user may be extracted, and the first user and the second user may be matched using this.

Here, the user information may mean information of a customer who is a first user, and the additional information is sales such as information of a seller who is a second user, information of a product, information related to sales behavior (e.g., business information), and situation information. It may refer to various information related to (eg, sales data), but is not limited thereto.

In various embodiments, the user matching device 100 receives at least one of user information (eg, customer information) and additional information (eg, seller information and product information) as an input of an artificial intelligence model, Valid data including at least one of information on a product to be sold to customers, information on a seller who will sell the product, and information on a customer who is a target for selling the product may be extracted (eg, FIG. 6 ).

Here, the user matching device 100 may extract valid data by using pre-stored user information and additional information input from the outside (eg, information input from a seller or manager) as an input of the artificial intelligence model, but is not limited thereto. Instead, user information may be directly input from the outside or an operation of directly extracting valid data in an environment in which a plurality of user information and a plurality of additional information are previously stored may be performed.

In one embodiment, the user terminal 200 (e.g., a customer's terminal, a seller's terminal, or an administrator's terminal) may be connected to the user matching device 100 through a network, and a user interface ( A user interface (UI) may be provided and output, and various information (eg, user information (customer information) and additional information) may be input through the output UI to provide it to the user matching device 100. For example, the user terminal 200 may receive a UI for performing a Myers-Briggs Type Indicator (MBTI) test from the user matching device 100, and a user (customer) derived by performing an MBTI test through the UI. And MBTI result information of the seller) may be provided to the user matching device 100. However, it is not limited thereto.

In various embodiments, the user terminal 200 may be at least one of a smartphone, a tablet PC, a desktop, and a laptop having a display on at least a portion of the user terminal 200, and from the user matching device 100 through the display. Provided UI can be displayed. However, it is not limited thereto.

In one embodiment, the external server 300 may be connected to the user matching device 100 through a network, and the operation of the user matching device 100 to learn and retrain the artificial intelligence model, effective using the artificial intelligence model. Various information (eg, a plurality of user information, a plurality of additional information, first learning data and actual matching result data, etc.) necessary to perform an operation of extracting data and an operation of matching a user may be stored and provided.

In addition, the external server 300 provides various data (eg, valid data, second learning data, actual matching result data) generated by the user matching device 100 performing an operation for extracting valid data and matching a user. You can receive and store. For example, the external server 300 may be a storage server separately provided outside the user matching device 100. However, it is not limited thereto. Hereinafter, a hardware configuration diagram of a user matching device 100 according to another embodiment of the present invention will be described with reference to FIG. 2.

2 is a hardware configuration diagram of a user matching device using an artificial intelligence model according to another embodiment of the present invention.

Referring to FIG. 2, a user matching device 100 (hereinafter, “computing device 100”) according to another embodiment of the present invention includes one or more processors 110 and a computer program executed by the processor 110 ( A memory 120 for loading the 151, a bus 130, a communication interface 140, and a storage 150 for storing the computer program 151 may be included. Here, only the components related to the embodiment of the present invention are shown in FIG. 2. Accordingly, those of ordinary skill in the art to which the present invention pertains can recognize that other general-purpose components may be further included in addition to the components shown in FIG. 2.

The processor 110 controls the overall operation of each component of the computing device 100. The processor 110 includes a CPU (Central Processing Unit), MPU (Micro Processor Unit), MCU (Micro Controller Unit), GPU (Graphic Processing Unit), or any type of processor well known in the technical field of the present invention. Can be.

In addition, the processor 110 may perform an operation on at least one application or program for executing the method according to embodiments of the present invention, and the computing device 100 may include one or more processors.

In various embodiments, the processor 110 temporarily and/or permanently stores a signal (or data) processed inside the processor 110. -Only Memory, not shown) may be further included. In addition, the processor 110 may be implemented in the form of a system on chip (SoC) including at least one of a graphic processor, RAM, and ROM.

The memory 120 stores various types of data, commands, and/or information. The memory 120 may load the computer program 151 from the storage 150 to execute a method/operation according to various embodiments of the present disclosure. When the computer program 151 is loaded in the memory 120, the processor 110 may perform the method/operation by executing one or more instructions constituting the computer program 151. The memory 120 may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

The bus 130 provides a communication function between components of the computing device 100. The bus 130 may be implemented as various types of buses such as an address bus, a data bus, and a control bus.

The communication interface 140 supports wired/wireless Internet communication of the computing device 100. In addition, the communication interface 140 may support various communication methods other than Internet communication. To this end, the communication interface 140 may be configured to include a communication module well known in the art. In some embodiments, the communication interface 140 may be omitted.

The storage 150 may non-temporarily store the computer program 151. In the case of performing the effective data extraction process and the user matching process through the computing device 100, the storage 150 extracts the valid data and stores various information necessary for matching users by using it (e.g., a plurality of users). Information and a plurality of additional information).

The storage 150 is a nonvolatile memory such as a ROM (Read Only Memory), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), flash memory, etc., a hard disk, a removable disk, or well in the technical field to which the present invention pertains. It may be configured to include any known computer-readable recording medium.

Computer program 151 may include one or more instructions that when loaded into memory 120 cause processor 110 to perform a method/operation in accordance with various embodiments of the present invention. That is, the processor 110 may perform the method/operation according to various embodiments of the present disclosure by executing the one or more instructions.

In one embodiment, the computer program 151 obtains a matching request between a first user and a second user, using pre-stored user information and additional information input from the outside as training data, and using a pre-learned artificial intelligence model. Including one or more instructions for performing a user matching method using an artificial intelligence model including the step of extracting valid data for the matching request and matching the first user and the second user based on the extracted valid data. I can.

The steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or a combination thereof. Software modules include Random Access Memory (RAM), Read Only Memory (ROM), Erasable Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), Flash Memory, hard disk, removable disk, CD-ROM, or It may reside on any type of computer-readable recording medium well known in the art to which the present invention pertains.

Components of the present invention may be implemented as a program (or application) and stored in a medium in order to be combined with a computer that is hardware to be executed. Components of the present invention may be implemented as software programming or software elements, and similarly, embodiments include various algorithms implemented with a combination of data structures, processes, routines or other programming elements, including C, C++. , Java, assembler, or the like may be implemented in a programming or scripting language. Functional aspects can be implemented with an algorithm running on one or more processors. Hereinafter, a user matching method using an artificial intelligence model performed by the computing device 100 will be described with reference to FIGS. 3 and 4.

3 is a flowchart of a user matching method using an artificial intelligence model according to another embodiment of the present invention, and FIG. 4 is a diagram illustrating a telemarketing system to which a user matching method using an artificial intelligence model can be applied in various embodiments. to be.

Referring to FIGS. 3 and 4, in various embodiments, the computing device 100 applies a method of extracting effective data (eg, FIGS. 5 to 10) for optimized matching between users to the telemarketing system. ) And the customer 20 can be optimized.

Here, the telemarketing system can connect the telemarketer 10 (the telemarketer's telephone terminal) and the customer 20 (the customer's telephone terminal) through telephone communication using the telephone network 401 as shown in FIG. 4, By connecting the telemarketer 10 and the customer 20 and the computing device 100 through the data communication network 402, it means a system capable of transmitting and receiving data between the telemarketer 10, the customer 20, and the computing device 100. I can.

In addition, here, with reference to Figs. 3 and 4, a method of performing optimized matching between the telemarketer 10 and the customer 20 that sells products and manages the sold products is disclosed, but this is only an example. However, it is not limited to the telemarketing field, and is applied to the field of matching customers with counselors, surveyors, guides, insurance counselors, telecommunications sellers, etc. who conduct counseling online or offline, providing optimized matching between customers and specific targets. You can also do it.

In addition, here, in the user matching method using the artificial intelligence model performed by the computing device 100, a method of extracting effective data for optimized matching between users is applied to the telemarketing system. Valid data indicating that a specific product is to be sold is extracted, and according to the extracted valid data, the description is based on the fact that it is matched with a specific customer and connected to the customer through the telephone network. However, this is only one example to more easily explain the effective data extraction method for optimized matching between users and the user matching method using an artificial intelligence model according to various embodiments of the present invention, and is limited to the above embodiments. It can be applied to various fields that connect specific customers and specific users (eg, sellers, counselors, surveyors, guides, telemarketers, insurance counselors, telecommunications sellers, etc.), such as online shopping and door-to-door sales. In addition, it is processed (changed, added, and deleted) into a form that can be easily derived from a person with ordinary skill in the technical field to which the present invention belongs, so that a technology that matches a specific person and a specific person, and connects the matched people It can be applied to all fields used (or available).

In step S110, the computing device 100 may obtain a matching request between a customer as a first user and a seller (eg, a telemarketer) as a second user. For example, the computing device 100 may provide a matching service providing UI to the seller's terminal in response to executing a matching service providing application previously installed on the seller's terminal, and suitable for sale through the matching service providing UI. It is possible to obtain a matching request for providing information on a product and a customer who will sell the product, and a matching request for providing information on a customer who is a suitable target for selling the product directly entered by the seller. In addition, the computing device 100 may obtain a matching request requesting a purchase for a specific product from a customer. However, it is not limited thereto.

Here, from the seller's point of view, the matching request is to determine the target customer to sell the product that the seller wants to sell (for example, a product directly entered from the seller) and request matching, or the seller's disposition and strength. The analysis may be performed to determine an appropriate product for the seller to sell, and to determine a customer who will sell the determined product to request matching, but is not limited thereto.

In addition, from the customer's point of view, the matching request may be a request for a product to be purchased by selecting a product that the customer wants to purchase or to provide information on a recommended product based on their personal information. , Is not limited thereto.

In step S120, the computing device 100 may extract valid data according to the matching request by using the pre-learned artificial intelligence model using pre-stored user information and additional information input from the outside as training data. Hereinafter, a method of extracting effective data for optimized matching between users performed by a computing device will be described with reference to FIGS. 5 to 10.

5 is a flowchart of a method of extracting effective data for optimized matching between users in various embodiments.

Referring to FIG. 5, in step S210, the computing device 100 includes pre-stored user information (eg, information of a customer who is a first user) and additional information input from the outside (eg, information of a seller who is a second user, and a product). The artificial intelligence model may be trained using the first training data) and information related to sales behavior.

Here, the user information may mean all information related to the customer. For example, customer information includes the customer's name, age, gender, address, occupation, bank in use, card type, points, disposition and personality (e.g., MBTI analysis results), interests, marital status and presence of children, and voice. It can include all information related to the customer, such as tone and propensity. However, the type of customer information is not limited thereto, and all fields and various types of information that can be used to grasp the customer's disposition can be applied.

In addition, here, the additional information may mean all information related to the seller and sales behavior, such as seller information, situation information, business information, and artificial intelligence model result strategy information. For example, sales information includes the seller's name, age, gender, address, disposition and personality (e.g., MBTI analysis results), interests, marital status and presence of children, voice tone and disposition, seller’s sales performance by product, customer response disposition. , Consultation time (total consultation time, consultation time by case, average consultation time), voice tone and propensity, effective success rate, customer response pattern, call pattern with customer, matching success pattern, weather by consultation, temperature and time, social and economic And cultural context information. However, the present invention is not limited thereto, and the additional information is applicable to all fields and various types of information as well as sales data related to sales.

In various embodiments, the computing device 100 includes customer information, seller information, and information related to a sales activity generated by matching the customer and the seller (eg, matching result information, consultation content information, etc.) generated by a sales activity. Labeling the result information (e.g., information on the success of product sales) derived from matching the customer and the seller to the first learning data including various information), and using a supervised learning method. The first training data labeled with the result may be trained in the artificial intelligence model. However, this is only an example of training an artificial intelligence model, and various techniques for training an artificial intelligence model can be applied.

In various embodiments, the computing device 100 collects a plurality of pre-stored customer information and a plurality of additional information related to sales from the outside (for example, the external server 300), or collects a plurality of customer information and a plurality of sales related information from a user. The artificial intelligence model may be trained by receiving additional information of, but is not limited thereto, and the computing device 100 loads the artificial intelligence model into an environment in which a plurality of customer information and a plurality of additional information related to sales are previously stored. The learning behavior of the artificial intelligence model can be driven so that the artificial intelligence model can be trained in the environment.

In step S220, the computing device 100 may extract valid data for matching between the customer as the first user and the seller as the second user by using the artificial intelligence model learned through the step S210.

Here, the valid data is result data extracted by inputting at least one of additional information including customer information, seller information, and product information as an input of the artificial intelligence model, and includes information on products and products to be sold to customers. It may include at least one of information on a seller to be sold and information on a customer to whom the product is to be sold, but is not limited thereto.

For example, the computing device 100 may use customer information as an input of an artificial intelligence model to extract information on a product to be sold to a corresponding customer and valid data on information on a seller who will sell the product. Through this, the computing device 100 can improve the success rate of product sales to a specific customer by determining a product suitable for sale to a specific customer and an optimal seller for selling the product to a customer (eg, a seller with a high sales success rate). have. However, it is not limited thereto.

In addition, the computing device 100 may use product information as an input of an artificial intelligence model to extract valid data regarding information on a seller who will sell a product and information on a customer who is a target for selling the product. Through this, the computing device 100 may improve a sales volume and a sales success rate for a specific product by determining an optimal customer for selling a specific product and an optimal seller for selling a specific product to the customer. However, it is not limited thereto.

In addition, the computing device 100 may extract information on a product suitable for sale by the seller and valid data on information on a customer to whom the product is to be sold by using the seller's information as an input of the artificial intelligence model. Through this, the computing device 100 identifies the propensity and strengths of a specific seller, and determines a product that a specific seller can sell well and a customer who will sell it, so that a specific seller can use his or her strengths to sell the product. You can induce to be able to. However, it is not limited thereto.

In addition, the computing device 100 may extract valid data regarding information on a product suitable for the seller to sell to the customer by using the customer information and the seller information as inputs of the artificial intelligence model. Through this, the computing device 100 may determine an optimal product to be sold between a specific customer and a specific seller, thereby improving the sales volume and sales success rate of the product. However, it is not limited thereto.

In addition, the computing device 100 may extract valid data on information of a seller who will sell a product to a customer by using customer information and product information as inputs of the artificial intelligence model. Through this, the computing device 100 may determine an optimal seller to sell a specific product to a specific customer, thereby improving a sales volume and a sales success rate of a specific product to a specific customer. However, it is not limited thereto.

In addition, the computing device 100 may extract valid data on customer information suitable for the seller to sell the product by using the information of the seller and the product as inputs of the artificial intelligence model. Through this, the computing device 100 may improve a sales success rate of a specific product by a specific seller by determining an optimal customer for selling a specific product. However, it is not limited thereto.

In addition, the computing device 100 may extract valid data on a sales success rate when the seller sells the product to the customer by using customer information, seller information, and product information as inputs of the artificial intelligence model. Through this, the computing device 100 predicts the likelihood that a specific seller will sell a specific product to a specific customer, thereby re-verifying the valid data so that the artificial intelligence model can extract more accurate result data, and thereby improve the performance of the artificial intelligence model. Can be evaluated. However, it is not limited thereto.

In various embodiments, the computing device 100 may determine additional information necessary to extract valid data based on the type of product to be sold and information on the customer who is the target of selling the product. In this case, the computing device 100 provides a UI to the user terminal 200 to extract valid data when the necessary additional information is omitted (eg, when it is not collected from the outside or input from the user, etc.). It can guide you to enter the necessary additional information.

In various embodiments, the computing device 100 uses at least one of additional information including customer information, seller information, and product information as an input of the artificial intelligence model to provide information on products and products to be sold to customers. Extract valid data including at least one of the information of the seller to be sold and the information of the customer to whom the product is to be sold, but guide information that can increase the sales success rate in the process of the seller selling the product to the customer (e.g.: Valid data indicating guidance phrases (e.g. phrases that increase the success rate, phrases that should be avoided), voice tone, intonation, product guidance method, etc.) can be extracted together.

In various embodiments, the computing device 100 may analyze a plurality of valid data extracted from an artificial intelligence model to calculate a correlation score of each of the plurality of customer information and the plurality of sales information, and the calculated correlation score may be calculated by artificial intelligence. By reflecting in the model, the performance of the effective data extraction operation can be improved. Hereinafter, it will be described with reference to FIGS. 7 and 8.

7 is a flowchart of a method of extracting valid data by filtering information according to a correlation between user information and additional information, according to various embodiments.

Referring to FIG. 7, in step S310, the computing device 100 may analyze a plurality of valid data extracted from the artificial intelligence model.

In various embodiments, the computing device 100 analyzes a plurality of input data including customer information and additional information and a plurality of valid data extracted according to the plurality of input data using a data analysis model, Correlation between additional information can be derived. For example, the computing device 100 may include first valid data extracted from input data including first customer information and first additional information, and second data extracted from input data including second customer information and first additional information. 2 By analyzing the valid data, it is possible to derive a correlation between the first customer information and the second customer information and the first additional information.

Here, the correlation between the plurality of customer information and the additional information may mean information on whether or not the customer information and the additional information mutually influence and the degree of mutual influence. However, it is not limited thereto.

That is, the computing device 100 generates a plurality of input data by changing some of the plurality of customer information and additional information included in the input data, and compares a plurality of valid data derived according to the generated plurality of input data. In addition, it is possible to determine whether there is a mutual influence between customer information and additional information.

In addition, the computing device 100 generates a plurality of input data by changing customer information and additional information that affect each other by varying the degree of change, and compares a plurality of valid data derived according to the generated plurality of input data. By doing so, it is possible to determine the degree of mutual influence between the customer information and the additional information. However, the present invention is not limited thereto, and various techniques for determining the correlation and correlation between a plurality of pieces of information may be applied.

In step S320, the computing device 100 may calculate a correlation score for each of the plurality of customer information and the additional information according to the correlation (or relevance) between the plurality of customer information and the additional information derived in step S310.

In this case, the computing device 100 determines the correlation between the plurality of customer information and the additional information as a preset level (eg, a first grade indicating high relevance, a second grade indicating intermediate relevance, and a third grade indicating low relevance). It can be classified, and a correlation score can be calculated according to the classified grade. For example, when the correlation between the additional information and the plurality of customer information is a third grade indicating low relevance, the computing device 100 may calculate a correlation score having a first value, and may calculate a second score indicating the middle of the relevance. In the case of a grade, a correlation score having a second value greater than the first value may be calculated, and in the case of a first grade indicating high relevance, a correlation score having a third value greater than the second value may be calculated.

In this case, various methods known in advance may be applied to the configuration for calculating the correlation score, but a method of setting a preset grade according to the correlation and assigning a preset score according to the set grade, or specific information according to a preset criterion It is preferable to calculate the correlation score according to any one of the methods of scoring the correlation between them and calculating the scoring result as a correlation score.

Here, the first value, the second value, and the third value may be values set by the administrator before the computing device 100 performs an operation of calculating a correlation score for each of the plurality of customer information and additional information, but is limited thereto. It doesn't work.

In step S330, the computing device 100 may filter only any one of the plurality of customer information based on the correlation score calculated in step S320. For example, the computing device 100 is a score for which a correlation score with additional information is set in advance (for example, a score value set in advance by an administrator or a score value set by itself through learning to maintain the reliability of an artificial intelligence model) Only one or more of the above customer information may be selected, and valid data may be extracted by using only the selected one or more customer information as an input of the artificial intelligence model.

In addition, the computing device 100 may filter only one of customer information among a plurality of additional information based on the correlation score calculated through step S320. For example, the computing device 100 is a score for which a correlation score with customer information is set in advance (for example, a score value set in advance by an administrator or a score value set by itself through learning to maintain the reliability of an artificial intelligence model) Only one or more of the above additional information may be selected, and valid data may be extracted by using only the selected one or more additional information as an input of the artificial intelligence model.

That is, in the process of extracting valid data, the computing device 100 selectively considers only customer information having high correlation with additional information, or selectively considers only additional information having high correlation with customer information, and extracts valid data. There is an advantage in that valid data can be extracted more quickly by filtering unnecessary information while maintaining the reliability of valid data extracted from an intelligent model.

In addition, the computing device 100 may extract valid data using an artificial intelligence model and continuously analyze previously stored valid data and new valid data to continuously update the correlation score between customer information and additional information, Accordingly, there is an advantage in that it is possible to extract effective data with higher reliability by using an artificial intelligence model by improving performance for an operation of filtering various types of information.

8 is a flowchart of a method of extracting valid data by assigning a weight to information according to a correlation between user information and additional information, according to various embodiments.

Referring to FIG. 8, in step S410, the computing device 100 may analyze a plurality of valid data extracted from the artificial intelligence model. Here, the plurality of valid data analysis operations performed by the computing device 100 in step S410 may be performed in the same or similar form as the plurality of valid data analysis operations performed by the computing device 100 in step S310 of FIG. 7. However, it is not limited thereto.

In step S420, the computing device 100 may calculate a correlation score of each of the plurality of customer information and the additional information according to the correlation (or relevance) between the plurality of customer information and the additional information derived in step S410. Here, the correlation score calculation operation performed by the computing device 100 in step S420 may be performed in the same or similar form as the correlation score calculation operation performed by the computing device 100 in step S320 of FIG. 7. Not limited.

In step S430, the computing device 100 may assign a weight to each of the plurality of customer information based on the correlation score calculated through the step S420, and add additional information and a plurality of weighted customer information to the artificial intelligence model. Valid data can be extracted as input. For example, the computing device 100 may assign a weight to each of a plurality of customer information in proportion to a score of association with additional information. In this case, the computing device 100 may set customer information having a relatively high correlation score to have a higher weight value than customer information having a relatively low correlation score.

In addition, the computing device 100 may assign a weight to each of the plurality of additional information based on the correlation score calculated through step S420, and use the customer information and the plurality of weighted additional information as an input of the artificial intelligence model. Thus, valid data can be extracted. For example, the computing device 100 may assign a weight to each of a plurality of additional information in proportion to a score of association with customer information. In this case, the computing device 100 may set the additional information having a relatively high correlation score to have a higher weight value than the additional information having a relatively low correlation score.

In this case, the weight value may be a value converted at a predetermined ratio in proportion to the association score, or a weight value corresponding to the association score based on the weight value data for each association score previously set and stored by the administrator, but is not limited thereto, Using the artificial intelligence model, an optimal weight value that can improve the performance of the artificial intelligence model can be automatically set.

That is, in the process of extracting valid data, the computing device 100 gives high weight to additional information including a plurality of sales information and highly relevant customer information to focus on highly relevant customer information, or There is an advantage in that it is possible to extract more accurate and reliable valid data by giving high weight to additional information that is highly relevant to customer information, and extracting valid data by focusing on additional information that is highly relevant.

In addition, the computing device 100 may extract valid data using an artificial intelligence model and continuously analyze previously stored valid data and new valid data to continuously update the correlation score between customer information and additional information, Accordingly, by updating the weights assigned to each information in real time to improve the effective data extraction performance of the artificial intelligence model, there is an advantage that more reliable valid data can be extracted using the artificial intelligence model.

Referring back to FIG. 5, in step S230, the computing device 100 uses at least one of the valid data extracted through the step S120 and the actual matching result data according to the extracted valid data as second training data to generate an artificial intelligence model. Can be retrained.

Here, the actual matching result data is a variety of data and information generated by matching customers and sellers based on valid data through a customer and seller matching service provided from the computing device 100 or an external matching server, and matching customers and sellers. (For example, it may mean information generated by a consultation procedure performed by a customer and a seller, such as information on consultation time, consultation content, number of consultations, progress status, and progress result (sales result)). However, the actual matching result data is not limited to data and information generated by matching customers and sellers, and customers and counselors, customers and surveyors, customers and guides, customers and telemarketers, customers and insurance agents, and customers and telecommunications sellers. It may include various types of data and information generated by performing matching between various users, such as.

In various embodiments, the computing device 100 filters a plurality of valid data extracted from the artificial intelligence model according to a preset rule, so that at least one valid data among the plurality of valid data and at least one valid data are actually matched. The result data may be determined, and the artificial intelligence model may be retrained using only the determined at least one valid data and actual matching result data according to the at least one valid data as second training data.

In general, when valid data is extracted using an artificial intelligence model, it is not a problem of the performance of the artificial intelligence model, but a variety of external factors such as operation errors of the artificial intelligence model, omissions and errors of input data, and communication problems. Therefore, valid data with low reliability may be extracted.

When an artificial intelligence model is trained using valid data with low reliability as training data, the performance of the artificial intelligence model previously trained using the previously highly reliable training data can be degraded, and accurate valid data is extracted accordingly. The problem of not being able to do it can arise.

Therefore, the computing device 100 does not retrain the artificial intelligence model by using all valid data extracted from the artificial intelligence model and the actual matching result data according to all valid data as training data, and does not retrain the artificial intelligence model. Valid data, at least part of the data damaged due to an error in the valid data extraction operation, valid data extracted while accurate information could not be collected due to communication problems, and valid data with low reliability due to less input information, etc. In the learning process for the data extraction operation, in order to filter the data that may cause obstacles, the valid data with an abnormality is filtered using a preset rule, which is a pre-set rule, and selectively relearn only valid data with no abnormality. There is an advantage in that the above problems can be solved and the reliability of the artificial intelligence model can be improved.

In various embodiments, the computing device 100 generates input data for matching an arbitrary customer and an arbitrary seller based on at least one of customer information and additional information, and uses the generated input data in an artificial intelligence model. Valid data can be extracted as input, and an artificial intelligence model can be trained by using the extracted valid data as second training data.

For example, even if there is no separate control command for extracting valid data (eg, a user input instructing to extract valid data for matching between a customer and a seller), the computing device 100 Every period set in advance by) extracting first valid data for matching between any first customer and any first seller, and second valid data for matching between any second customer and any second seller, Using this as the second training data, the artificial intelligence model can be retrained.

At this time, the computing device 100 delays or interferes with the second operation of extracting valid data for matching between the actual customer and the seller by the first operation of extracting valid data for matching an arbitrary customer and an arbitrary seller. To prevent this, the first operation may be performed only when the second operation is not performed or the number of times the second operation to be performed is less than or equal to a preset value. However, it is not limited thereto.

That is, the computing device 100 automatically generates training data for training of the artificial intelligence model at regular intervals even if there is no separate instruction, and retrains the artificial intelligence model using this automatically, thereby continuously improving the performance of the artificial intelligence model. There is an advantage that it can be improved.

In various embodiments, the computing device 100 may extract a plurality of valid data by repeatedly inputting one input data for matching a customer and a seller into an artificial intelligence model, and convert the plurality of valid data as second training data. Thus, the artificial intelligence model can be retrained.

That is, the computing device 100 may increase the reliability of the artificial intelligence model by increasing the number of times of learning the artificial intelligence model by generating a plurality of training data for learning the artificial intelligence model with only one input data.

In various embodiments, the computing device 100 may analyze valid data extracted from the artificial intelligence model, and determine the intensity of training and retraining of the artificial intelligence model according to the analysis result of the effective data. Hereinafter, it will be described with reference to FIGS. 9 and 10.

9 is a flowchart of a method of determining a learning strength of an artificial intelligence model according to a degree of similarity between a plurality of valid data in various embodiments.

Referring to FIG. 9, in step S510, the computing device 100 may extract valid data by inputting one first input data artificial intelligence model for matching between a customer as a first user and a seller as a second user. have. Here, the effective data extraction operation performed by the computing device 100 in step S510 may be performed in the same or similar form as the effective data extraction operation performed by the computing device 100 in step S220 of FIG. 5. Not limited.

In step S520, the computing device 100 performs an operation (for example, step S510) of extracting valid data by using one first input data as an input of an artificial intelligence model, and the number of times N is a reference value (for example, the manager It can be determined whether the value is higher than or equal to a preset value).

In step S530, when it is determined that the number of times (N) that the operation of extracting valid data is performed is less than the reference value, the computing device 100 extracts valid data by using one first input data as an input of the artificial intelligence model. Counting (N = N + 1) the number of times the operation was performed (N = N + 1), and retrying the operation of extracting valid data by using one first input data as the input of the artificial intelligence model (e.g., step S510). Can be done.

That is, the computing device 100 repeatedly performs an operation of extracting valid data until the number of times N of performing the operation of extracting valid data (for example, step S510) is equal to or greater than the reference value, A plurality of valid data can be extracted by using the input data.

In step S540, when the number of times (N) of performing the operation of extracting valid data is equal to or greater than the reference value, the computing device 100 determines the similarity between the plurality of extracted valid data by repeatedly performing steps S510 to S530. Can be analyzed.

Here, the degree of similarity between the plurality of valid data means a measure indicating how much of the same or similar information is included in the plurality of valid data by individually comparing a plurality of detailed information included in each of the plurality of valid data. I can. For example, in various embodiments, the computing device 100 includes detailed information including the same or similar content relative to the total number of detailed information included in each of the first valid data and the second valid data. A ratio of is calculated, and a degree of similarity between the first valid data and the second valid data may be calculated using the calculated ratio. However, the present invention is not limited thereto, and various techniques for calculating and determining the degree of similarity between a plurality of data may be applied.

In step S550, the computing device 100 may determine the learning strength for the artificial intelligence model according to the similarity between the plurality of valid data calculated in step S440.

Here, the learning intensity may mean a measure indicating a degree to which retraining of the artificial intelligence model is to be performed. For example, the learning intensity is set to a certain extent for retraining the second training data used for retraining of the artificial intelligence model, how short or long the retraining period of the artificial intelligence model is to be performed, and the artificial intelligence model. It may include information on how many times the number of times of relearning is to be performed, but is not limited thereto.

In various embodiments, the computing device 100 may calculate a valid data reliability score for an artificial intelligence model according to a similarity between a plurality of valid data.

Here, the valid data reliability score may mean a measure indicating a degree of reliability of valid data extracted from the artificial intelligence model. For example, the computing device 100 selects an effective data reliability score value corresponding to the calculated similarity based on the valid data reliability score value data for each similarity previously set and stored by the administrator, thereby selecting the effective data for the artificial intelligence model. Reliability scores can be calculated.

In this case, the valid data reliability score value data may be set so that the case where the similarity between the plurality of valid data is high has a higher valid data reliability score than the case where the similarity is low, but is not limited thereto.

In addition, the computing device 100 may determine at least one of the range of the second training data, the retraining period, and the number of times of the artificial intelligence model according to the valid data reliability score. For example, the computing device 100 may set a case having a high valid data reliability score to have a lower learning intensity than a case having a low valid data reliability score. However, it is not limited thereto.

Here, the detailed setting value of the learning intensity may be set based on the learning intensity data for each valid data reliability score previously set and stored by the administrator, but is not limited thereto, and the performance of the artificial intelligence model is determined by using the artificial intelligence model. An optimal setting value that can be improved can be automatically set.

In general, an artificial intelligence model with good effective data extraction performance extracts the same result data when the same input data is input. Therefore, when multiple result data are extracted by repeatedly inputting the same input data into the artificial intelligence model, the case where the same result data is repeatedly extracted or the result data with high similarity is extracted is compared to the case of extracting different result data. It can be judged to have high reliability.

That is, when the same input data is repeatedly input to the artificial intelligence model, the computing device 100 determines the learning strength of the artificial intelligence model so that the same result data can be repeatedly extracted or result data having high similarity can be extracted. By doing so, there is an advantage in that an artificial intelligence model can be trained so that more reliable results can be extracted.

10 is a flowchart of a method of determining a learning strength of an artificial intelligence model according to an actual matching result according to valid data, in various embodiments.

Referring to FIG. 10, in step S610, the computing device 100 may collect a plurality of valid data extracted from the artificial intelligence model and a plurality of actual matching result data according to each of the plurality of valid data. For example, actual matching result data generated by matching a customer and a seller based on valid data through a customer and seller matching service provided from the computing device 100 or an external matching server Data including information) can be collected. However, it is not limited thereto.

In step S620, the computing device 100 may calculate a matching success rate for the artificial intelligence model based on the plurality of valid data and the plurality of actual matching result data according to each of the plurality of valid data collected through the step S610. . For example, the computing device 100 calculates a matching success rate indicating the ratio of the valid data successfully sold to the total number of valid data by using the success of each of the plurality of valid data (eg, whether or not a product is sold) can do. However, it is not limited thereto.

In step S630, the computing device 100 may calculate a valid data reliability score for the artificial intelligence model based on the matching success rate calculated in step S620. Here, the calculation of the valid data reliability score performed by the computing device 100 is performed in the same or similar form as the calculation of the valid data reliability score performed by the computing device 100 in step S550 of FIG. 9 (e.g., matching The case where the success rate is high may yield a higher valid data reliability score than the case where the matching success rate is low, but is not limited thereto.

In step S640, the computing device 100 may determine the learning strength for the artificial intelligence model based on the valid data reliability score calculated in step S630. Here, the operation of determining the learning strength of the artificial intelligence model performed by the computing device 100 may be performed in the same or similar form as the operation of determining the learning strength performed in step S550 of FIG. 9, but is not limited thereto.

Accordingly, the computing device 100 provides valid data extracted using an artificial intelligence model for matching between customers and sellers, and obtains actual matching result data for customers and sellers as feedback on the provided valid data. In other words, by using this to determine the learning strength of the artificial intelligence model, that is, by determining the learning intensity of the artificial intelligence model by reflecting the actual result, there is an advantage that the artificial intelligence model can be trained to extract more reliable results.

Referring back to FIG. 3, in various embodiments, the computing device 100 may include user information including customer information, seller information, product information, and additional information including information related to product sales behavior, Valid data on information of one or more sellers who will sell products to customers can be extracted.

In various embodiments, in order to sell a product to a customer, the computing device 100 determines a seller to perform each of the plurality of consultation procedures when a sales process including a plurality of consultation procedures is to be performed, and Valid data about information can be extracted.

For example, when the computing device 100 needs to perform two consultation procedures (e.g., pre-consultation procedure, this consultation procedure) according to the type of product to be sold to customers, the information of the seller who will perform the pre-consultation procedure and Valid data on the information of the seller who will perform this consultation procedure can be extracted.

In this case, when the sellers who will perform each of the two consultation procedures are different from each other, the computing device 100 may respectively extract valid data regarding information of different sellers and information of a consultation procedure to be performed by the corresponding seller.

On the other hand, when the seller who will perform each of the two consultation procedures is the same seller, that is, when all of the plurality of consultation procedures can be performed by only one seller, the computing device 100 One valid data about the seller's information can be extracted. However, it is not limited thereto.

In various embodiments, in order to sell a product to a customer, the computing device 100 needs to perform a sales process including a sales procedure for selling products and a counseling procedure for performing counseling at preset periods for a predetermined period of time. , It is possible to extract valid data on the information of the seller who will perform the sales procedure and the information of the seller who will perform the consultation procedure.

For example, when the product to be sold to a customer is a rental product that is rented for a certain period, the computing device 100 is a seller (or manager) who continuously consults with a seller who sells the rental product. Can be determined, and valid data for this can be extracted.

At this time, the computing device 100 divides a predetermined period for performing a continuous consultation procedure into a plurality of unit periods, and determines an optimized seller for each period based on the attributes of each of the divided plurality of unit periods, and , It is possible to extract valid data on the information of the seller to be matched for each period.

For example, the computing device 100 includes a first period for explaining and guiding a product purchased for a predetermined period, a second period for performing maintenance and management of the product, and a third period for inducing a renewal of a rental product. It is possible to divide into periods, and determine a seller who will perform counseling in each of the divided first period, second period, and third period. As an example, the computing device 100 matches a first seller who shows strengths in detailed and friendly descriptions of products during the first period, and shows strengths in meticulously performing maintenance and management of products during the second period. 2 It is possible to extract valid data for matching sellers and matching third sellers with a high success rate of renewal during the third period. However, it is not limited thereto.

In various embodiments, the computing device 100 matches a customer and a seller according to a matching operation (for example, step S630) of a customer and a seller to be described later, based on valid data extracted from the artificial intelligence model, and The consultation data generated as the seller performs the consultation procedure is collected and analyzed, and it is possible to determine whether a change of the seller matching the customer is necessary according to the analysis result of the consultation data.

Thereafter, when it is determined that the change of the seller matched with the customer is necessary according to the determination result of whether the change of the seller is necessary, the computing device 100 re-determines the seller to be matched with the customer, and Valid data can be extracted. In this case, the computing device 100 may re-determine a seller to be newly matched to the customer by using the analysis result of the consultation data as well as the user information and additional information as an input of the artificial intelligence model.

For example, the computing device 100 analyzes the consultation data generated as the customer and the first seller perform the consultation procedure, and uses the customer's voice tone, intonation, and the contents of the customer's remarks to the first seller. The preference (or likelihood) of can be evaluated, and when it is determined that the preference of the customer is less than the reference value, it can be determined that the change of the seller matched with the customer is necessary.

Here, the customer's preference for the seller is calculated according to the customer's voice tone, intonation, counseling time, and remark content according to the previously stored preference, and the attribute of each of the plurality of keywords included in the remark is positive or negative ( Negative) may be a value obtained by scoring the content of speech and correcting it by reflecting the calculated result value, but is not limited thereto.

Thereafter, the computing device 100 may re-determine a seller (eg, a second seller) to be matched with the customer based on the customer information and the additional information, and extract valid data regarding the re-determined second seller information. In this case, the computing device 100 may determine a seller who is predicted to have high customer preference among the plurality of sellers as a seller to perform the second consultation procedure, but is not limited thereto.

In various embodiments, when the number of products to be sold to the customer is plural, the computing device 100 determines a seller to sell each of the plurality of products to the customer, but at least one such that the number of sellers matching the customer has a minimum value. It is possible to determine a seller and extract valid data regarding information of one or more determined sellers.

For example, when the product to be sold to a customer is a first product, a second product, and a third product, and the number of products is three, the computing device 100 may provide the customer with the first product, the second product, and the third product. One seller who is evaluated as suitable for sale (eg, a success rate of selling three products according to an artificial intelligence model is equal to or greater than a predetermined value) may be determined, and valid data regarding the determined seller's information may be extracted.

At this time, the computing device 100 sells three products in the absence of one seller evaluated as suitable for selling the first product, the second product, and the third product to the customer, or by matching with one seller and the customer. When it is predicted that the expected success rate in the case of proceeding with is less than a predetermined value, two sellers who are suitable to sell the first product, the second product and the third product to the customer are determined (e.g., the first product and the second product). Determining a seller suitable for selling the product and a seller suitable for selling a third product, or determining a seller suitable for selling the first product and a seller suitable for selling the second product and the third product), and Effective data related to information can be extracted.

In addition, when the computing device 100 does not have two sellers suitable for selling the first product, the second product, and the third product to the customer, or when it matches the two sellers and the customer to sell three products, If the expected success rate of is predicted to be less than a predetermined value, three sellers who are suitable to sell the first product, the second product, and the third product to the customer are determined (e.g., a seller suitable for selling the first product, A seller suitable for selling the second product and a seller suitable for selling the third product are individually determined for each product), and valid data related to information of the determined three sellers may be extracted.

In general, when a customer wants to purchase multiple products, if many sellers are matched to sell each of the products, the seller changes every time the product purchase procedure is performed, resulting in confusion for the customer or the seller. As is changed, there is a problem that you may experience discomfort such as having to re-consult the contents you consulted.

In addition, in order to solve this inconvenience, if you want to sell multiple products to customers by simply matching only one seller with a customer, one seller can successfully sell a specific product showing strength, but each There is a problem in that the possibility of selling is lowered when considering the sale of all products, because the expertise is poor in preparation for matching a professional seller for each product.

In particular, if there is no relationship between products to be sold to customers (e.g., electronic products and insurance products) when simply matching only one seller to a customer, it may be impossible to provide accurate information about the product or to sell itself. there is a problem.

That is, when the number of products to be sold to the customer is plural, the computing device 100 determines the number of sellers who will sell the plurality of products to the customer by sequentially increasing the number of sellers from one person to the minimum number of sellers matching the customer. By allowing customers to match with multiple sellers, confusion or discomfort can be minimized. In addition, by sequentially increasing the number of matching sellers so that customers and optimized sellers can be matched, various problems that may occur when only one seller is matched for the convenience of customers as described above can be supplemented. I can.

In step S130, the computing device 100 may match the customer and the seller based on the valid data extracted in step S120.

Here, matching the customer and the seller is the terminal of the telemarketer 10 and the terminal of the customer 20 through the telephone network 401 in response to obtaining a predetermined input from the telemarketer 10 (eg, an input indicating a connection request). Or automatically connect the terminal of the telemarketer 10 and the terminal of the customer 20 through the telephone network 401, or the terminal of the telemarketer 10 through the data communication network 402 and sell information about the customer and to the customer By providing information on a product to be done, it may mean providing information so that the telemarketer 10 can directly check the information and directly connect with the customer 20, but is not limited thereto.

In various embodiments, the computing device 100 refers to each of the determined sellers each time a plurality of consultation procedures is performed, based on valid data on information of a seller (second user) who will perform each of the plurality of consultation procedures. It can match (first user).

At this time, the computing device 100 recognizes the start and end points of the counseling procedure between the customer and the specific seller (eg, recognizes the start of the call, the call is in progress, the end of the call), and when each counseling procedure ends, the order of the counseling procedure is performed. Accordingly, the seller who will perform the next consultation procedure can be automatically matched with the customer.

In various embodiments, after selling the product to the customer, the computing device 100 provides valid data extracted in response to the need to continuously perform the consultation procedure at preset periods for a predetermined period of time (for example, to perform a sales procedure). Valid data on the seller's information and the seller's information to perform the consultation procedure) Initially, the customer and the first seller are matched to perform the sales procedure for the product, and according to the sales procedure, the product is provided in response to the customer's purchase of the product. 2 By setting the seller as a designated consultant for the customer, the customer and the second seller can be automatically matched at each preset period for a predetermined period of time.

In various embodiments, the computing device 100 matches the customer and the first seller based on the valid data extracted in step S620, but in the effective data for changing the seller matched with the customer from the first seller to the second seller. Accordingly, the seller matching the customer can be changed from the first seller to the second seller and matched.

At this time, the computing device 100 identifies the time point at which the matching between the customer and the first seller is terminated (eg, the point at which the first consultation procedure performed between the customer and the first seller is terminated), and the first consultation procedure is terminated. After that, at the time of starting the second consultation procedure, the customer and the second seller may be automatically matched.

In various embodiments, the computing device 100 may sequentially match the customer and the seller based on valid data on information about the seller for selling a plurality of products to the customer.

At this time, when the number of sellers matching the customer is plural, the computing device 100 satisfies a preset priority (e.g., sales priority for each product, order of high sales success rate, order of low sales success rate, etc.) and preset conditions. A plurality of sellers may be sequentially matched with customers depending on whether or not (eg, whether the sales and consultation procedure for a specific product has ended).

In various embodiments, when a customer and a plurality of sellers are matched based on valid data, the computing device 100 may group the plurality of sellers into one group to match the customer. At this time, the computing device 100 automatically matches the first seller belonging to one group and the customer to match the consultation data generated by performing the consultation procedure to other second sellers and third sellers belonging to one group. Can provide.

This enables other sellers to become familiar with the contents of consultation between a customer and a specific seller, thereby preventing the customer from experiencing discomfort by repeatedly asking the same question or requesting the same to the customer.

In step S140, the computing device 100 may collect the actual matching result data generated by matching the customer and the seller through step S130, and using this, according to the valid data extracted from the artificial intelligence model. Matching success rate can be calculated. Here, the matching success rate calculation operation performed by the computing device 100 may be performed in the same or similar form as the matching success rate calculation operation performed by the computing device 100 in step S620 of FIG. 10, but is not limited thereto. .

Thereafter, the computing device 100 includes valid data extracted through step S120, various data and information generated by matching a customer and a seller through step S130 (eg, consultation data, product sales, etc.), and calculated in step S140. The AI model can be retrained using the success rate according to the matching result as training data. Here, the artificial intelligence model re-learning operation may be performed in the same or similar form as the re-learning operation of the artificial intelligence model performed by the computing device 100 in step S230 of FIG. 5, but is not limited thereto.

The above-described user matching method using the artificial intelligence model between users has been described with reference to the flowchart shown in the drawings. For the sake of simplicity, the user matching method using the artificial intelligence model has been described by showing a series of blocks, but the present invention is not limited to the order of the blocks, and some blocks are performed in a different order from that shown and described in this specification. Or can be performed simultaneously. In addition, new blocks not described in the specification and drawings may be added, or some blocks may be deleted or changed.

In the above, embodiments of the present invention have been described with reference to the accompanying drawings, but those of ordinary skill in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features. You will be able to understand. Therefore, the embodiments described above are illustrative in all respects, and should be understood as non-limiting.

100: user matching device (computing device)
200: user terminal
300: external server
400: network

Claims (10)

In the method performed by the computing device,
Obtaining a matching request between the first user and the second user;
Extracting valid data for the matching request through a pre-learned artificial intelligence model using pre-stored user information and additional information input from the outside as training data; And
Matching the first user and the second user based on the extracted valid data,
The step of extracting the valid data,
In order to sell the product to the first user, when a sales process including a sales procedure for selling the product and a consultation procedure for performing consultation at preset periods for a predetermined period is to be performed, the sales procedure is to be performed. A second step of extracting valid data on information of a second user, and dividing the predetermined period into a plurality of unit periods, and optimally performing the consultation procedure based on attributes of each of the divided plurality of unit periods. Extracting valid data on user information;
The first user and the second user matched according to the extracted valid data analyze counseling data generated by performing a counseling procedure to determine the preference of the first user for the second user, and the determination Determining that a change of the second user matched with the first user is necessary when the preference of the first user is less than the reference value;
If it is determined that the change of the second user is necessary, re-determining a second user to be matched with the first user, and extracting valid data regarding the re-determined second user information; And
When the number of products to be sold to the first user is plural, a second user to sell each of the plurality of products to the first user is determined, but the number of second users matching the first user has a minimum value. Determining one or more second users, and extracting valid data regarding information of the determined one or more second users; including,
The step of matching the first user and the second user,
Based on the extracted valid data, the sales procedure is performed by matching the first user and a second user to perform the sales procedure, and when the first user purchases the product through the sales procedure, the Automatically matching a second user to perform the counseling procedure with the first user every preset period;
In response to extracting valid data regarding the re-determined second user information, changing and re-matching the second user matched with the first user to the re-determined second user; And
Based on the extracted valid data, the first user and the determined one or more second users are matched, and the determined one or more second users and the second user according to a preset product sales priority and whether a preset condition is satisfied. 1 step of sequentially matching users; containing,
User matching method using artificial intelligence model. The method of claim 1,
The user information includes information on a customer as a first user, and the additional information includes at least one of information on a seller as a second user, information on a product, or information related to a sales activity,
The step of extracting the valid data,
Selling information of a product to be sold to the customer, information of a seller who will sell the product, and the product by using at least one of the customer information, the seller information, and the product information as input of the artificial intelligence model Extracting valid data including at least one of the information of the target customer;
Extracting valid data regarding information of a first product suitable for sale by the first seller and information of a first customer who will sell the first product by using the information of the first seller as an input of the artificial intelligence model ;
Valid data regarding information on a second product suitable for selling to the second customer and information on a second seller who will sell the second product to the second customer by using the information of the second customer as the input of the artificial intelligence model Extracting; And
The step of extracting valid data on information of a third seller who will sell the third product and information of a third customer who will be selling the third product by using the information of the third product as an input of the artificial intelligence model. Included,
User matching method using artificial intelligence model. The method of claim 1,
The step of extracting the valid data,
In order to sell a product to the first user, when a sales process including a plurality of consultation procedures is to be performed, a second user to perform each of the plurality of consultation procedures is determined, and the determined second user's information And extracting valid data for,
The step of matching the first user and the second user,
Comprising the step of matching each of the determined second users with the first user each time the plurality of consultation procedures are performed based on the extracted valid data,
User matching method using artificial intelligence model. delete delete delete The method of claim 1,
The step of matching the first user and the second user,
If the first user and the plurality of second users are matched based on the extracted valid data, grouping the plurality of second users into one group and matching the first user; And
Providing matching result data generated by matching any one second user included in the one group and the first user to other second users included in the one group,
User matching method using artificial intelligence model. The method of claim 1,
Further comprising the step of retraining the artificial intelligence model by using the extracted valid data and actual matching result data according to the extracted valid data as training data,
User matching method using artificial intelligence model. A memory for storing one or more instructions; And
A processor that executes the one or more instructions stored in the memory,
The processor executes the one or more instructions,
An apparatus for performing the method of claim 1. A computer program combined with a computer as hardware and stored in a recording medium readable by a computer to perform the method of claim 1.

Images