System for providing automatic response shopping service using artificial intelligence

Abstract

The present invention relates to an automatic response shopping service provision system using artificial intelligence, comprising: a management server that stores and manages product information generated and received from a seller terminal, performs text conversion on voice input information generated and received from a buyer terminal to generate text information, and outputs and provides a webpage for product information guidance or purchase corresponding to the text information based on the generated text information using an integrated artificial intelligence platform. Through this, the automatic response shopping service providing system using artificial intelligence of the present invention can provide a shopping service that further satisfies the needs of the buyer in a user-responsive manner by considering the buyer's purchase history and preferences, as well as event details related to the buyer's product shopping, when providing product information that the buyer intends to purchase.

Description

Technology Field

[0001] The present invention relates to an automated response shopping service provision system using artificial intelligence. Background Technology

[0002] Artificial intelligence technology is a technology that implements human capabilities such as vision, perception, learning, reasoning, and natural language processing into programs so that computers can execute them. It is achieving rapid development by moving beyond simple data processing and utilizing raw data based on cloud computing and big data.

[0003] Recently, research on such artificial intelligence technology has moved beyond simple recognition and perception of input data to continuously advance toward a stage where it derives optimal solutions and countermeasures within an environment based on input data, performs knowledge-based reasoning and prediction through self-learning, and independently predicts and resolves problems expected to occur in the future.

[0004] Accordingly, there are increasing attempts to incorporate artificial intelligence technology into business areas related to the provision of shopping services. In particular, there are attempts to automate tasks such as predicting and detecting customer behavior and improving user experience, as well as to apply it to various tasks such as user personalization and enhancing brand reputation.

[0005] Specifically, by utilizing artificial intelligence technology to drive technological changes in various areas such as product and image search, demand forecasting, order management, consultation, recommendations, and logistics / delivery, it is inducing the attraction of new consumers and the expansion of consumption by providing new customer shopping experiences and improving user convenience.

[0006] In this regard, prior art for a system for providing shopping services built to improve consumer satisfaction by recommending the latest trendy items to consumers, and for operating such a system based on artificial intelligence, includes the "AI-based goods shopping mall provision system" (hereinafter referred to as "prior art") of Korean Registered Patent Publication No. 10-2455777.

[0007] However, existing shopping service provision systems using artificial intelligence, including conventional technology, still failed to significantly improve the convenience of users attempting to make a purchase, and there was a problem in that they were limited to building business models for the younger generation who are familiar with using IT devices and do not experience difficulties in using them. The problem to be solved

[0008] The present invention was created to solve the aforementioned problems, and the objective of the present invention is to provide an automated response shopping service provision system using artificial intelligence that allows users of all age groups to easily and conveniently receive various functions efficiently through voice alone when receiving a shopping service to purchase specific products, and naturally and effectively induces purchases preferred by the buyer by taking into account the buyer's purchase history or service usage history. means of solving the problem

[0009] To achieve the above objective, the automatic response shopping service provision system using artificial intelligence according to the present invention comprises: a management server that stores and manages product information generated and received from a seller terminal, performs text conversion on voice input information generated and received from a buyer terminal to generate text information, and outputs and provides a webpage for guidance on product information or purchase corresponding to the text information based on the generated text information using a linked artificial intelligence platform; wherein the management server comprises: a product information management unit that receives product information including product details, product category information, and product price information from the seller terminal; an information storage unit that stores and manages product information received through the product information management unit; a natural language indexer model management unit in which a first natural language indexer model configured and established is repeatedly trained using the product information received through the product information management unit; and a user event information management unit that receives voice input information from the buyer terminal and receives user event information generated when the buyer terminal performs a configured specific signal input operation within a webpage for guidance on specific product information or purchase provided through the management server. A user preference product recommendation model management unit that establishes and constructs a second natural language indexer model based on a natural language indexer model learned and stored through the natural language indexer model management unit and user event information received through the user event information management unit to enable iterative learning, and generates user product preference information using the user event information received through the user event information management unit and the second natural language indexer model;A purchase probability prediction model management unit that generates user purchaseability preference information using the user event information received through the user event information management unit and the user event information generated through the user event information management unit, wherein a pre-configured purchase probability prediction model is iteratively trained using user product preference information generated through the user preferred product recommendation model management unit and the pre-trained purchase probability prediction model; and a shopping service providing unit that extracts product information corresponding to the text information within the information storage unit and lists it using text information generated by text conversion processing of voice input information received through the user event information management unit and the pre-linked artificial intelligence platform, and controls the output form of a webpage for product information guidance or purchase using the number of listed product information, the user preferred product recommendation model, and the purchase probability prediction model.

[0010] Here, the natural language indexer model management unit comprises: a product metadata extraction unit that extracts and generates metadata for each product information based on product information received through the product information management unit; and a first natural language indexer model management unit that enables the first natural language indexer model to perform deep learning through iterative learning under pre-set conditions using the metadata for each product information extracted and generated through the product metadata extraction unit, and enables the learned first natural language indexer model to be stored and managed in a separate database provided within the information storage unit.

[0011] Additionally, the user event information management unit comprises: a first information receiving unit that receives voice input information from the buyer terminal; a second information receiving unit that receives user event information from the buyer terminal; a first event metadata management unit that extracts and generates metadata for each user event information based on the user event information received through the second information receiving unit, and stores and manages it in a separate database provided within the information storage unit; and a second event metadata management unit that generates user profiling data by processing the metadata for each user event information extracted and generated through the first event metadata management unit using data profiling, and stores and manages it in a separate database provided within the information storage unit.

[0012] In addition, the user preferred product recommendation model management unit includes: an embedded text information generation unit that generates embedded text information by performing word embedding processing using metadata for each event information stored in the information storage unit after extraction and generation through the first event metadata management unit; and a second natural language indexer model management unit that establishes and constructs the second natural language indexer model using the embedded text information generated through the embedded text information generation unit and the first natural language indexer model stored in the information storage unit after learning through the first natural language indexer model management unit, thereby performing deep learning through iterative learning under pre-set conditions, and ensures that the learned second natural language indexer model is stored and managed in a separate database prepared in the information storage unit.

[0013] Additionally, the user preferred product recommendation model management unit further includes a user product preference information generation unit that transmits user profiling data, which is generated through the second event metadata management unit and stored in the information storage unit, and the second natural language indexer model, which is trained through the second natural language indexer model management unit and stored in the information storage unit, to the linked artificial intelligence platform, and receives user product preference information that is extracted and generated when user profiling data regarding specific product information is input using the second natural language indexer model as a user preferred product recommendation model.

[0014] Additionally, the purchase probability prediction model management unit comprises: a purchase history information extraction unit that receives user product preference information provided after extraction and generation from the artificial intelligence platform through the user product preference information generation unit, and generates purchase history information by extracting metadata of event information generated according to the performance of a signal input operation regarding the execution of a purchase within a webpage for purchasing by a buyer terminal from among the metadata for each event information stored in the information storage unit after extraction and generation through the first event metadata management unit; and a purchase probability prediction model management unit that causes the purchase probability prediction model to perform deep learning through iterative learning under preset conditions using the user product preference information received through the purchase history information extraction unit and the purchase history information extracted and generated through the purchase history information extraction unit, and ensures that the learned purchase probability prediction model is stored and managed in a separate database prepared within the information storage unit. and includes a user purchaseability preference information generation unit that generates user purchaseability preference information by inputting user profiling data regarding specific product information among the user profiling data stored in the information storage unit after generating it through the second event metadata management unit, learning it through the purchase probability prediction model management unit, and inputting it into the purchase probability prediction model stored in the information storage unit.

[0015] In addition, the shopping service providing unit comprises: a text conversion management unit that processes voice input information received through the user event information management unit into text to generate text information corresponding to the voice input information; and an artificial intelligence management unit that uses the linked artificial intelligence platform to extract and list product information corresponding to the text information generated through the text conversion management unit from among the product information stored and managed within the information storage unit.

[0016] Additionally, the shopping service providing unit is further included in a first shopping service providing unit that is activated when there is no product information extracted and listed in correspondence with text information through the artificial intelligence management unit, provides first guidance information guiding the extraction and listing results to the buyer terminal, and provides a recommended product guidance webpage to the buyer terminal for providing a list of at least one product information within a pre-set ranking with high user product preference based on the second natural language indexer model; and after providing the recommended product guidance webpage to the buyer terminal, the first shopping service providing unit provides a product information guidance webpage to the buyer terminal for providing detailed information of specific product information corresponding to the recommended product selection signal based on the recommended product selection signal received after input generation from the buyer terminal.

[0017] delete

[0018] delete Effects of the invention

[0019] According to the present invention, the following effects are achieved.

[0020] First, buyers can easily and conveniently make a purchase by receiving guidance on the product they wish to buy using only voice input.

[0021] Second, in providing product information that the buyer intends to purchase, it is possible to provide a shopping service that further satisfies the buyer's needs in a user-responsive manner by considering the buyer's purchase history and preferences, as well as event details related to the buyer's product shopping.

[0022] Third, even if the buyer has not selected the product information they wish to purchase, it is possible to automatically respond with shopping services and provide information on recommended products that would be helpful to the buyer by considering the buyer's purchase history and preferences, as well as event details related to the buyer's shopping. Brief explanation of the drawing

[0023] FIG. 1 is a configuration diagram illustrating an automatic response shopping service provision system using artificial intelligence according to the present invention. FIG. 2 is a block diagram illustrating the detailed configuration within an automatic response shopping service provision system using artificial intelligence according to the present invention. FIG. 3 is a flowchart illustrating the process of performing a management function through a natural language indexer model management unit within an automatic response shopping service providing system using artificial intelligence according to the present invention. FIG. 4 is a flowchart illustrating the process of performing a management function through a user event information management unit within an automatic response shopping service provision system using artificial intelligence according to the present invention. FIG. 5 is a flowchart illustrating the process of performing a management function through a user preferred product recommendation model management unit within an automatic response shopping service provision system using artificial intelligence according to the present invention. FIG. 6 is a flowchart illustrating the process of performing a management function through a purchase probability prediction model management unit within an automatic response shopping service provision system using artificial intelligence according to the present invention. FIG. 7 is a flowchart illustrating the process of performing a service provision function through a shopping service provision unit within an automatic response shopping service provision system using artificial intelligence according to the present invention. Specific details for implementing the invention

[0024] Preferred embodiments of the present invention will be described in more detail with reference to the attached drawings, provided that technical details that are already well known are omitted or compressed for the sake of brevity.

[0025] <Description of an automated response shopping service provision system using artificial intelligence >

[0026] Referring to FIG. 1, the automatic response shopping service provision system (100) using artificial intelligence of the present invention is a system that not only allows a buyer to receive guidance on a product they wish to purchase among various products registered by a seller using only voice, but also further guides the buyer to the optimal product that can satisfy their needs according to their purchase history and product preference information, thereby increasing purchase satisfaction. To this end, it includes a buyer terminal (100); a seller terminal (200); a management server (300); a voice-to-text conversion platform (400) and an artificial intelligence platform (500); and the actual configuration is based on the management server (300).

[0027] The management server (110) stores and manages product information generated and received from the seller terminal (200), generates text information by performing text conversion of voice input information generated and received from the buyer terminal (100) in conjunction with the voice-to-text conversion platform (400), and outputs and provides a webpage for product information guidance or purchase corresponding to the text information based on the generated text information using the linked artificial intelligence platform (500).

[0028] That is, the management server (110) provides access to a shopping mall-type website (which can be implemented in various ways, such as programs, applications, etc.) to the buyer terminal (100) and the seller terminal (200), thereby enabling the seller terminal (200) to access the website and register product information so that the purchase of the product can be made, and the buyer terminal (100) can access the website and receive a shopping service that allows the entire process from providing product information guidance to purchasing to be done conveniently and simply using only voice.

[0029] To this end, the management server (110) includes a product information management unit (310), an information storage unit (320), a natural language index model management unit (330), a user event information management unit (340), a user preferred product recommendation model management unit (350), a purchase probability prediction model management unit (360), and a shopping service provision unit (370).

[0030] First, the product information management unit (310) receives product information including product details, product category information, and product price information from the seller terminal (200), and builds a database within the information storage unit (320) so that it can be stored and managed.

[0031] Here, the product information registered by the seller terminal (200) includes detailed information containing various information such as the detailed form, description, image, and precautions of the product, product category information indicating the category as a major classification of the product, and product price information including various information related to the physical value of the product, such as the price and discount details of the product, and in addition to this, various information that helps the buyer decide to purchase the product may be included.

[0032] In addition, the information storage unit (320) stores and manages product information received through the product information management unit (310) as described above, and various information and metadata, including the first natural language indexer model, the second natural language indexer model, and the purchase probability prediction model to be described later, are stored and managed separately by constructing individual databases.

[0033] Next, the natural language index model management unit (330) enables the first natural language indexer model (Indexer Model), which has been configured and built, to perform iterative learning using product information received through the product information management unit (310).

[0034] To this end, the natural language index model management unit (330) includes a product metadata extraction unit (331) and a first natural language indexer model management unit (332), and has an operating form as shown in FIG. 3.

[0035] First, the product metadata extraction unit (331) extracts and generates metadata for each product information based on the product information received through the product information management unit (310) (S110).

[0036] Following this, the first natural language indexer model management unit (332) uses the metadata for each product information extracted and generated through the product metadata extraction unit to enable the first natural language indexer model to perform deep learning through iterative learning under pre-set conditions (specific date, specific time, metadata generation time, etc.) (S130).

[0037] In addition, the first natural language indexer model management unit (332) ensures that the learned first natural language indexer model is stored and managed (S130) by providing a separate database within the information storage unit (320), and repeats this series of processes daily (S140).

[0038] Next, the user event information management unit (340) receives voice input information from the buyer terminal (100) and receives user event information generated when the buyer terminal (100) performs a pre-set specific signal input operation within a webpage for guidance or purchase of specific product information provided through the management server (300).

[0039] In this regard, the user event information management unit (340) includes a first information receiving unit (341), a second information receiving unit (342), a first event metadata management unit (343), and a second event metadata management unit (344), and has an operating form as shown in FIG. 4.

[0040] First, the user event information management unit (340) inserts pixels (S210) into web pages that can be outputted on the website, and by tracking the usage history of the buyer terminal (100) that receives the web pages and uses them via signal input, user event information is generated as specific actions such as page views, clicks, and purchases are performed, and the user event information generated in this way is continuously provided.

[0041] Next, the first information receiving unit (341) receives voice input information from the buyer terminal (100) (S220) and provides it to the text conversion management unit (371), which will be explained later, so that it can be converted into text and utilized.

[0042] Additionally, the second information receiving unit (342) receives user event information from the buyer terminal (100) (S220).

[0043] Here, the first information receiving unit (341) and the second information receiving unit (342) can be implemented in the form of a General Purpose API (ECS, Elastic Container Service).

[0044] Next, the first event metadata management unit (343) extracts and generates metadata for each user event information based on the user event information received through the second information receiving unit (342), and then prepares a separate database within the information storage unit (320) to store and manage it (S230).

[0045] Additionally, the second event metadata management unit (344) processes the metadata for each user event information extracted and generated through the first event metadata management unit (343) by data profiling to generate user profiling data, and then stores and manages the data by providing a separate database within the information storage unit (320) (S240).

[0046] Specifically, the second event metadata management unit (344) sets the generated user profiling data as a key using separate identification information such as a user ID and stores and manages it in a NoSQL DB within the information storage unit (320) (S240).

[0047] Additionally, the user preferred product recommendation model management unit (350) establishes and builds a second natural language indexer model based on the natural language indexer model learned and stored through the natural language indexer model management unit (330) and user event information received through the user event information management unit (340), thereby enabling iterative learning.

[0048] In addition, the user preferred product recommendation model management unit (350) generates user product preference information using user event information received through the user event information management unit (340) and a second natural language indexer model.

[0049] To this end, the user preferred product recommendation model management unit (350) includes an embedded text information generation unit (351), a second natural language index model management unit (352), and a user product preference information generation unit (353), and has an operating form as shown in FIG. 5.

[0050] First, the embedded text information generation unit (351) generates embedded text information (S310) by performing word embedding processing using metadata for each event information stored in the information storage unit after extraction generation through the first event metadata management unit (343).

[0051] Next, the second natural language index model management unit (352) establishes and constructs the second natural language index model using the embedded text information generated through the embedded text information generation unit (351) and the first natural language index model stored in the information storage unit (320) after learning through the first natural language index model management unit (332).

[0052] Here, if the first natural language indexer model is a model regarding products, the second natural language indexer model is used as a model regarding users and products for the purpose of a user-preferred product recommendation model.

[0053] The second natural language index model management unit (352) enables the second natural language indexer model, which has been configured, to perform deep learning through repeated learning under pre-configured conditions, and enables the learned second natural language indexer model to be stored and managed in a separate database within the information storage unit (S330), and this series of processes is repeated daily (S340).

[0054] Finally, the user product preference information generation unit (353) can be implemented in the form of an artificial intelligence API (ECS), and the user profiling data stored in the information storage unit (320) after generation through the second event metadata management unit (344) and the second natural language indexer model stored in the information storage unit (320) after learning through the second natural language indexer model management unit (352) are transmitted to the linked artificial intelligence platform (500).

[0055] Through this, an artificial intelligence platform (500), such as GPT linked with a management server (300), uses a second natural language indexer model as a model for recommending user-preferred products, and when user profiling data regarding specific product information is input, extracts a predicted result value regarding user product preference to generate user product preference information (S350).

[0056] In this way, user product preference information generated in conjunction with the artificial intelligence platform (500) is provided again to the user product preference information generation unit (353), and the user product preference information generation unit (353) transmits the provided user product preference information to the shopping service providing unit (370) to be described below.

[0057] Next, the purchase probability prediction model management unit (360) repeatedly learns the pre-configured purchase probability prediction model using user product preference information generated through the user preferred product recommendation model management unit (350) and user event information received through the user event information management unit (340), and generates user purchaseability preference information using the user event information received through the user event information management unit (340) and the pre-trained purchase probability prediction model.

[0058] To this end, the purchase probability prediction model management unit (360) includes a purchase history information extraction unit (361), a purchase probability prediction model management unit (362), and a user purchaseability preference information generation unit (363), and has an operating form as shown in FIG. 6.

[0059] First, the purchase history information extraction unit (361) receives user product preference information provided after extraction from the artificial intelligence platform (500) through the user product preference information generation unit (353), and generates purchase history information (S410) by extracting metadata of event information generated according to the signal input operation regarding the purchase execution within the webpage for purchasing on the buyer's terminal from among the metadata for each event information stored in the information storage unit (320) after extraction through the first event metadata management unit (343).

[0060] Next, the purchase probability prediction model management unit (362) uses the user product preference information received through the purchase history information extraction unit (361) and the purchase history information extracted and generated through the purchase history information extraction unit (361) to enable the purchase probability prediction model to perform deep learning through iterative learning under pre-set conditions (S420).

[0061] In addition, the purchase probability prediction model management unit (362) ensures that the learned purchase probability prediction model is stored and managed (S430) in a separate database within the information storage unit (320), and that this series of processes is repeated daily (S440).

[0062] Additionally, the user purchaseability preference information generation unit (363) generates user purchaseability preference information (S450) by inputting user profiling data regarding specific product information among the user profiling data stored in the product information storage unit (320) after generating it through the second event metadata management unit (344), into the purchase probability prediction model stored in the information storage unit (320) after learning through the purchase probability prediction model management unit (362).

[0063] Finally, the shopping service providing unit (370) processes the voice input information received through the user event information management unit (340) into text to generate text information and uses the linked artificial intelligence platform (500) to extract product information corresponding to the text information within the information storage unit (320) and list it, and controls the output form of the webpage for product information guidance or purchase using the number of listed product information, the user preferred product recommendation model, and the purchase probability prediction model.

[0064] To this end, the shopping service providing unit (370) includes a text conversion management unit (371), an artificial intelligence management unit (372), a first shopping service providing unit (373), a second shopping service providing unit (374), and a third shopping service providing unit (375) as shown in FIG. 2, and has an operating form as shown in FIG. 7.

[0065] First, the text conversion management unit (371) processes the voice input information received through the user event information management unit (340) into text and generates text information corresponding to the voice input information (S510).

[0066] Specifically, the text conversion management unit (371) can perform text conversion using a module or software provided in-house to the voice input information received through the user event information management unit (340), or, depending on the implementation, can provide the received voice input information through linkage with a voice-to-text conversion platform (400), such as an external STT (Speech To Text) engine linked as shown in FIGS. 1 and 2, and then receive the converted text information after text conversion processing.

[0067] Next, the artificial intelligence management unit (372) uses the linked artificial intelligence platform (500) to extract product information corresponding to text information generated through the text conversion management unit (371) among the product information stored and managed in the information storage unit (320), and lists it (S520).

[0068] Here, the product information extracted for listing is classified into cases of 0, 1, or 2 or more items, and takes the form of a shopping service provided according to each case.

[0069] Specifically, the first shopping service providing unit (373) is activated when there is no product information that has been extracted and listed in correspondence with text information through the artificial intelligence management unit (372), and provides the first guidance information (S530) to the buyer terminal (100) that provides the results of extraction and listing (for example, information including a message such as "We could not find a product that matches what you said. Please try again" so that the user of the buyer terminal can recognize that they could not find product information corresponding to text information based on voice input information).

[0070] At the same time, the first shopping service provider (373) outputs and provides (S530) a webpage for providing recommended products to the buyer terminal (100) to provide a list of at least one product information within a pre-set ranking (e.g., top 5 products) with high user preference for all products, based on a second natural language indexer model for the user corresponding to the buyer terminal (100).

[0071] Through this, the user of the buyer terminal (100) can choose whether to attempt to generate voice input information again, or to proceed with the purchase of a product among the recommended products that they intended to purchase.

[0072] Accordingly, the first shopping service provider (373) provides a webpage for guiding recommended products to the buyer terminal (100), and then, based on the recommended product selection signal received after input generation from the buyer terminal (100), provides a webpage for guiding detailed information of specific product information corresponding to the recommended product selection signal to the buyer terminal (100) (S560), thereby providing detailed information about the product so that it can lead to a purchase.

[0073] Next, the second shopping service provision unit (374) is activated when there is one product information that is extracted and listed in correspondence with text information through the artificial intelligence management unit (372), and requests and receives user product preference information and user purchaseability preference information regarding the extracted and listed product information from the user product preference information generation unit (353) and the user purchaseability preference information generation unit (363) to perform a comparison judgment (S540).

[0074] delete

[0075] delete

[0076] delete

[0077] delete

[0078] delete

[0079] delete

[0080] delete

[0081] delete

[0082] delete

[0083] The embodiments disclosed in this invention are intended to illustrate, not limit, the technical concept of the invention, and the scope of the technical concept of the invention is not limited by these embodiments. The scope of protection shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of this invention. Explanation of the symbols

[0084] 10: Automated response shopping service provision system using artificial intelligence 100 : Buyer's terminal 200 : Seller's terminal 300 : Management Server 310 : Product Information Management Unit 320 : Information storage unit 330 : Natural Language Index Model Management Unit 331 : Product Metadata Extraction Section 332 : First Natural Language Indexer Model Management Department 340 : User Event Information Management Unit 341: 1st Information Receiving Unit 342: 2nd Information Receiving Unit 343 : 1st Event Metadata Management Department 344 : 2nd Event Metadata Management Department 350 : User Preferred Product Recommendation Model Management Unit 351 : Embedded Text Information Generation Unit 352 : Second Natural Language Index Model Management Department 353 : User Product Preference Information Generation Section 360: Purchase Probability Prediction Model Management Unit 361 : Purchase History Information Extraction Unit 362 : Purchase Probability Prediction Model Management Department 363 : User Purchase Preference Information Generation Section 370 : Shopping Service Provision Unit 371 : Text Conversion Management Department 372 : Artificial Intelligence Management Department 373 : 1st Shopping Service Provision Department 374 : Second Shopping Service Provision Department 375 : Third Shopping Service Provision Department 400 : Speech-to-Text Conversion Platform 500: Artificial Intelligence Platform

Claims

Claim 1 A management server that stores and manages product information generated and received from a seller terminal, generates text information by performing text conversion on voice input information generated and received from a buyer terminal, and outputs and provides a webpage for guidance on product information or purchase corresponding to the text information based on the generated text information using a linked artificial intelligence platform; wherein the management server includes: a product information management unit that receives product information including product details, product category information, and product price information from the seller terminal; an information storage unit that stores and manages product information received through the product information management unit; a natural language indexer model management unit in which a first natural language indexer model configured and established is repeatedly trained using the product information received through the product information management unit; and a user event information management unit that receives voice input information from the buyer terminal and receives user event information generated when the buyer terminal performs a configured specific signal input operation within a webpage for guidance on specific product information or purchase provided through the management server. A user preference product recommendation model management unit that establishes and constructs a second natural language indexer model based on a natural language indexer model learned and stored through the natural language indexer model management unit and user event information received through the user event information management unit, and generates user product preference information using the user event information received through the user event information management unit and the second natural language indexer model; and a purchase probability prediction model management unit that repeatedly learns a pre-established purchase probability prediction model using user product preference information generated through the user preference product recommendation model management unit and user event information received through the user event information management unit, and generates user purchaseability preference information using the user event information received through the user event information management unit and the pre-learned purchase probability prediction model.and a shopping service providing unit that extracts and lists product information corresponding to the text information within the information storage unit using text information generated by text conversion processing of voice input information received through the user event information management unit and the linked artificial intelligence platform, and controls the output form of a webpage for product information guidance or purchase using the number of listed product information, the user preferred product recommendation model, and the purchase probability prediction model; wherein the natural language indexer model management unit includes: a product metadata extraction unit that extracts and generates metadata for each product information based on product information received through the product information management unit; and a first natural language indexer model management unit that causes the first natural language indexer model to perform deep learning through iterative learning under preset conditions using the metadata for each product information extracted and generated through the product metadata extraction unit, and ensures that the learned first natural language indexer model is stored and managed in a separate database prepared within the information storage unit; and wherein the user event information management unit includes a first information receiving unit that receives voice input information from the buyer terminal; A second information receiving unit that receives user event information from the buyer terminal; a first event metadata management unit that extracts and generates metadata for each user event information based on the user event information received through the second information receiving unit, and stores and manages it in a separate database provided within the information storage unit; and a second event metadata management unit that generates user profiling data by processing the metadata for each user event information extracted and generated through the first event metadata management unit using data profiling, and stores and manages it in a separate database provided within the information storage unit.A system for providing an automated response shopping service using artificial intelligence, comprising: an embedded text information generation unit that generates embedded text information by performing word embedding processing using metadata for each event information stored in the information storage unit after extraction and generation through the first event metadata management unit; and a second natural language indexer model management unit that establishes and constructs the second natural language indexer model using the embedded text information generated through the embedded text information generation unit and the first natural language indexer model stored in the information storage unit after learning through the first natural language indexer model management unit, thereby performing deep learning through iterative learning under pre-set conditions, and ensuring that the learned second natural language indexer model is stored and managed in a separate database prepared in the information storage unit. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 The system for providing an automatic response shopping service using artificial intelligence according to claim 1, wherein the user preferred product recommendation model management unit transmits user profiling data stored in the information storage unit after being generated through the second event metadata management unit and the second natural language indexer model stored in the information storage unit after being learned through the second natural language indexer model management unit to the linked artificial intelligence platform, and receives user product preference information that is extracted and generated when user profiling data regarding specific product information is input using the second natural language indexer model as a model for recommending user preferred products. Claim 6 In claim 5, the purchase probability prediction model management unit comprises: a purchase history information extraction unit that receives user product preference information provided after extraction and generation from the artificial intelligence platform through the user product preference information generation unit, and generates purchase history information by extracting metadata of event information generated according to the performance of a signal input operation regarding the performance of a purchase within a webpage for purchasing by a buyer terminal from among the metadata for each event information stored in the information storage unit after extraction and generation through the first event metadata management unit; and a purchase probability prediction model management unit that causes the purchase probability prediction model to perform deep learning through iterative learning under preset conditions using the user product preference information received through the purchase history information extraction unit and the purchase history information extracted and generated through the purchase history information extraction unit, and causes the learned purchase probability prediction model to be stored and managed in a separate database prepared within the information storage unit. A system for providing an automated response shopping service using artificial intelligence, characterized by comprising: a user purchaseability preference information generation unit that generates user purchaseability preference information by inputting user profiling data regarding specific product information among user profiling data stored in the information storage unit after generating it through the second event metadata management unit and learning it through the purchase probability prediction model management unit and inputting it into the purchase probability prediction model stored in the information storage unit. Claim 7 In claim 6, the shopping service providing unit comprises: a text conversion management unit that processes voice input information received through the user event information management unit into text to generate text information corresponding to the voice input information; and an artificial intelligence management unit that uses the linked artificial intelligence platform to extract and list product information corresponding to the text information generated through the text conversion management unit from among the product information stored and managed within the information storage unit. Claim 8 In claim 7, the shopping service providing unit is further comprising: a first shopping service providing unit that is activated when there is no product information extracted and listed in correspondence with text information through the artificial intelligence management unit, outputs and provides first guidance information guiding the extraction and listing results to the buyer terminal, and outputs and provides a recommended product guidance webpage to the buyer terminal for guiding a list of at least one product information within a pre-set ranking with high user product preference based on the second natural language indexer model; and the first shopping service providing unit outputs and provides a recommended product guidance webpage to the buyer terminal for guiding detailed information of specific product information corresponding to the recommended product selection signal based on the recommended product selection signal received after input generation from the buyer terminal. Claim 9 delete Claim 10 delete

Images