Voice-based live product registration and automatic comment response system

Abstract

The present invention relates to a voice-based live product registration and automatic comment response system that enables real-time product registration and sales by processing voice commands from a seller during a live broadcast using artificial intelligence, and enables the provision of customized responses in real-time by analyzing comments using artificial intelligence during a live broadcast. The live broadcasting device is implemented by including a voice recognition unit that recognizes a product registration voice transmitted from the seller terminal during a live broadcast; an artificial intelligence-based analysis unit that classifies information based on the voice recognized by the voice recognition unit, automatically extracts essential information, and automatically generates product information based on the extracted essential information to perform registration and sales; and a product information provision unit that provides the product information registered by the artificial intelligence-based analysis unit to a viewer terminal or provides it on the screen.

Description

Technology Field

[0001] The present invention relates to voice-based live product registration and automatic comment response, and more specifically, to a voice-based live product registration and automatic comment response system that enables real-time product registration and sales by processing voice commands from a seller during a live broadcast using artificial intelligence, and enables the provision of customized responses in real-time by analyzing comments using artificial intelligence during a live broadcast. Background Technology

[0002] Online e-commerce has been growing rapidly alongside the development of the Internet. Furthermore, with the rapid advancement of mobile internet and smartphone technology, the e-commerce market is rapidly being reorganized into a mobile environment, and recently, online e-commerce in the form of live commerce is gaining attention.

[0003] Live commerce refers to a form of e-commerce that introduces and sells products to viewers through real-time video streaming. Combining the term "live," meaning real-time, with "e-commerce," it can be defined as a combination of online shopping and broadcasting.

[0004] Through live commerce, consumers access product information while watching internet broadcasts, and the host and viewers exchange opinions. Additionally, they can make purchases directly on the broadcast page.

[0005] In other words, live commerce is the online transition of home shopping channels. However, there are differences from traditional home shopping channels. The biggest difference lies in the directionality. In home shopping channels, the delivery of information is unilateral. Of course, some channels collect and respond to viewers' opinions via text messages, for instance. However, live commerce involves a much more active and dynamic exchange of information than home shopping channels.

[0006] For example, in live commerce selling laptops, games are run right in front of everyone at the request of viewers. Additionally, the weight of the product is measured on a scale right away. Sometimes, if a streamer provides incorrect information, the error is immediately corrected by the viewers.

[0007] Another difference in live commerce is that the composition of broadcast content is more flexible compared to home shopping channels. Home shopping channels curate shows hosted by a few famous show hosts. In contrast, live commerce allows sellers to freely customize their broadcasts according to their preferences.

[0008] In addition, while home shopping channels are subject to legal regulations due to the nature of broadcasting, live commerce is relatively free from legal regulations, like other online services.

[0009] Therefore, live commerce is more entertaining and interesting than home shopping channels. Rather than professional broadcasters, various experts from all walks of life, such as brand representatives and influencers, appear on the show, and their remarks are free and tend to be quite risqué.

[0010] However, such live broadcasting has limitations in that it is difficult for general users who are not familiar with the technology or infrastructure for internet broadcasting to independently set up and conduct real-time streaming environments.

[0011] For example, for live broadcasting and live commerce, there is the inconvenience of sellers having to manually register all product information in advance, and in particular, there is a disadvantage in that product registration based on voice is impossible.

[0012] In addition, viewers can post comments during a live broadcast, but it is very difficult for the host to respond to these comments in real time.

[0013] Therefore, the present invention requires a technology that enables a seller to easily register product information during a live broadcast and to provide real-time responses to comments registered by viewers. Prior art literature

[0014] Republic of Korea Published Patent 10-2007-0105348 (Published Oct. 30, 2007) (Method and apparatus for performing real-time conversation with multiple viewers during a live broadcast program) Republic of Korea Registered Patent 10-2139348 (Registered Jul. 23, 2020) (Method for providing real-time product information) Republic of Korea Registered Patent 10-2053128 (Registered Dec. 02, 2019) (Method and apparatus for generating live streaming video, method and apparatus for providing live service, and live streaming system) Republic of Korea Registered Patent 10-2283232 (Registered Jul. 23, 2021) (Method and apparatus for providing information about products in real-time and producing real-time product information through live video broadcasting in a video commerce system) Republic of Korea Registered Patent 10-2407276 (Registered Jul. 03, 2022) (Apparatus and method for selling products based on two-way communication) Republic of Korea Published Patent 10-2023-0120899 (Published Aug. 17, 2023) (Method and system for providing live commerce through an AI human model) The problem to be solved

[0015] Therefore, the present invention is proposed to solve the disadvantages of not being able to register products in real time during general live broadcasts (live commerce) as described above and the problem of difficulty in providing real-time responses to registered comments. Its purpose is to provide a voice-based live product registration and automatic comment response system that enables real-time product registration and sales by processing voice commands from a seller during a live broadcast using artificial intelligence, and to provide customized responses in real time by analyzing comments using artificial intelligence during the live broadcast. means of solving the problem

[0016] In order to achieve the above-mentioned purpose, the "voice-based live product registration and automatic comment response system" according to the present invention is,

[0017] A seller terminal that requests product registration using voice during a live broadcast; a viewer terminal that purchases products sold during a live broadcast or registers comments; and a live broadcasting device that processes product information input via voice from the seller terminal during a live broadcast based on artificial intelligence to generate real-time product information and register it as a product for sale.

[0018] The above live broadcasting device is,

[0019] A voice recognition unit that recognizes product registration voice transmitted from the seller terminal during a live broadcast;

[0020] An AI-based analysis unit that classifies information based on voice recognized by the voice recognition unit, automatically extracts essential information, and automatically generates product information based on the extracted essential information to perform registration and sales;

[0021] It is characterized by including a product information providing unit that provides product information registered by the above-mentioned artificial intelligence-based analysis unit to a viewer terminal or provides it on the screen.

[0022] Preferably, the artificial intelligence-based analysis unit is,

[0023] A data receiving module that receives voice data recognized by the voice recognition unit above;

[0024] A classification module that classifies product data received from the above data receiving module by word, and extracts and classifies the classified word information according to pre-set essential items for generating product information;

[0025] It is characterized by including a product information generation module that generates product information based on essential items classified by the above classification module.

[0026] Preferably, the classification module is,

[0027] It is characterized by extracting at least one of category, product name, price, stock, and option as a required item from classified word information.

[0028] Preferably, the product information generation module is,

[0029] When generating product information using extracted essential items, if the generated product information is insufficient with only the extracted essential items, it is characterized by performing a supplementary query to derive the missing product information.

[0030] Preferably, the artificial intelligence-based analysis unit is,

[0031] A product information display module that displays product information generated by the above product information generation module on a seller terminal;

[0032] It is characterized by including a product information registration module that registers the generated product information as a product for sale in a product database when a request for product information registration is made from the seller terminal.

[0033] Preferably, the live broadcasting device is,

[0034] It is characterized by including a product information supplementation unit that requests product information supplementation by transmitting supplementary query information for product information generation to the seller terminal in conjunction with the above-mentioned artificial intelligence-based analysis unit.

[0035] Preferably, the live broadcasting device is,

[0036] It is characterized by including an artificial intelligence (AI) manager that collects comments registered by viewer terminals during a live broadcast and provides answers generated in real time through an AI-based analysis unit.

[0037] Preferably, the artificial intelligence manager above,

[0038] A comment collection unit that collects comments registered by viewer terminals in real time during a live broadcast;

[0039] It is characterized by including an answer providing unit that provides an answer generated for the above-mentioned collected comments.

[0040] Preferably, the artificial intelligence-based analysis unit is,

[0041] A comment collection module that collects comments registered by viewers in real time during a live broadcast;

[0042] A comment analysis module that analyzes comments collected by the above comment collection module by linking them with product information, analyzing them based on pre-registered sales policies, or analyzing them based on the user's past comment patterns;

[0043] It is characterized by including an answer generation module that generates and provides real-time answers based on the results analyzed by the comment analysis module above.

[0044] Preferably, the above-mentioned answer generation module is,

[0045] The above comment analysis module is characterized by automatically generating comment responses based on registered product information when the analysis method analyzes comments in conjunction with product information, automatically generating comment responses based on sales policy information including shipping, return, and payment methods when the analysis method analyzes based on sales policies, and automatically generating user-customized responses based on response information regarding the user's past comment patterns when the analysis method analyzes based on the user's past comment patterns. Effects of the invention

[0046] According to the present invention, when a seller gives a voice command during a live broadcast, it is processed based on artificial intelligence to enable real-time product registration and sales, thereby resolving the difficulty of the seller registering product information and simultaneously enabling real-time product registration during a live broadcast.

[0047] In addition, according to the present invention, comments can be analyzed based on artificial intelligence during a live broadcast to provide customized responses in real time, thereby resolving the problem where it is difficult for the host to respond to comments in real time during a live broadcast, while simultaneously providing the advantage of being able to respond to comments in real time during a live broadcast. Brief explanation of the drawing

[0048] FIG. 1 is a schematic diagram of a voice-based live product registration and automatic comment response system according to the present invention, and FIG. 2 is a configuration diagram of an exemplary embodiment of the seller terminal and live broadcasting device of FIG. 1, and FIG. 3 is a block diagram of an exemplary embodiment of the artificial intelligence-based analysis unit of FIG. 2, and FIG. 4 is a configuration diagram of an exemplary embodiment of the seller terminal, live broadcasting device, and viewer terminal of FIG. 1, and Figure 5 is a block diagram of an exemplary embodiment of the artificial intelligence-based analysis unit of Figure 4. Specific details for implementing the invention

[0049] A voice-based live product registration and automatic comment response system according to a preferred embodiment of the present invention will be described in detail below with reference to the attached drawings.

[0050] The terms or words used in the present invention described below should not be interpreted as being limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the present invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention.

[0051] Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely preferred embodiments of the present invention and do not represent all technical concepts of the present invention; thus, it should be understood that various equivalents and modifications that can replace them may exist at the time of filing this application.

[0052] FIG. 1 is a schematic diagram of a voice-based live product registration and automatic comment response system according to a preferred embodiment of the present invention, FIG. 2 is a diagram of an exemplary configuration of a seller terminal and a live broadcasting device of FIG. 1, FIG. 3 is a block diagram of an exemplary configuration of an artificial intelligence-based analysis unit of FIG. 2, FIG. 4 is a diagram of an exemplary configuration of a seller terminal, a live broadcasting device, and a viewer terminal of FIG. 1, and FIG. 5 is a block diagram of an exemplary configuration of an artificial intelligence-based analysis unit of FIG. 4.

[0053] The voice-based live product registration and automatic comment response system according to the present invention may include a live broadcasting device (100), a seller terminal (200), and a viewer terminal (300).

[0054] Here, the live broadcasting device (100), the seller terminal (200), and the viewer terminal (300) can perform mutual voice and data interfaces in real time through a network during the live broadcast.

[0055] The seller terminal (200) plays the role of requesting product registration using voice during a live broadcast. This seller terminal (200) can be implemented as a mobile device, such as a smartphone, with a product registration app installed.

[0056] As illustrated in FIG. 2, such a seller terminal (200) may include a product information input unit (210) that receives product information transmitted by voice from a seller and transmits it to a live broadcasting device (100), and a product information display unit (220) that displays product information generated from the live broadcasting device (100) on a screen.

[0057] The viewer terminal (300) serves the role of purchasing products sold in the live broadcast or registering comments. This viewer terminal (300) can be implemented as a personal computer (PC) or a mobile device such as a smartphone that includes a comment registration app.

[0058] As illustrated in FIG. 4, such a viewer terminal (300) may include a comment registration unit (310) for registering comments (chat) and a response display unit (320) for receiving and displaying a response to the registered comments. Hereinafter, the term "comment" should be understood as having the same meaning as "chat conversation."

[0059] Here, comments registered by viewers can be displayed on the video screen of the live broadcasting device (100), and answers generated in real time in response to them can also be displayed on the video screen.

[0060] The live broadcasting device (100) is a device that conducts live broadcasting (live commerce) via the internet, etc., and performs the role of processing sales product information input via voice from the seller terminal (200) based on artificial intelligence (AI) to generate real-time product information and register it as a sales product. In addition, the live broadcasting device (100) also performs the role of collecting comments registered by the viewer terminal (300) during the live broadcast and providing real-time generated comment responses through the AI-based analysis unit (120).

[0061] The live broadcasting device (100) may include a voice recognition unit (110) that recognizes a product registration voice transmitted from the seller terminal (200) during a live broadcast, an artificial intelligence-based analysis unit (120) that classifies information based on the voice recognized by the voice recognition unit (110), automatically extracts essential information, and automatically generates product information based on the extracted essential information to perform registration and sales, a product information provision unit (140) that provides the product information registered by the artificial intelligence-based analysis unit (120) to a viewer terminal (300) or displays it on a screen, a product database (150) in which product information, product sales policy information, user's past comment pattern information, data classification, product information generation, and comment analysis models are registered, and a product information supplementation unit (130) that requests product information supplementation by transmitting supplementary query information for product information generation to the seller terminal (200) in conjunction with the artificial intelligence-based analysis unit (120).

[0062] The above voice recognition unit (110) can recognize voice transmitted by a seller using conventional voice recognition and natural language processing (NLP) technology.

[0063] Additionally, the artificial intelligence-based analysis unit (120) may include a data receiving module (121) that receives voice data recognized by the voice recognition unit (110), a classification module (122) that classifies product data received by the data receiving module (121) by word and extracts and classifies it by pre-set essential items for generating product information from the classified word information, and a product information generation module (123) that generates product information based on the essential items classified by the classification module (122).

[0064] Here, the classification module (122) can extract and classify at least one of the categories, product names, prices, stock, and options from the classified word information as required items.

[0065] In addition, when the product information generation module (123) generates product information using extracted essential items, if the generation of product information is insufficient with only the extracted essential items, it can perform a supplementary query to derive the insufficient product information.

[0066] Meanwhile, the artificial intelligence-based analysis unit (120) may include a product information display module (124) that displays product information generated by the product information generation module (123) on a seller terminal (200), and a product information registration module (125) that registers the generated product information as a product for sale in a product database (150) when a request for product information registration is made from the seller terminal (200).

[0067] Additionally, the live broadcasting device (100) may include an artificial intelligence (AI) manager (160) that collects comments registered by a viewer terminal (300) during a live broadcast and provides answers generated in real time through an artificial intelligence-based analysis unit (120).

[0068] This artificial intelligence manager (160) may include a comment collection unit (161) that collects comments registered by a viewer terminal (300) during a live broadcast in real time, and an answer providing unit (162) that provides an answer generated for the collected comments.

[0069] Additionally, the artificial intelligence-based analysis unit (120) may include a comment collection module (126) that collects comments registered by viewers in real time during a live broadcast, a comment analysis module (127) that analyzes the comments collected by the comment collection module (126) in conjunction with product information, analyzes them based on pre-registered sales policies, or analyzes them based on the user's past comment patterns, and an answer generation module (128) that generates and provides a real-time answer based on the results analyzed by the comment analysis module (127).

[0070] The above-mentioned answer generation module (128) can automatically generate a comment answer based on registered product information when the analysis method of the above-mentioned comment analysis module (127) analyzes comments in conjunction with product information, automatically generate a comment answer based on sales policy information including delivery, return, and payment methods when the analysis method analyzes based on sales policy, and automatically generate a user-customized answer based on response information that responded to the user's past comment patterns when the analysis method analyzes based on the user's past comment patterns.

[0071] The operation of the voice-based live product registration and automatic comment response system according to the present invention configured as described above is specifically explained as follows.

[0072] First, the present invention may largely include a function for a seller to register products via voice in real time during a live broadcast, and a function to provide real-time responses by analyzing comments, such as chat messages, during a live broadcast.

[0073] The process of registering a product during an electronic live broadcast is as follows.

[0074] While the live broadcasting device (100) is conducting a live broadcast via the internet or the like, the product seller inputs product information (product category, product name, price, stock, options, etc.) to be determined by voice using the seller terminal (200). The product sales information input by voice is input through the product information input unit (210), after which voice signal processing is performed, and subsequently transmitted to the live broadcasting device (100) via the network.

[0075] When a voice is input from the seller terminal (200) while a live broadcast (live commerce) is being conducted using a network such as the internet, the live broadcasting device (100) recognizes the voice in a normal speech-to-text (STT) manner at the voice recognition unit (110), interprets the meaning using natural language processing (NLP) technology, generates appropriate voice recognition data, and transmits it to the artificial intelligence-based analysis unit (120).

[0076] The artificial intelligence-based analysis unit (120) classifies information based on recognized voice data, automatically extracts essential information, and automatically generates product information based on the extracted essential information to perform registration and sales. Here, the artificial intelligence-based analysis unit (120) may use a learning model composed of any one of a conventional forward neural network (FF), recurrent neural network (RNN), long short-term memory (LSTM), convolutional neural network (CNN), and generative adversarial network (GAN).

[0077] For example, the data receiving module (121) of the artificial intelligence-based analysis unit (120) receives voice data recognized by the voice recognition unit (110) and transmits it to the classification model (122).

[0078] The classification module (122) classifies the product data received from the data receiving module (121) by word and extracts and classifies the product information from the classified word information according to pre-set essential items for generating product information. Here, the essential items may be basic product information for selling products in typical live commerce, etc.

[0079] For example, the classification module (122) classifies word information by extracting at least one item among category, product name, price, stock, and option as a required item from the classified word information, and transmits the classified required item information to the product information generation module (123).

[0080] That is, when a seller inputs product information by voice, such as "This T-shirt costs 29,900 won, comes in two colors, black and white, and comes in sizes M and L," the T-shirt is extracted and classified into the clothing category and simultaneously extracted and classified into the product name, 29,900 won is extracted and classified into the price item, black and white are extracted and classified into color information, and M and L are extracted and classified into size.

[0081] The above product information generation module (123) generates product information based on the essential items classified by the above classification module (122).

[0082] When registering a product of a category based on the categories classified here, there may be required items for product registration. These required items for product registration are items that must be disclosed to the viewer (consumer) when selling the product. These required items for product registration for each category are pre-registered in the product database (150) for each category and product. Accordingly, the product information generation module (123) looks at the category and product name of the product information requested by the seller, extracts the corresponding required items for product registration, compares the two, and determines whether the sales information entered by the seller via voice is insufficient. If the sales information is not insufficient as a result of this determination, product information is generated immediately using the extracted and classified required item information.

[0083] In this method of generating product information using required item information, an AI-based product information generation model is registered in the product database in advance, and after optimizing the model through training, the required items are input into the model to simply and automatically generate the product information.

[0084] In contrast, if there is a lack of sales information, a supplementary inquiry is made to the product information supplementation unit (130) to derive the missing parts of the product information to be generated through the seller.

[0085] The above product information supplementation unit (130) links with the above artificial intelligence-based analysis unit (120) and transmits supplementary query information for generating product information to the above seller terminal (200) to request product information supplementation.

[0086] These supplementary inquiries may be transmitted as voice through the speaker of the seller terminal (200) or displayed in text format on the product information display unit (220).

[0087] After confirming the supplementary inquiry, the seller responds to it verbally.

[0088] The response voice input for the supplementary inquiry is recognized again through the voice recognition unit (110) and then transmitted to the product information generation module (123) via the data receiving module (121) and classification module (1220) of the artificial intelligence-based analysis unit (120).

[0089] The above product information generation module (123) automatically generates product information based on the previously obtained essential items and the essential items obtained through a supplementary query if it is possible to generate product information.

[0090] The product information generated in this way is transmitted to the seller terminal (200) through the product information display module (124) and is visually displayed through the product information display unit (220) of the seller terminal (200).

[0091] The seller views the product information displayed on the screen through the product information display unit (220) and requests registration if the product information is correct (this can be processed via a response using an input device or by voice). Of course, even if the product information is incorrect, the product information can be modified through a voice or response requesting a change or modification.

[0092] Meanwhile, when the artificial intelligence-based analysis unit (120) receives a product registration request from the seller terminal (200), the product information registration module (125) registers the generated product information as a product for sale in the product database (150).

[0093] The product information registered in the product database (150) in this way is displayed through the screen of the viewer terminal (300) or the live broadcasting device (100), and sales are made.

[0094] In this way, product information is automatically generated and registered based on AI through the seller's voice input during a live broadcast, thereby enabling sales and registration, making it possible to conveniently register product information in real time during the live broadcast.

[0095] Next, the latter, comment analysis and real-time automated responses, will be explained as follows.

[0096] Viewers watching a live broadcast (live commerce) via the internet, etc., can use a viewer terminal (300) to ask questions about the products being sold via comments (chat). That is, the comment register (310) of the viewer terminal (300) registers conversations such as questions about the products being sold through chat, etc., to the live broadcasting device (100).

[0097] When a chat conversation occurs from a viewer terminal (300) during a live broadcast, the AI ​​manager (160) of the live broadcasting device (100) collects the comments registered by the viewer terminal (300), transmits them to an artificial intelligence-based analysis unit (120), and provides a response generated in real time through the artificial intelligence-based analysis unit (120).

[0098] For example, the comment collection unit (161) of the artificial intelligence manager (160) collects comments registered by the viewer terminal (300) during a live broadcast in real time and transmits them to the artificial intelligence-based analysis unit (120).

[0099] The comment collection module (126) of the above artificial intelligence-based analysis unit (120) collects comments registered by viewers in real time during a live broadcast and transmits them to the comment analysis module (127).

[0100] The comment analysis module (127) analyzes the comments collected by the comment collection module (126) in conjunction with product information, analyzes them based on pre-registered sales policies, or analyzes them based on the user's past comment patterns.

[0101] For example, the comment analysis module (127) analyzes the collected comments, and if product information exists in the comments, it extracts product information registered in the product database (150) and transmits it to the answer generation module (128).

[0102] In addition, if the collected comments are analyzed and the comments contain major sales policy information such as shipping, returns, and payment methods, the sales policy information registered in the product database (150) is extracted and transmitted to the answer generation module (128).

[0103] Additionally, if the collected comments are analyzed and the conversation contains information about the user's previous purchase, the user's past comment patterns registered in the product database (150) are extracted and transmitted to the answer generation module (128).

[0104] The above-mentioned answer generation module (128) automatically generates a response to a comment based on the transmitted product information when product information is transmitted. That is, when the comment is a conversation such as "What sizes is this available?", size information (e.g., M, L, XL, 95, 100, 105, 110, etc.) is extracted from the above-mentioned product information, and the extracted size information is generated as response information.

[0105] In addition, when sales policy information is transmitted, the answer generation module (128) automatically generates a response to a comment based on information such as shipping, returns, and payment methods included in the sales policy information. That is, if the comment is policy information such as "How long does shipping take?", it extracts policy information such as "usually shipped within 2 to 3 days" from the policy information of the corresponding product that is registered in advance, and generates the extracted policy information as the response information.

[0106] Additionally, when the user's past comment pattern is transmitted, the answer generation module (128) analyzes the user's past comment pattern and automatically generates a customized answer. That is, if a viewer who is an existing customer comments (chat) "When will it be restocked?", the module analyzes the user's past comment information and automatically generates a user-customized answer such as "Is this the product X you previously purchased? It is scheduled to be restocked next week."

[0107] The answer to the comment generated in this way is provided through the answer providing unit (162).

[0108] The response to the comment can be displayed in real time on the broadcast screen of the live broadcasting device or provided to the viewer terminal (300).

[0109] When an answer is provided to the viewer terminal (300), it is displayed on the screen through the answer display unit (320) of the viewer terminal (300), allowing the viewer to check the answer to the comment registered by the viewer in real time.

[0110] In this way, when a viewer registers a comment about a product for sale during a live broadcast, an AI-based response is automatically generated in real time and provided immediately. This enables the host or seller to respond to real-time comments without the inconvenience of having to generate a separate response during the live broadcast.

[0111] Although the invention made by the inventors has been specifically described according to the above embodiments, it is obvious to those skilled in the art that the invention is not limited to the above embodiments and can be modified in various ways without departing from the gist thereof. Explanation of the symbols

[0112] 100: Live broadcasting equipment 110: Voice recognition unit 120: AI-based analysis unit 121: Data receiving module 122: Classification Module 123: Product Information Generation Module 124: Product Information Display Module 125: Product Information Registration Module 126: Comment Collection Module 127: Comment Analysis Module 128: Answer generation module 130: Product Information Security Section 140: Product Information Provision Department 150: Product Database 160: AI Manager 161: Comment Collection Department 162: Answer Provider 200: Seller's terminal 210: Product Information Input Section 220: Product Information Display 300: Viewer Terminal

Claims

Claim 1 A voice-based live product registration and automatic comment response system comprising: a seller terminal requesting product registration using voice during a live broadcast; a viewer terminal purchasing products sold during a live broadcast or registering comments; and a live broadcasting device that processes product information input via voice from the seller terminal during a live broadcast based on artificial intelligence to generate real-time product information and register it as a product for sale, wherein the live broadcasting device comprises: a voice recognition unit that recognizes product registration voice transmitted from the seller terminal during a live broadcast; an artificial intelligence-based analysis unit that classifies information based on the voice recognized by the voice recognition unit, automatically extracts essential information, and automatically generates product information based on the extracted essential information to perform registration and sales; and a product information providing unit that provides the product information registered by the artificial intelligence-based analysis unit to the viewer terminal or provides it on the screen. Claim 2 In claim 1, the artificial intelligence-based analysis unit comprises: a data receiving module that receives voice data recognized by the voice recognition unit; a classification module that classifies product data received by the data receiving module by word and classifies it by extracting pre-set essential items from the classified word information for generating product information; and a product information generation module that generates product information based on the essential items classified by the classification module, characterized in that it is a voice-based live product registration and automatic comment response system. Claim 3 In claim 2, the classification module is characterized by extracting at least one item among category, product name, price, stock, and option as a required item from the classified word information, in a voice-based live product registration and automatic comment response system. Claim 4 In claim 2, the voice-based live product registration and automatic comment response system is characterized in that the product information generation module performs a supplementary query to derive insufficient product information when generating product information using only the extracted essential items, if the generation of product information is insufficient with only the extracted essential items. Claim 5 A voice-based live product registration and automatic comment response system, wherein, in claim 1, the live broadcasting device includes a product information supplementation unit that requests product information supplementation by transmitting supplementary query information for generating product information to the seller terminal in conjunction with the artificial intelligence-based analysis unit. Claim 6 In claim 1, the voice-based live product registration and automatic comment response system is characterized in that the live broadcasting device includes an artificial intelligence (AI) manager that collects comments registered by a viewer terminal during a live broadcast and provides answers generated in real time through an AI-based analysis unit. Claim 7 In claim 6, the voice-based live product registration and automatic comment response system is characterized in that the artificial intelligence manager comprises: a comment collection unit that collects comments registered by a viewer terminal during a live broadcast in real time; and an answer providing unit that provides an answer generated for the collected comments. Claim 8 In claim 1, the artificial intelligence-based analysis unit comprises: a comment collection module that collects comments registered by viewers in real time during a live broadcast; a comment analysis module that analyzes the comments collected by the comment collection module in conjunction with product information, analyzes them based on pre-registered sales policies, or analyzes them based on the user's past comment patterns; and an answer generation module that generates and provides a real-time answer based on the results analyzed by the comment analysis module, characterized in that it is a voice-based live product registration and automatic comment response system.

Images