Intelligent agent system and method for short video marketing customer acquisition based on large language model

Abstract

The invention provides an intelligent agent system and method for short video marketing customer obtaining based on a large language model.The intelligent agent system comprises an AI short video planning device, an AI short video making device and an AI short video operating device, the AI short video planning device captures topics from a public media platform and performs topic selection recommendation in combination with commercial positioning of customers, and the AI short video making device makes the selected topics according to the selected topics; the customer completes confirmation of short video topic selection; the AI short video making device generates a client IP short video according to the selected question confirmed by the client, and publishes the client IP short video on a whole-network short video media platform; the AI short video operation device captures video playing data and net friend comment content, performs data analysis and summary report on video release conditions, and replies net friend comments. Through an agent application technology driven by a large language model, three key agent devices are constructed in a client, and a new employment normal form which only provides exclusive short video production service for the client for 7 * 24 hours is started.

Description

Technical Field

[0001] The present invention relates to the field of AI application technology, and in particular to an intelligent system and method for acquiring customers through short video marketing based on a large language model. Background Art

[0002] Currently, in content marketing, especially in the field of short video marketing, traditional GUI (Graphic user interface) software tools are basically used to implement the topic selection, copywriting and post-shooting editing functions of short videos. The user entry threshold is very high, so it is basically through self-built content teams or entrusted third-party operation teams to continuously produce IP (Intellectual Property) short videos for customers. After the emergence of AI (Artificial Intelligence) large language model technology, because the semantic understanding and content production capabilities have reached the expert level of output, it is a future technology trend to train and use AI agents as real corporate employees. At present, with the rise of short video marketing, more and more business owners hope to achieve short video marketing and customer acquisition through IP exposure, and the output of such short videos through natural dialogue is the closest way for bosses to assign tasks to employees. However, there are no AI-native IP short video creation products yet.

[0003] After short video platforms became the main way for Internet users to obtain content, customers paid more and more attention to how to reach their users through short video platforms and influence these users to form commercial conversions. However, the entire chain of short video production includes planning, topic selection, copywriting, shooting, editing, operation and service. For small and medium-sized enterprises that lack market content teams, each link faces the challenges of lack of knowledge system, lack of time and energy, lack of funds, and lack of team. It is very difficult to combine them together. It is basically difficult to achieve high-frequency updates every day. If you use an external operation team, you will face high operation fees, and the company cannot precipitate the methodology of short video operation. In order to enable more companies to have their own content teams more conveniently, quickly and cheaply, we have invented a new intelligent system for short video marketing and customer acquisition based on a large language model. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent system for acquiring customers through short video marketing based on a large language model, which can enable enterprises to have their own content team more conveniently, quickly and cheaply.

[0005] The objectives of the present invention can be achieved through the following technical measures: an intelligent system for short video marketing and customer acquisition based on a large language model, the intelligent system for short video marketing and customer acquisition based on a large language model includes an AI short video planning device, an AI short video production device and an AI short video operation device, the AI ​​short video planning device captures topics from public media platforms, and makes topic recommendations based on the customer's business positioning, and the customer completes the confirmation of the short video topic; the AI ​​short video production device generates a customer IP short video based on the topic confirmed by the customer, and publishes it on the short video media platform of the entire network; the AI ​​short video operation device captures video playback data and netizen comment content, makes data analysis and summary reports on the video release situation, and responds to netizen comments.

[0006] The purpose of the present invention can also be achieved by the following technical measures:

[0007] The AI ​​short video planning device includes a hot topic selection module and a hit topic selection module. The hot topic selection module obtains a large number of real-time hot spots based on the hot lists of each platform by crawling information from the entire network media platform, and then analyzes and summarizes the hot titles through a large language model to form industry topic tags. The industry topic tags are further compared and analyzed with the customer's industry and business operations, and analyzed and judged in combination with the portrait of the customer's target customer group, and precise hot topics that are in line with the customer's oral broadcast are extracted for the customer to use as a reference for topic selection; the hit topic selection module uses the customer's preset industry keywords to search for keywords on the short video platform, and sorts the hit popularity according to the time dimension, retrieves the number of hits preset by the customer, and extracts the title and content summary of the hit video for the customer to use as a reference for topic selection.

[0008] The AI ​​short video production device uses the topic confirmed by the AI ​​short video planning device, and uses the prompt word engineering technology in combination with the customer's business positioning and user positioning, as well as the oral style, to generate a video script with the customer's unique personal style. It combines the digital human reproduction technology and intelligent editing technology of the customer's exclusive image to generate the customer's IP short video and publish it on the short video media platform across the entire network.

[0009] The AI ​​short video production device includes a link parsing module, a copywriting module, a product placement module, a digital population broadcast video generation module, a short video automatic editing module, a short video poster generation module and a full-platform automatic publishing module; the link parsing module performs in-depth analysis of the link content, including the capture of text links and the analysis of video / audio ASR, through the topic selection and captured links received by the AI ​​short video planning device, so as to generate original materials;

[0010] The copy rewriting module is trained on the prompt words for the customer's business positioning, customer group positioning, and personality positioning preset by the big language model, and then combines the AIGC capability of the big model to rewrite the oral copy that matches the customer's IP.

[0011] The product placement module uses RAG technology and vector database technology to analyze the customer's knowledge base documents, and inserts the details of the customer's business or product that the customer wants to embed in the external video, forming some core materials to assist the customer in enhancing the explanation in the external short video;

[0012] The digital human voice video generation module is divided into a training phase and an inference phase. In the training phase, the digital human voice video generation module trains a video of the customer to form a digital human that is a 1:1 replica of the customer's image and voice; in the inference phase, the digital human voice video generation module will combine the output of the oral text to drive the output of the digital human's lip shape and expression to form a video file that plays the customer's oral broadcast in a 1:1 ratio.

[0013] The short video automatic editing module collects the video files output by the digital voice broadcast video module, the copywriting output by the copywriting rewriting module, and the core materials output by the product placement module, and generates a short video with subtitles and effects as a final film;

[0014] The short video poster cover generation module generates poster covers for short video platforms in three ways. The first is to add the title text generated by AI summary based on the photos uploaded by the customer. The second is to collect the first frame of the short video and add the title summary text generated by AI. The third is to generate the poster cover based on AI and match the video summary text generated by AI.

[0015] After obtaining the initial scan code authorization from the customer's account, the full-platform automatic publishing module automatically links to the customer's multi-platform account. Before each release, the video will be automatically published on multiple platforms based on the platform selected by the customer.

[0016] The AI ​​short video operation device includes a data capture module, a data analysis module and an intelligent customer response module. The data capture module uses RPA robot technology to capture the data of likes, reposts, favorites and comments on short videos posted by customer accounts on the entire video platform and transmits them back to the intelligent system for short video marketing and customer acquisition based on the large language model; the data analysis module uses the semantic understanding of the large model and the small model of NL2SQL to perform operational analysis on the collected data for customers to adjust their operational strategies; the intelligent customer response module uses the large language model to accurately screen the captured comments from netizens, then generates reply content that conforms to the customer's personality, and responds to the comments.

[0017] The purpose of the present invention can also be achieved by the following technical measures: a short video marketing customer acquisition method based on a large language model, the short video marketing customer acquisition method based on a large language model includes:

[0018] Step 1: Collect the customer’s video and audio files for digital human pre-training;

[0019] Step 2: Capture topics from public media platforms and make topic recommendations based on the client’s business positioning. The client then confirms the topic of the short video.

[0020] Step 3: Generate a short video based on the topic confirmed by the customer and publish it on the short video media platform of the whole network;

[0021] Step 4: Capture video playback data and netizen comments, make data analysis and summary reports on video releases, and respond to netizen comments.

[0022] The purpose of the present invention can also be achieved by the following technical measures:

[0023] Step 1 specifically includes:

[0024] Step 11, shooting a video of the customer speaking to the camera;

[0025] Step 12, collect the audio material of the customer and pre-train the digital human to train the timbre;

[0026] Step 13, providing the materials of steps 11 and 12 to the digital human training platform for training, so as to complete a 1:1 replica of the real person;

[0027] Step 14: Produce the content positioning of the short video based on the client’s personality, business, and user positioning, and input the prompt words into the large language model;

[0028] Step 15: Organize and upload the company's external marketing-oriented knowledge documents separately, and use the large language model to learn and interpret them to prepare for the subsequent generation of short video materials.

[0029] Step 2 specifically includes:

[0030] Step 21: The customer actively provides industry keywords through a conversation. Based on the industry keywords provided by the customer, RPA+API technology is used to search for video content on the short video platform.

[0031] Step 22: Perform ASR on the retrieved video files, extract the text of the popular videos, and use the large language model to generate titles and summaries;

[0032] Step 23: crawl the hot news titles and copywriting of the mainstream media platforms on the entire network for the customer to decide whether to use;

[0033] Step 24: The client selects topics from the topic list that are highly relevant to their products and services;

[0034] Step 25, based on the topic selected by the customer, give priority to hot videos and generate a video schedule.

[0035] Step 3 includes:

[0036] Step 31, rewrite the copy by using the text generation capability of the large language model and combining the content of the preset customer personality and business positioning prompt words;

[0037] Step 32, through RAG search enhancement generation technology, the details of relevant business or products in the customer knowledge base are found out, and the graphic materials for product implantation are formed, as well as the corresponding relationship between the pictures and the spoken texts are formed;

[0038] Step 33, reorganize the copywriting produced in steps 31 and 32 to generate a set of copywriting that meets the requirements in terms of structure and logic;

[0039] Step 34: Use the summarization and text-generated image capabilities of the large language model to help customers generate the cover image required for each short video;

[0040] Step 35: Use TTS technology to drive the trained customer digital human to make a spoken announcement video.

[0041] Step 36, based on the copy generated in step 33 and the oral video generated in step 35, a video with both IP image and core material is formed;

[0042] Step 37: The large language model automatically completes sentence segmentation and pauses based on the semantic understanding of the text, calls the default or customer-preset subtitle style, generates subtitles, and highlights some key subtitles;

[0043] Step 38, generating the title, topic and cover of the video to be published according to the copy;

[0044] Step 39, obtain the publishing permission based on the platform authorized by the customer. When each video work is released, the customer checks the platform where he or she publishes it, and publishes the short video to the global mainstream short video platform through interface calls.

[0045] Step 4 includes:

[0046] Step 41: Use RPA robot technology to capture the data of likes, reposts, favorites, and comments on the short videos posted by the customer account and transmit the data back;

[0047] Step 42, after the data is transmitted back, the structured data is organized in a preset database, and the columns and fields such as the video work name, release time, number of views, number of likes, number of reposts, number of favorites, number of comments, and three-second completion rate are formed in the database, and the data is updated;

[0048] Step 43: Through semantic understanding of the large language model and the support of the NL2SQL small model, the data report is created in the dialogue interface;

[0049] Step 44, through the semantic understanding and text generation capabilities of the large language model, private domain conversion is performed on users in the comments who may be converted into leads, and the reply to the user's comment is completed.

[0050] The intelligent agent system for short video marketing and customer acquisition based on the large language model in the present invention, through the intelligent agent application technology driven by the large language model, helps customers build three key intelligent agent devices of the content production team, and opens a new employment paradigm that provides exclusive short video production services to customers 24 hours a day, 7 days a week. Through the application development of three intelligent agent devices, the present invention replaces the editor, director, shooter and operator of the actual enterprise market short video team, and realizes the use of short video creation with zero threshold for customers. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 This is a structural diagram of a specific embodiment of the intelligent system for acquiring customers through short video marketing based on a large language model of the present invention;

[0052] Figure 2 It is a flowchart of a specific embodiment of the short video marketing customer acquisition method based on a large language model of the present invention;

[0053] Figure 3 This is a flow chart of initializing an intelligent system for acquiring customers through short video marketing based on a large language model in a specific embodiment of the present invention;

[0054] Figure 4 This is a workflow diagram of an AI short video planning device in a specific embodiment of the present invention;

[0055] Figure 5 A flowchart of copy material generation of an AI short video production device in a specific embodiment of the present invention;

[0056] Figure 6 A video generation and publishing flow chart of an AI short video production device in a specific embodiment of the present invention;

[0057] Figure 7 This is a workflow diagram of an AI short video operation device in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0058] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0059] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations and / or combinations thereof.

[0060] Figure 1 This is a module diagram of the intelligent system for short video marketing and customer acquisition based on a large language model of the present invention, including an AI short video planning device 10, an AI short video production device 20, and an AI short video operation device 30.

[0061] The AI ​​short video planning device 10 captures topics in real time on public media platforms such as Douyin, Baidu, and Weibo through popular topics and market hot topics in the customer's industry, and makes topic recommendations based on the customer's business positioning. The customer completes the confirmation of the short video topic for the day or the week.

[0062] The AI ​​short video production device 20 uses the topic confirmed by the AI ​​short video planning device, and uses the prompt word engineering technology to combine the customer's business positioning and user positioning, as well as the oral style, to generate a video script with the customer's unique personality style, and combines the digital human reproduction technology and intelligent editing technology of the customer's exclusive image to generate a complete customer IP short video, and automatically publish it on the short video media platform of the entire network.

[0063] The AI ​​short video operation device 30 captures video playback data and netizen comments by combining RPA (Robotic Process Automation) and API (Application Programming Interface) for videos released on the entire network, performs data analysis and summary reports on daily / weekly video releases in a conversational manner, and automatically responds to netizen comments.

[0064] The AI ​​short video planning device 10 includes a hot topic selection module 11 and a hot topic selection module 12.

[0065] The hot topic selection module 11 obtains the hot lists of each platform for half an hour / 15 minutes (such as Baidu hot search, Douyin hot list, Weibo hot list, etc.) by crawling information from all media platforms on the entire network, and obtains a large number of real-time hot spots. Then, the hot topic titles are analyzed and summarized through a large language model to form industry topic tags; these tags will be further compared and analyzed with the customer's industry and business operations, and analyzed and judged in combination with the portrait of the customer's target customer group, to extract accurate hot topics that are in line with the customer's oral broadcast, and provide AI short video production devices with video generation.

[0066] The hot-selling topic selection module 12 searches for keywords on the short video platform through the industry keywords preset by the customer, and sorts the popularity of hot-selling videos according to the time dimension (7 days, 30 days, 90 days, 180 days, etc.), retrieves the number of hot-selling videos preset by the customer (for example, 2-10), and extracts the titles and content summaries of the hot-selling videos for the customer's reference in topic selection.

[0067] The AI ​​short video production device includes seven modules: a link parsing module 21, a text rewriting module 22, a product implantation module 23, a digital voice broadcast video generation module 24, a short video automatic editing module 25, a short video poster generation module 26 and a full-platform automatic publishing module 27.

[0068] The link analysis module 21 receives the selected topics and captured links from the AI ​​short video planning device in a timely manner, and performs in-depth analysis of the link content, including the capture of text links and ASR (audio to text analysis) analysis of video / audio links, so as to generate original materials.

[0069] After training the prompt words of the customer's business positioning (business, business model), customer positioning (customer portrait) and personality positioning (personality, speaking style) preset by the large language model, the copy rewriting module 22 can combine the AIGC (Artificial Intelligence Generated Content) capability of the large model to rewrite the oral copy that conforms to the customer's IP.

[0070] The product implantation module 23 refers to the use of RAG (retrieval enhanced generation) technology and vector database technology to parse the company's knowledge base documents (pdf, jpg, word, ppt, etc.), and to embed the details of the company's business or products (such as highlights, cases, brands, comparison charts) in external videos, so as to form some core materials to assist customers in enhancing explanations in short videos released to the outside world. For example, when talking about the topic of the comparison of the number of immigrants, a trend chart of the number of immigrants in the past 10 years in the company's knowledge base and the number of people of various identities can be used to present details in the lower middle part of the video.

[0071] The digital human voice broadcast video generation module 24 is divided into a training phase and an inference phase. In the training phase, only a 1-3 minute video of the customer is needed for training to form a digital human that is a 1:1 replica of the customer's image and voice. In the inference phase, the digital human voice broadcast module will combine the output of the oral broadcast copy to drive the output of the digital human's lip shape and expression to form a video file that plays the customer's oral broadcast in a 1:1 ratio.

[0072] The short video automatic editing module 25 can automatically collect the video output of the digital voice broadcast video module, the text output of the text rewriting module and the output of the patch material of the product placement module, and can automatically generate a short video with subtitles and effects as a final film.

[0073] The short video poster cover generation module 26 can generate poster covers for short video platforms in three ways. The first is to automatically add title text generated by AI summary based on the photos uploaded by the customer. The second is to automatically collect the first frame of the short video and add the title summary text generated by AI. The third is to generate the poster cover completely based on AI and match it with the video summary text generated by AI.

[0074] The full-platform automatic publishing module 27 can automatically link the customer's multi-platform accounts (Tik Tok, Kuaishou, Video Account, Xiaohongshu, etc.) after obtaining the initial scan code authorization of the customer's account. Before each release, the video will be automatically published on multiple platforms according to the platform selected by the customer, and the customer will not be aware of the entire publishing process.

[0075] The AI ​​short video operation device 30 is divided into a data capture module 31, a data analysis module 32 and an intelligent customer response module 33.

[0076] The data capture module 31 uses RPA (Robotic Process Automation) robot technology to capture the likes, reposts, favorites and comments of short videos posted by customer accounts on the entire video platform and returns them to the system.

[0077] The data analysis module 32 uses the semantic understanding of the large model and the small model of NL2SQL to perform operational analysis on the collected data for customers to adjust their operational strategies.

[0078] The intelligent customer response module 33 uses a large language model to accurately screen the captured comments, then generates response content that matches the customer's personality, and automatically responds to comments to guide private domain conversions.

[0079] Figure 2 The flowchart of the short video marketing customer acquisition method based on the large language model of the present invention. The short video marketing customer acquisition method based on the large language model specifically includes:

[0080] Step 101, collect the customer's video and audio files for digital human pre-training; after providing positioning consulting services to the customer, generate the word prompts of the large language model. Make the system have the customer's unique style, and collect the customer's video / audio files for digital human pre-training.

[0081] Step 102, capture topics from public media platforms, and make topic recommendations based on the customer's business positioning, and the customer completes the confirmation of the short video topic; the AI ​​short video planning device searches and analyzes hot events / explosive videos on the entire network, reads the content, and combines the customer's preset business positioning, personality and user positioning to give a short video topic plan through the AI ​​large language model.

[0082] Step 103: Generate a client IP short video based on the topic selected by the client and publish it on the short video media platform of the entire network;

[0083] The AI ​​short video production device obtains the topic to be produced by asking the planning specialist, and then rewrites the copy based on the topic, while implanting the company's products / cases / services to generate the final spoken copy, patch materials and video cover image.

[0084] The AI ​​short video production device drives the digital human to generate spoken videos, and automatically inserts corporate knowledge as materials through AI editing, and then publishes them to short video platforms across the entire network according to customer needs.

[0085] Step 104, capture video playback data and netizen comments, make data analysis and summary report on video release, and reply to netizen comments.

[0086] The I short video operation device automatically tracks the video playback data (likes, favorites, reposts, comments) of the entire platform and presents the data in the form of data reports. It can also combine the design of the company's short video commercial monetization path to automatically reply to customers and make the next step of business conversion.

[0087] Before the system is put into use, the system initialization work must be completed through consulting services, such as Figure 3 As shown, it mainly includes two parts: customer digital human training and customer content positioning input.

[0088] Step S1011, shoot a 1-3 minute video of the customer in a certain scene. For example, according to the theory that people, goods and places are consistent in short videos, you should choose a scene that is suitable for the customer to sell goods and shoot a 1-3 minute video. For example, for immigration agencies, the best scene is to shoot a 1-3 minute video in their own office or conference room with neatly arranged bookcases and archives behind, and the customer faces the camera to make a 1-3 minute video.

[0089] In step S1012, in order to achieve better sound restoration, 1-3 minutes of audio material from the customer can be collected separately to pre-train the digital human, mainly training the timbre.

[0090] In step S1013, the materials of the previous two steps are uniformly provided to the digital human training platform for training to complete a 1:1 replica of a real person.

[0091] In step S1014, the content positioning of the short video is generated based on the personality, business and user positioning completed face-to-face with customers offline, and these inputs are given to the prompt words of the large language model, so that the three AI devices can better tailor their services to customers.

[0092] In step S1015, in order to better integrate with the client’s own business, we also need to separately organize and upload the client’s external marketing-oriented knowledge documents, and use the large language model to learn and interpret them, so as to prepare for the subsequent generation of short video materials.

[0093] The AI ​​short video planning device includes two system modules: hot topic selection and hot topic selection. Its workflow is as follows: Figure 4 shown.

[0094] S1021, the user actively outputs an industry keyword to the AI ​​short video planning device through a dialogue. For example, if the industry keyword is immigration, the requirement is to find the popular videos with the keyword immigration that have the highest number of likes in the past 180 days on Douyin. The planning assistant will activate the search capability, use RPA+API technology to search for video content that meets the requirements on the short video platform, and retrieve 10 popular videos.

[0095] In step S1022, the system will perform ASR (Automatic Speech Recognition) on the retrieved video files, extract the texts of the 10 most popular videos about emigration in the past 180 days, and use the large language model capabilities to generate titles and summaries.

[0096] In step S1023, the system will automatically crawl the hot news headlines and copywriting of the mainstream media platforms on the entire network, and after collection, it will label the large model and generate recommendation scores. For example, education and investment are relatively relevant, and similar news headlines and content will be recommended to the topic list for customers to decide whether to use them.

[0097] In step S1024, customers can use topics in the topic list that are highly relevant to their products / services. For example, if there is a recommendation for university education in the topic list, it is a good opportunity to implant investment products. In this case, you only need to mark the relationship between the topic and the investment product.

[0098] In step S1025, the AI ​​short video planning device will give priority to hot videos, and then plan the popular videos, and pass the information to the AI ​​short video production device.

[0099] The short video AI short video production device includes seven system modules: link analysis, copywriting, product placement, digital voice broadcast video generation, short video automatic editing, short video poster generation module and full platform automatic publishing. Its workflow is as follows: Figure 5 and Figure 6 shown.

[0100] In step S1031, through data communication between intelligent entities, the AI ​​short video production device already knows that a hot topic of policy changes will be released at 9 o'clock this morning, and goes online to start work on time at 9 o'clock.

[0101] In step S1032, the copy rewriting module uses the text generation capability of the large language model and combines it with the content of the preset customer personality and business positioning prompts to automatically rewrite the copy. For example, it will emphasize that the author is an immigration expert with 10 years of service in the industry and has never failed in any case, and the language style will be close to the customer's personality.

[0102] In step S1033, the product placement module uses RAG search enhancement generation technology to find all the cases and service highlights related to investment immigration in the enterprise knowledge base, and forms graphic materials for product placement, as well as the corresponding relationship between pictures and spoken text. For example, when our investment immigration helped the boss of Enterprise A get his identity card in as fast as 50 days, it corresponded to a photo of the boss of Enterprise A thanking the company after he got his identity card.

[0103] In step S1034, the system will reorganize the documents produced in the above two steps to generate a set of documents that meet the structural logic requirements.

[0104] In step S1035, because all short video platforms require the upload of a cover image when publishing a video, we use the summarization and text-based image capabilities of the large language model to help customers generate the cover image required for each short video. For example, the base image is still an image of the IP of the owner of an immigration agency, and the title above reads how investment immigrants can obtain status within 2 months.

[0105] In step S1041, using the digital human voice broadcast video generation module, all oral texts will be driven by TTS technology (Text To Speech) to drive the trained customer digital human to make oral broadcast videos. For example, if there is a text of about one minute with 300 words generated, then the digital human can broadcast this one-minute oral video.

[0106] In step S1042, because the customer wants to complete more effective information transmission in the short video, the text materials of the product and service categories generated in S1033 above will be uniformly loaded into the oral video generated by S1041 through the short video automatic editing module at this time, forming a video with an IP image and material patches.

[0107] In step S1043, the large language model will automatically complete sentence breaks and pauses based on the semantic understanding of the text, call the system default or customer preset subtitle style, generate subtitles, and highlight some key subtitles, such as the three pitfalls of investment immigration to emphasize such information.

[0108] In step S1044, the short video poster generation module will automatically generate the title, topic and cover of the video according to the copy, and pass the completed cover to the full platform automatic publishing module. For example, the cover will be equipped with the client's IP image photo and the large words "Immigration law has changed dramatically" and the small words "You can get investment immigration in 30 days".

[0109] In step S1045, the full-platform automatic publishing module will obtain publishing permissions based on the platforms authorized by the customer. When each video work is released, the customer only needs to check the platform where he or she publishes it (such as TikTok, Video Account, and Kuaishou), and publish the short video to the world's mainstream short video platforms through interface calls.

[0110] The short video AI short video operation device includes three system modules: data capture module, data analysis module and intelligent customer response module. Its workflow is as follows: Figure 7 shown.

[0111] In step S1051, the data capture module uses RPA robot technology to simulate the clicks of real users on various platforms. For example, it first retrieves the customer's own video account, then captures the details of each work in the creator center, clicks on the details of each work one by one, and sees the playback data (key indicators such as playback, forwarding, collection, likes, and completion rate). It can also capture the comment data of real users' messages in the user comment area and send it back in the format of json (JavaScript Object Notation) data.

[0112] In step S1052, after the data is transmitted back, the system organizes the structured data in a preset database, forming columns such as the video work name, release time, number of views, number of likes, number of reposts, number of favorites, number of comments, and three-second completion rate in the database, and completes 15 minutes of data update.

[0113] In step S1053, through the semantic understanding of the large language model and the support of the NL2SQL small model, data reports can be created on the dialogue interface. For example, if a customer asks which of my videos brought more leads in the past seven days, the data analysis module will automatically conduct a comprehensive analysis of the data of the 10 videos released within 7 days and tell the customer that the policy released on the third night generated more than 30 leads, which is much higher than other videos. This gives customers a clearer signal that they should spread the message based on hot spots.

[0114] In step S1054, through the semantic understanding and text generation capabilities of the large language model, we can automatically conduct private domain conversion for users in the comments who are likely to be converted into leads. For example, we can automatically reply to customers and add XX WeChat account to obtain the top ten guides and top ten cases for avoiding investment immigration pitfalls, and then guide customers to convert to the private domain to realize lead business opportunities.

[0115] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions recorded in the aforementioned embodiments or replace some of the technical features therein with equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0116] Except for the technical features described in the specification, all other technical features are known technologies to those skilled in the art.

Claims

1. An intelligent system for acquiring customers through short video marketing based on a large language model, characterized by: The intelligent system for short video marketing and customer acquisition based on a large language model includes an AI short video planning device, an AI short video production device and an AI short video operation device. The AI ​​short video planning device captures topics from public media platforms, and makes topic recommendations based on the customer's business positioning, and the customer completes the confirmation of the short video topic; the AI ​​short video production device generates the customer IP short video based on the topic confirmed by the customer, and publishes it on the short video media platform of the entire network; the AI ​​short video operation device captures video playback data and netizens' comments, makes data analysis and summary reports on the video release situation, and responds to netizens' comments.

2. The intelligent agent system for acquiring customers through short video marketing based on a large language model according to claim 1 is characterized in that: The AI ​​short video planning device includes a hot topic selection module and a hit topic selection module. The hot topic selection module obtains a large number of real-time hot spots based on the hot lists of each platform by crawling information from the entire network media platform, and then analyzes and summarizes the hot titles through a large language model to form industry topic tags. The industry topic tags are further compared and analyzed with the customer's industry and business operations, and analyzed and judged in combination with the portrait of the customer's target customer group, and precise hot topics that are in line with the customer's oral broadcast are extracted for the customer to use as a reference for topic selection; the hit topic selection module uses the customer's preset industry keywords to search for keywords on the short video platform, and sorts the hit popularity according to the time dimension, retrieves the number of hits preset by the customer, and extracts the title and content summary of the hit video for the customer to use as a reference for topic selection.

3. The intelligent agent system for acquiring customers through short video marketing based on a large language model according to claim 1 is characterized in that: The AI ​​short video production device uses the topic confirmed by the AI ​​short video planning device, and uses the prompt word engineering technology in combination with the customer's business positioning and user positioning, as well as the oral style, to generate a video script with the customer's unique personal style. It combines the digital human reproduction technology and intelligent editing technology of the customer's exclusive image to generate the customer's IP short video and publish it on the short video media platform across the entire network.

4. The intelligent agent system for acquiring customers through short video marketing based on a large language model according to claim 3 is characterized in that: The AI ​​short video production device includes a link parsing module, a copywriting module, a product placement module, a digital population broadcast video generation module, a short video automatic editing module, a short video poster generation module and a full-platform automatic publishing module; the link parsing module performs in-depth analysis of the link content, including the capture of text links and the analysis of video / audio ASR, through the topic selection and captured links received by the AI ​​short video planning device, so as to generate original materials; The copy rewriting module is trained on the prompt words for the customer's business positioning, customer group positioning, and personality positioning preset by the big language model, and then combines the AIGC capability of the big model to rewrite the oral copy that matches the customer's IP. The product placement module uses RAG technology and vector database technology to analyze the customer's knowledge base documents, and inserts the details of the customer's business or product that the customer wants to embed in the external video, forming some core materials to assist the customer in enhancing the explanation in the external short video; The digital person broadcasting video generation module is divided into a training phase and an inference phase. In the training phase, the digital person broadcasting video generation module trains a video of the customer to form a digital person that is a 1:1 replica of the customer's image and voice. In the inference stage, the digital human voice-over video generation module will combine the output of the voice-over copy to drive the output of the digital human's lip shape and expression, forming a video file that plays the voice-over in a 1:1 ratio with the customer; The short video automatic editing module collects the video files output by the digital voice broadcast video module, the copywriting output by the copywriting rewriting module, and the core materials output by the product placement module, and generates a short video with subtitles and effects as a final film; The short video poster cover generation module generates poster covers for short video platforms in three ways. The first is to add the title text generated by AI summary based on the photos uploaded by the customer. The second is to collect the first frame of the short video and add the title summary text generated by AI. The third is to generate the poster cover based on AI and match the video summary text generated by AI. After obtaining the initial scan code authorization from the customer's account, the full-platform automatic publishing module automatically links to the customer's multi-platform account. Before each release, the video will be automatically published on multiple platforms based on the platform selected by the customer.

5. The intelligent agent system for acquiring customers through short video marketing based on a large language model according to claim 1, characterized in that: The AI ​​short video operation device includes a data capture module, a data analysis module and an intelligent customer response module. The data capture module uses RPA robot technology to capture the data of likes, reposts, favorites and comments on short videos posted by customer accounts on the entire video platform and transmits them back to the intelligent system for short video marketing and customer acquisition based on the large language model; the data analysis module uses the semantic understanding of the large model and the small model of NL2SQL to perform operational analysis on the collected data for customers to adjust their operational strategies; the intelligent customer response module uses the large language model to accurately screen the captured comments from netizens, then generates reply content that conforms to the customer's personality, and responds to the comments.

6. A short video marketing customer acquisition method based on a large language model, characterized in that: The short video marketing customer acquisition method based on the large language model includes: Step 1: Collect the customer’s video and audio files for digital human pre-training; Step 2: Capture topics from public media platforms and make topic recommendations based on the client’s business positioning. The client then confirms the topic of the short video. Step 3: Generate a short video based on the topic confirmed by the customer and publish it on the short video media platform of the whole network; Step 4: Capture video playback data and netizen comments, make data analysis and summary reports on video releases, and respond to netizen comments.

7. The short video marketing customer acquisition method based on a large language model according to claim 6 is characterized in that: Step 1 specifically includes: Step 11, shooting a video of the customer speaking to the camera; Step 12, collect the audio material of the customer and pre-train the digital human to train the timbre; Step 13, providing the materials of steps 11 and 12 to the digital human training platform for training, so as to complete a 1:1 replica of the real person; Step 14: Produce the content positioning of the short video based on the client’s personality, business, and user positioning, and input the prompt words into the large language model; Step 15: Organize and upload the company's external marketing-oriented knowledge documents separately, and use the large language model to learn and interpret them to prepare for the subsequent generation of short video materials.

8. The short video marketing customer acquisition method based on a large language model according to claim 6 is characterized in that: Step 2 specifically includes: Step 21: The customer actively provides industry keywords through a conversation. Based on the industry keywords provided by the customer, RPA+API technology is used to search for video content on the short video platform. Step 22: Perform ASR on the retrieved video files, extract the text of the popular videos, and use the large language model to generate titles and summaries; Step 23: crawl the hot news titles and copywriting of the mainstream media platforms on the entire network for the customer to decide whether to use; Step 24: The client selects topics from the topic list that are highly relevant to their products and services; Step 25, based on the topic selected by the customer, give priority to hot videos and generate a video schedule.

9. The short video marketing customer acquisition method based on a large language model according to claim 6, characterized in that: Step 3 includes: Step 31, rewrite the copy by using the text generation capability of the large language model and combining the content of the preset customer personality and business positioning prompt words; Step 32, through RAG search enhancement generation technology, the details of relevant business or products in the customer knowledge base are found out, and the graphic materials for product implantation are formed, as well as the corresponding relationship between the pictures and the spoken texts are formed; Step 33, reorganize the copywriting produced in steps 31 and 32 to generate a set of copywriting that meets the requirements in terms of structure and logic; Step 34: Use the summarization and text-generated image capabilities of the large language model to help customers generate the cover image required for each short video; Step 35: Use TTS technology to drive the trained customer digital human to make a spoken announcement video. Step 36, based on the copy generated in step 33 and the oral video generated in step 35, a video with both IP image and core material is formed; Step 37: The large language model automatically completes sentence segmentation and pauses based on the semantic understanding of the text, calls the default or customer-preset subtitle style, generates subtitles, and highlights some key subtitles; Step 38, generating the title, topic and cover of the video to be published according to the copy; Step 39, obtain the publishing permission based on the platform authorized by the customer. When each video work is released, the customer checks the platform where he or she publishes it, and publishes the short video to the global mainstream short video platform through interface calls.

10. The short video marketing customer acquisition method based on a large language model according to claim 6, characterized in that: Step 4 includes: Step 41: Use RPA robot technology to capture the data of likes, reposts, favorites, and comments on the short videos posted by the customer account and transmit the data back; Step 42, after the data is transmitted back, the structured data is organized in a preset database, and the columns and fields such as the video work name, release time, number of views, number of likes, number of reposts, number of favorites, number of comments, and three-second completion rate are formed in the database, and the data is updated; Step 43: Through semantic understanding of the large language model and the support of the NL2SQL small model, the data report is created in the dialogue interface; Step 44, through the semantic understanding and text generation capabilities of the large language model, private domain conversion is performed on users in the comments who may be converted into leads, and the reply to the user's comment is completed.

Images