Automatic live broadcast method and system applied to Internet platform

Abstract

The invention relates to an automatic live broadcast method and system applied to an internet platform, and the method comprises the steps: creating a live broadcast task, generating a complete verbal skill draft of the live broadcast in a combined manner, carrying out the cutting and segmentation, controlling an intelligent mobile robot to walk to a live broadcast shooting point of a first explanation commodity, and then starting the live broadcast; performing live explanation on the cut and segmented live verbal skills in sequence, generating a new live verbal skill according to a live verbal skill strategy and bullet screen area information acquired in real time, and after playing of one cut and segment of a currently played live verbal skill draft is completed, inserting the current segment of strategy or interactive verbal skill for sound playing; and judging whether the current live broadcast has a commodity to be explained by the live broadcast, if so, controlling the robot to move to the live broadcast shooting point of the next explained commodity, otherwise, closing the live broadcast after playing the live broadcast. According to the invention, live broadcast operation capability and marketing capability can be continuously improved through upgrading of a software platform and a system, high-load live broadcast tasks can be realized at the same time, artificial anchors are completely replaced, and cost reduction and efficiency improvement are realized for enterprises.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to an automatic live broadcast method and system applied to an Internet platform. Background Art

[0002] Now all industries are live-streaming on various Internet platforms (TikTok, Xiaohongshu, Kuaishou, JD.com and Meituan, etc.), mainly implementing the following two marketing methods through company introduction, product display, preferential benefits explanation, etc. One is to attract customers with purchasing intention to leave their phone and WeChat contact information on the platform, and the company will contact them offline to invite them to the company's sales point for further interviews to promote transactions; the other is to directly facilitate customers to place orders on the platform. Live broadcasts are divided into live broadcasts with the host going abroad and live broadcasts without going abroad, but both require at least one anchor. During the live broadcast, they walk to different displayed products, explain the products, publicize preferential policies, and respond to user inquiries, etc., to attract users to leave their information or place orders. In order to have better live broadcast traffic and customer conversion effects in the live broadcast room, the anchor needs to have certain live broadcast operation capabilities, certain marketing experience, and rich knowledge of companies and products. In order for the company to have more Internet exposure, the company also needs to conduct live broadcasts frequently (such as multiple live broadcasts per day), which requires more anchors or entrusts them to professional third-party organizations, which adds certain costs to the company.

[0003] However, the current live broadcasting method has the following problems: 1. The high turnover of anchors will not only bring personnel management burden to the company, but the replacement of anchors will also affect the labels and exposure of the live broadcast rooms on the Internet platform; 2. The levels of anchors are uneven, and companies need to invest in training and guidance for anchors on corporate knowledge, product knowledge, live broadcasting operation capabilities, etc.; 3. Staff energy is limited, and it is impossible to broadcast live for a long time or high frequency every day, and live broadcasts on various platforms cannot bring more performance growth to the company. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide an automatic live broadcast method and system applied to an Internet platform, thereby solving the deficiencies of the prior art.

[0005] The object of the present invention is achieved by the following technical solution: an automatic live broadcast method applied to an Internet platform, the automatic live broadcast method comprising:

[0006] Step 1: Create a live broadcast task based on the pre-set and collected data, and after detecting the live broadcast task, generate a complete script for the live broadcast according to the requirements of the live broadcast task;

[0007] Step 2: Cut the complete script into segments, control the intelligent mobile robot to go to the first live broadcast shooting point to explain the product, and then start the live broadcast;

[0008] Step 3: Live broadcast the segmented live broadcast scripts one by one, and generate new live broadcast scripts according to the live broadcast script strategy and the real-time acquired bullet screen area information. After a segment of the currently playing live broadcast script is played, insert the audio of this segment of live broadcast script to play;

[0009] Step 4: Determine whether there are other products to be explained in this live broadcast. If so, control the intelligent mobile robot to move to the live shooting point of the next product to be explained. If not, play the next broadcast and then close this live broadcast.

[0010] The step 1 specifically includes:

[0011] Create enterprise knowledge base and product knowledge base in software platforms and systems. The product knowledge base includes product name, product introduction, configuration rules, product advantages, product price, product benefits and competitive product comparison, etc.

[0012] Select the gender, personality style, and timbre of the virtual voice host provided by the system in the software platform or system, or use the third-party large model voice technology to replicate the voice based on the provided voice sample;

[0013] The intelligent mobile robot's built-in control system is used to collect and create walking maps of the company's live broadcast locations;

[0014] The control system of the intelligent mobile robot is used to calibrate the live shooting points of different products or corporate display content in the enterprise live broadcast venue, and the point information is synchronously saved in the software platform and system;

[0015] Create live broadcast tasks on the Internet platform in the software platform and system, set the live broadcast theme, live broadcast cover, live broadcast cycle, live broadcast time, product explanation and virtual voice anchor who will not leave the country;

[0016] After the software platform and system detects a new live broadcast task, it will provide the enterprise knowledge base, live broadcast product knowledge base and virtual voice anchor personality style in advance according to the requirements of the live broadcast task, and combine them into prompt words to generate a complete script for this live broadcast through the third-party LLM large language model.

[0017] The step 2 specifically includes:

[0018] The software platform and system cut and segment the generated live broadcast script;

[0019] The software platform and system control the intelligent mobile robot to move to the first live broadcast shooting point for explaining the product through the intelligent mobile robot control system interface;

[0020] When the live broadcast time of the live broadcast task arrives, the software platform and system use RPA technology to start the live broadcast on the live broadcast tools of various Internet platforms on the computer or the APPs of various Internet platforms on the mobile phone.

[0021] The step three specifically includes:

[0022] The software platform and system will sequentially cut the live broadcast into segments according to the gender and voice of the virtual anchor selected for the live broadcast, and call the third-party speech synthesis technology to play the sound and conduct the live broadcast explanation;

[0023] During the live broadcast, the software platform and system periodically generate a new live broadcast script through the third-party LLM large language model based on the live broadcast room data and the live broadcast script strategy within the software platform and system. After a segmented script in the live broadcast script that is being played is played, the third-party speech synthesis technology is called to insert the sound of the current live broadcast script.

[0024] During the live broadcast, the software platform and system obtain questions from the audience in the barrage area in real time, and use intent recognition technology or keyword matching to generate an interactive response script through a third-party LLM large language model. After a cut segment of the live broadcast script is played, the third-party speech synthesis technology is called to insert the sound of this interactive response script.

[0025] An automatic live broadcast system applied to an Internet platform, the system comprising a data collection and speech generation module, a speech cutting and live broadcast start module, a live broadcast explanation and speech playback module and a judgment module;

[0026] The data collection and speech generation module is configured to create a live broadcast task according to the pre-set and collected data, and after detecting the live broadcast task, generate a complete speech script for the live broadcast according to the requirements of the live broadcast task;

[0027] The script cutting and live broadcast start module is configured to cut and segment the complete script, and control the intelligent mobile robot to go to the first live broadcast shooting point for explaining the product and then start the live broadcast;

[0028] The live explanation and script playback module is configured to sequentially explain the segmented live scripts, and generate new live scripts according to the live script strategy and the real-time acquired bullet screen area information, and insert the audio playback of the current live script after a segment of the currently playing live script is played;

[0029] The judgment module is configured to judge whether there are other products to be explained in the live broadcast. If so, the intelligent mobile robot is controlled to move to the next live broadcast shooting point for explaining the product. If not, the live broadcast is closed after playing the next broadcast speech.

[0030] The data collection and speech generation module specifically includes the following contents:

[0031] Create enterprise knowledge base and product knowledge base in software platforms and systems. The product knowledge base includes product name, product introduction, configuration rules, product advantages, product price, product benefits and competitive product comparison;

[0032] Select the gender, personality style, and timbre of the virtual voice host provided by the system in the software platform or system, or use the third-party large model voice technology to replicate the voice based on the provided voice sample;

[0033] The intelligent mobile robot's built-in control system is used to collect and create walking maps of the company's live broadcast locations;

[0034] The control system of the intelligent mobile robot is used to calibrate the live shooting points of different products or corporate display content in the enterprise live broadcast venue, and the point information is synchronously saved in the software platform and system;

[0035] Create live broadcast tasks on the Internet platform in the software platform and system, set the live broadcast theme, live broadcast cover, live broadcast cycle, live broadcast time, product explanation and virtual voice anchor who will not leave the country, etc.;

[0036] After the software platform and system detects a new live broadcast task, it will provide the enterprise knowledge base, live broadcast product knowledge base and virtual voice anchor personality style in advance according to the requirements of the live broadcast task, and combine them into prompt words to generate a complete script for this live broadcast through the third-party LLM large language model.

[0037] The speech cutting and live broadcast start module specifically includes the following contents:

[0038] The software platform and system cut and segment the generated live broadcast script;

[0039] The software platform and system control the intelligent mobile robot to move to the first live broadcast shooting point for explaining the product through the intelligent mobile robot control system interface;

[0040] When the live broadcast time of the live broadcast task arrives, the software platform and system use RPA technology to start the live broadcast on the live broadcast tools of various Internet platforms on the computer or the APPs of various Internet platforms on the mobile phone.

[0041] The live explanation and speech playing module specifically includes the following contents:

[0042] The software platform and system will sequentially cut the live broadcast into segments according to the gender and voice of the virtual anchor selected for the live broadcast, and call the third-party speech synthesis technology to play the sound and conduct the live broadcast explanation;

[0043] During the live broadcast, the software platform and system periodically generate a new live broadcast script through the third-party LLM large language model based on the live broadcast room data and the live broadcast script strategy within the software platform and system. After a segmented script in the live broadcast script that is being played is played, the third-party speech synthesis technology is called to insert the sound of the current live broadcast script.

[0044] During the live broadcast, the software platform and system obtain questions from the audience in the barrage area in real time, and use intent recognition technology or keyword matching to generate an interactive response script through a third-party LLM large language model. After a cut segment of the live broadcast script is played, the third-party speech synthesis technology is called to insert the sound of this interactive response script.

[0045] The present invention has the following advantages: an automatic live broadcast method and system applied to an Internet platform, which can continuously improve the live broadcast operation ability and marketing ability through the upgrade of the software platform and system, and at the same time can realize high-load live broadcast tasks, completely replace manual anchors, and reduce costs and increase efficiency for enterprises. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the detailed description of the embodiments of the present application provided below in conjunction with the drawings is not intended to limit the scope of protection of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application. The present invention is further described below in conjunction with the drawings.

[0048] like Figure 1As shown, the present invention specifically relates to an automatic live broadcast method applied to an Internet platform, which realizes the automatic live broadcast of enterprises on various Internet platforms (such as new media: Douyin, Video Number, Xiaohongshu, Kuaishou, etc., e-commerce platforms: Taobao, JD.com, Tmall, Pinduoduo, Meituan, etc.) through the setting and control of software platforms and systems; it can continuously improve the live broadcast operation ability and marketing ability through the upgrade of software platforms and systems, and at the same time can realize high-load live broadcast tasks, completely replace manual anchors, and reduce costs and increase efficiency for enterprises. Specifically including the following contents:

[0049] S1. Create enterprise knowledge base and product knowledge base in the software platform and system. The product knowledge base includes product name, product introduction, configuration rules, product advantages, product price, product benefits and competitive product comparison;

[0050] S2. Select the gender, personality style (such as expert, humorous, sales elite, etc.), and timbre (such as gentle, sweet, friendly, sunny, etc.) of the virtual voice anchor provided by the system in the software platform or system, or use the third-party large model voice technology to replicate the voice based on the provided voice sample;

[0051] S3. Use the control system of the intelligent mobile robot to collect and create a walking map of the enterprise live broadcast venue (it can also be implemented in the software platform and system through the intelligent mobile robot interface);

[0052] Among them, the intelligent mobile robot is a robot system with mobile functions composed of sensors, remote control operators and automatic controllers.

[0053] S4. Use the control system of the intelligent mobile robot to calibrate the live shooting points of different products or enterprise display content in the enterprise live broadcast venue, and save the point information synchronously to the software platform and system;

[0054] S5. Create a live broadcast task on the Internet platform in the software platform and system, set the live broadcast theme, live broadcast cover, live broadcast cycle, live broadcast time, product explanation and virtual voice anchor who will not leave the country, etc.;

[0055] S6. After the software platform and system detect a new live broadcast task, they will provide the enterprise knowledge base, live broadcast product knowledge base and virtual voice anchor personality style in advance according to the requirements of the live broadcast task, and combine them into prompt words to generate a complete script for the live broadcast through the third-party LLM large language model.

[0056] Among them, the LLM large language model, also known as the large language model, is an artificial intelligence model designed to understand and generate human language.

[0057] S7. The software platform and system divide the generated live broadcast script into segments;

[0058] S8, the software platform, and the system control the intelligent mobile robot to move to the first live broadcast shooting point for explaining the product through the intelligent mobile robot control system interface;

[0059] S9. When the live broadcast time of the live broadcast task arrives, the software platform and system use RPA technology to start the live broadcast on the live broadcast tools of various Internet platforms on the computer or the APPs of various Internet platforms on the mobile phone.

[0060] S10, the software platform and the system sequentially play the live broadcast scripts cut into segments according to the gender and timbre of the virtual anchor selected for the live broadcast, and call the third-party speech synthesis technology to realize the sound playback and conduct the live broadcast explanation;

[0061] S11. During the live broadcast, the software platform and system periodically generate a new live broadcast script through a third-party LLM large language model based on the live broadcast room data and the live broadcast script strategy within the software platform and system. After a segmented script in the live broadcast script currently being played is played, the third-party speech synthesis technology is called to insert the sound of the current live broadcast script;

[0062] S12. During the live broadcast, after the software platform and system obtain questions from the audience in the barrage area in real time, they use intent recognition technology or keyword matching and a third-party LLM large language model to generate an interactive response script. After a segmented script is played in the live broadcast script being played at that time, the third-party speech synthesis technology is called to insert the sound of this interactive response script.

[0063] S13. Determine whether there are other products to be explained in the live broadcast. If yes, control the intelligent mobile robot to move to the next live broadcast shooting point for explaining the product. If no, play the next broadcast speech and then close the live broadcast. When the live broadcast ends, the software platform and system control the intelligent mobile robot to walk to the charging position for charging and standby through the intelligent mobile robot control system interface.

[0064] The above is only a preferred embodiment of the present invention. It should be understood that the present invention is not limited to the form disclosed herein, and should not be regarded as excluding other embodiments, but can be used for various other combinations, modifications and improvements, and can be modified within the scope of the concept described herein through the above teachings or the technology or knowledge of the relevant field. The changes and modifications made by those skilled in the art do not deviate from the spirit and scope of the present invention, and should be within the scope of protection of the claims attached to the present invention.

Claims

1. An automatic live broadcast method applied to an Internet platform, characterized in that: The automatic live broadcast method comprises: Step 1: Create a live broadcast task based on the pre-set and collected data, and after detecting the live broadcast task, generate a complete script for the live broadcast according to the requirements of the live broadcast task; Step 2: Cut the complete script into segments, control the intelligent mobile robot to go to the first live broadcast shooting point to explain the product, and then start the live broadcast; Step 3: Live broadcast the segmented live broadcast scripts one by one, and generate new live broadcast scripts according to the live broadcast script strategy and the real-time acquired bullet screen area information. After a segment of the currently playing live broadcast script is played, insert the audio playback of this segment strategy or interactive script; Step 4: Determine whether there are other products to be explained in this live broadcast. If so, control the intelligent mobile robot to move to the live shooting point of the next product to be explained. If not, play the next broadcast and then close this live broadcast.

2. The automatic live broadcast method applied to the Internet platform according to claim 1, characterized in that: The step 1 specifically includes: Create enterprise knowledge base and product knowledge base in software platforms and systems. The product knowledge base includes product name, product introduction, configuration rules, product advantages, product price, product benefits and competitive product comparison; Select the gender, personality style, and timbre of the virtual voice host provided by the system in the software platform or system, or use the third-party large model voice technology to replicate the voice based on the provided voice sample; The intelligent mobile robot's built-in control system is used to collect and create walking maps of the company's live broadcast locations; The control system of the intelligent mobile robot is used to calibrate the live shooting points of different products or corporate display content in the enterprise live broadcast venue, and the point information is synchronously saved in the software platform and system; Create live broadcast tasks on the Internet platform in the software platform and system, set the live broadcast theme, live broadcast cover, live broadcast cycle, live broadcast time, product explanation and virtual voice anchor who will not leave the country; After the software platform and system detects a new live broadcast task, it will combine prompt words in advance according to the requirements of the live broadcast task, the enterprise knowledge base, the live broadcast product knowledge base and the style of the virtual voice anchor, and use the third-party LLM large language model to generate a complete script for the live broadcast.

3. The automatic live broadcast method applied to the Internet platform according to claim 1, characterized in that: The step 2 specifically includes: The software platform and system cut and segment the generated live broadcast script; Before the live broadcast time of the live broadcast task arrives, the software platform and system control the intelligent mobile robot to move to the first live broadcast shooting point for explaining the product through the intelligent mobile robot control system interface; When the live broadcast time of the live broadcast task arrives, the software platform and system use RPA technology to start the live broadcast on the computer-side live broadcast tools of various Internet platforms or the mobile APPs of various Internet platforms.

4. The automatic live broadcast method applied to the Internet platform according to claim 1, characterized in that: The step three specifically includes: The software platform and system will sequentially cut the live broadcast into segments according to the gender and voice of the virtual anchor selected for the live broadcast, and call the third-party speech synthesis technology to play the sound and conduct the live broadcast explanation; During the live broadcast, the software platform and system periodically generate a new live broadcast script through the third-party LLM large language model based on the live broadcast room data and the live broadcast script strategy within the software platform and system. After a segmented script in the live broadcast script that is being played is played, the third-party speech synthesis technology is called to insert the sound of the current live broadcast script. During the live broadcast, the software platform and system obtain questions from the audience in the barrage area in real time, and use intent recognition technology or keyword matching to generate an interactive response script through a third-party LLM large language model. After a cut segment of the live broadcast script is played, the third-party speech synthesis technology is called to insert the sound of this interactive response script.

5. An automatic live broadcast system applied to an Internet platform, characterized in that: The system includes a data collection and speech generation module, a speech cutting and live broadcast start module, a live broadcast explanation and speech playback module and a judgment module; The data collection and speech generation module is configured to create a live broadcast task according to the pre-set and collected data, and after detecting the live broadcast task, generate a complete speech script for the live broadcast according to the requirements of the live broadcast task; The script cutting and live broadcast start module is configured to cut and segment the complete script, and control the intelligent mobile robot to go to the first live broadcast shooting point for explaining the product and then start the live broadcast; The live explanation and speech playing module is configured to sequentially explain the segmented live speech, and generate new live speech according to the live speech strategy and the real-time acquired bullet screen area information, and insert the sound playback of the current strategy or interactive speech after a segment of the live speech currently being played is finished; The judgment module is configured to judge whether there are other products to be explained in the live broadcast. If so, the intelligent mobile robot is controlled to move to the next live broadcast shooting point for explaining the product. If not, the live broadcast is closed after playing the next broadcast speech.

6. The automatic live broadcast system applied to the Internet platform according to claim 5, characterized in that: The data collection and speech generation module specifically includes the following contents: Create enterprise knowledge base and product knowledge base in software platforms and systems. The product knowledge base includes product name, product introduction, configuration rules, product advantages, product price, product benefits and competitive product comparison; Select the gender, personality style, and timbre of the virtual voice host provided by the system in the software platform or system, or use the third-party large model voice technology to replicate the voice based on the provided voice sample; The intelligent mobile robot's built-in control system is used to collect and create walking maps of the company's live broadcast locations; The control system of the intelligent mobile robot is used to calibrate the live shooting points of different products or corporate display content in the enterprise live broadcast venue, and the point information is synchronously saved in the software platform and system; Create live broadcast tasks on the Internet platform in the software platform and system, set the live broadcast theme, live broadcast cover, live broadcast cycle, live broadcast time, product explanation and virtual voice anchor who will not leave the country; After the software platform and system detects a new live broadcast task, it will provide the enterprise knowledge base, live broadcast product knowledge base and virtual voice anchor personality style in advance according to the requirements of the live broadcast task, and combine them into prompt words to generate a complete script for this live broadcast through the third-party LLM large language model.

7. The automatic live broadcast system applied to the Internet platform according to claim 5, characterized in that: The speech cutting and live broadcast start module specifically includes the following contents: The software platform and system cut and segment the generated live broadcast script; The software platform and system control the intelligent mobile robot to move to the first live broadcast shooting point for explaining the product through the intelligent mobile robot control system interface; When the live broadcast time of the live broadcast task arrives, the software platform and system will start the live broadcast through the computer-side live broadcast tools of various Internet platforms or the mobile APPs of various Internet platforms through RPA technology.

8. The automatic live broadcast system applied to the Internet platform according to claim 5, characterized in that: The live explanation and speech playing module specifically includes the following contents: The software platform and system will sequentially cut the live broadcast into segments according to the gender and voice of the virtual anchor selected for the live broadcast, and call the third-party speech synthesis technology to play the sound and conduct the live broadcast explanation; During the live broadcast, the software platform and system periodically generate a new live broadcast script through the third-party LLM large language model based on the live broadcast room data and the live broadcast script strategy within the software platform and system. After a segmented script in the live broadcast script that is being played is played, the third-party speech synthesis technology is called to insert the sound of the current live broadcast script. During the live broadcast, the software platform and system obtain questions from the audience in the barrage area in real time, and use intent recognition technology or keyword matching to generate an interactive response script through a third-party LLM large language model. After a cut segment of the live broadcast script is played, the third-party speech synthesis technology is called to insert the sound of this interactive response script.

Images