Plug-in type AI assistant system and use method thereof

Abstract

The invention discloses a plug-in type AI assistant system and a using method thereof. The plug-in type AI assistant system comprises a scene capturing and intelligent monitoring module, a question predicting and answering module, an output feedback module and an active interaction interface, and the modules are integrated into a plug-in type; through introduction of a scene capturing and intelligent monitoring module, capturing of operation obstacles is automatically completed, user consulting time is shortened, the flow processing speed is increased, and the business handling efficiency is improved. The user experience is remarkably improved, and the situation that the user gives up business handling due to difficult operation is reduced through active arousing after the obstacle scene is captured; due to the flexible plugging capacity, the integration efficiency of the AI assistant and the service system is greatly improved, and rapid expansion of the capacity of the AI assistant is facilitated.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a plug-in AI assistant system and a method for using the same. Background Art

[0002] With the development of Internet technology, online service platforms have been widely used in all walks of life, especially in the business processes of some products, where users need to complete various business processes through complex operation steps. However, users often face problems such as unclear processes, complex operations, and long system response times during the operation process. Especially in cases involving strict authentication mechanisms and interdependence of functional modules, the user experience is seriously affected. At present, the existing system mainly relies on the following ways to assist users: Static help documents provide operation guides and answers to frequently asked questions, but lack real-time interaction and cannot provide personalized help based on the user's specific operation status; Manual customer service, which answers user questions through manual service personnel, can provide customized help, but the response speed is slow and it is difficult to meet the needs of a large number of users in high concurrency situations; Intelligent customer service uses natural language processing technology to help users solve problems through a preset question-and-answer mechanism. However, as an independent tool, it lacks deep linkage with business processes, which means that users need to actively seek help and has high requirements on the user's ability to ask questions. The above-mentioned auxiliary methods all exist in the form of independent tools and lack close integration with the business operating system, which affects the efficiency of solving user problems and the overall user experience. Summary of the invention

[0003] In view of the defects of the prior art, the present invention provides a plug-in AI assistant system and a method of using the same, which deeply integrates intelligent customer service with a business operating system, intelligently identifies problems that users may encounter by detecting user operation status and operation breakpoints in real time, and proactively provides users with personalized help and guidance within the business process. By utilizing a "pluggable" architecture design, efficient integration of AI assistants in different business operating systems is achieved.

[0004] Based on the above objectives, in a first aspect, the present invention provides a plug-in AI assistant system, comprising: Scene capture and intelligent monitoring module, problem prediction and answering module, output feedback module, active interactive interface, the system uses a plug-in mechanism to achieve the integration of various modules; The scene capture and intelligent monitoring module is integrated with the business operating system through the API interface, receives and processes the data information transmitted by the business system, and performs user behavior tracking and data collection; The question prediction and answering module receives the behavior parameter information from the scene capture and intelligent monitoring modules, and analyzes in real time whether the user's current operation hits the pre-trained scene. If so, it generates answer rules based on the knowledge base content; The output feedback module sets flexibly configurable trigger conditions. When the conditions are met, prompt information and inquiry dialogues are actively sent. Based on the problem prediction results, relevant answer information is retrieved from the knowledge base, and natural language generation technology is used to convert the answers into easy-to-understand text and operation buttons to help users complete operations quickly.

[0005] The proactive interactive interface guides users to seek help in a step-by-step manner, avoiding disturbing users who do not need help.

[0006] Furthermore, each module acts as an independent plug-in and interacts with the business operating system through a standardized API interface to achieve plug-and-play of the module. The plug-in management system is responsible for loading, unloading and parameter configuration of the plug-in to ensure the flexibility and scalability of the system.

[0007] Furthermore, the scene capture and intelligent monitoring module captures and monitors the user's operation behavior (such as clicking, scrolling, staying, inputting, etc.) and system status information (such as response time, error log, etc.) in real time.

[0008] Furthermore, the problem prediction and answering module uses machine learning algorithms and pre-trained models to predict problems that users may encounter based on user historical behavior data, system status information, system logs and external knowledge bases, and generates targeted solutions.

[0009] Furthermore, the triggering conditions of the output feedback module include the operation time exceeding a threshold, the appearance of a specified type of error prompt, the user repeating the same operation multiple times, etc.

[0010] Furthermore, the active interaction interface reminds users through "AI assistant entrance highlight + prompt", and they can click the entrance to enter the human-computer chat interface, supporting multiple devices and platforms (PC, mobile, Web, etc.).

[0011] In a second aspect, the present invention further provides a method for using a plug-in AI assistant system, comprising: S1, data interaction between the business operating system and the AI ​​assistant system; S2, data flow and processing between modules of the AI ​​assistant system; S21, according to the previous definition, the business operating system collects data related to user operation behavior at an appropriate frequency when operating the business system. These data are transmitted to the scene capture and intelligent monitoring module of the AI ​​assistant in a decentralized form; S22, after receiving the information, the scene capture and intelligent monitoring module arranges the field information according to the preset scene rules, performs data cleaning, removes the field information that is invalid for scene recognition, and generates a standardized array; S23, the integrated standardized array will be submitted to the question prediction and question answering module, and the scene recognition will be performed through the AI ​​model. If the recognition does not hit any scene, it is determined that the user does not need assistance; if the scene is hit, the obtained scene rules will be submitted to the output feedback module; S24, the output feedback module receives the scene information hit by the AI ​​model, matches the processing result preset in the system for the scene, and feeds back the information to the user side through the active interaction interface of the front end; S25, after the user has collected system prompt information in the active interaction interface, he or she learns solutions to current operational difficulties according to the guidance of the AI ​​assistant. For the quick business processing function provided by the AI ​​assistant, the user can quickly complete the business processing on the AI ​​assistant chat page; S26, the user's response information in the interactive interface will be transmitted to the business operating system in real time, including the solution results of the current scenario and the node update status of the business processing.

[0012] Furthermore, in step S1, the data interaction is divided into two parts. The first part is the necessary information that is not related to business operations and meets the prerequisites for plug-in application, such as plug-in authentication information, plug-in configuration information, security policy information, version and compatibility information; the second part is strongly related to business platform operations and is used to capture user operation scenarios. This type of information is standardized by the AI ​​assistant field set, including but not limited to page routing, page hot spots, page button definitions, front paths and back paths, regular operation time thresholds, error codes, etc.

[0013] Compared with the prior art, the plug-in AI assistant system and its use method proposed in the present invention have the following beneficial effects: Through the introduction of scene capture and intelligent monitoring modules, operational obstacles can be captured automatically, reducing user consultation time, speeding up process processing, and improving business handling efficiency. The user experience is significantly improved, and the active reminder after obstacle scene capture is completed helps reduce the situation where users give up business handling due to operational difficulties. The flexible plug-in and unplug capabilities greatly improve the integration efficiency of AI assistants and business systems, which helps to quickly expand the capabilities of AI assistants. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 A plug-in AI assistant system and a flow chart of its use method provided by the present invention;.

[0015] Figure 2A schematic diagram of data flow between modules of a plug-in AI assistant system provided by the present invention. DETAILED DESCRIPTION

[0016] The specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings to further illustrate the technical solutions and implementation details of the present invention.

[0017] 1. Pluggable in-process AI assistant system architecture The system as a whole adopts a pluggable architecture design, and realizes the independent deployment and management of each module of the AI ​​assistant through a plug-in mechanism. The system mainly includes the following modules: Scene capture and intelligent monitoring module: integrated into the business operating system through the API interface, responsible for real-time monitoring and capturing user operation behaviors and system status information in the business operating system; Problem prediction and answering module: uses machine learning algorithms and pre-trained models to predict problems that users may encounter based on collected user behavior data, system status information, and external knowledge bases, and generates corresponding solutions; Output feedback module: provides help and guidance to users through active interactive interfaces, including trigger mechanisms, interactive interfaces, and answer generation; Active interaction interface: guide users to seek help in a step-by-step manner to avoid disturbing users who do not need help; Each module communicates with the business operating system and other modules through a standardized API interface to ensure efficient collaboration of the system.

[0018] 2. Plug-in integration method When the business operating system needs to integrate the AI ​​assistant, it calls the plug-in of the AI ​​assistant through the API interface. The plug-in management system is responsible for loading, unloading and parameter configuration of the plug-in to achieve plug-and-play of the module. The specific steps include: Plug-in loading: Load the relevant plug-in modules of the AI ​​assistant through the API interface according to the requirements of the business operating system; Parameter configuration: configure the running parameters of the plug-in, including trigger conditions, knowledge base path, API interface address, etc.; Module initialization: Each plug-in module is initialized and establishes a communication connection with the business operating system to start real-time monitoring and data interaction; Dynamic management: During the operation of the business operating system, the plug-in management system can dynamically load or unload the AI ​​assistant module according to demand to ensure the flexibility and scalability of the system.

[0019] 3. User operation behavior monitoring and problem prediction The system collects user operation behavior data through scene capture and intelligent monitoring modules, including click, scroll, stay, input and other operations, as well as system response time and error log information. After cleaning and standardization, these data are submitted to the question prediction and answering module for analysis; Data collection: collect user operation behavior and system status information in real time through API interface to ensure the real-time and accuracy of data; Data processing: The scene capture and intelligent monitoring modules clean the data according to the preset scene rules, remove invalid information, and form a standardized data set; Problem prediction: The problem prediction and answering module uses a pre-trained machine learning model to predict the operational obstacles that users may currently encounter based on historical data and knowledge base content, and generates corresponding solutions.

[0020] 4. Active interaction and user guidance When the system identifies operational obstacles that the user may encounter, the output feedback module actively provides help to the user through the interactive interface. The specific process is as follows: Trigger condition detection: The system detects the user's current operation status based on preset trigger conditions (such as operation timeout, frequent error prompts, etc.); Prompt message sending: When the trigger conditions are met, the system proactively sends prompt messages and inquiry scripts to invite users to seek help; Interactive interface display: After the user clicks on the AI ​​assistant entrance, the user enters the human-computer chat interface. The system uses natural language generation technology to display detailed operation instructions and solutions; Operational guidance execution: Users complete the corresponding operation steps according to the guidance provided by the system. The system updates the progress status of the business operation system in real time based on the user's response.

[0021] For example: When a user is placing an order and paying, he clicks the [Pay] button several times and is prompted "Please read the agreement first". However, since the page content is too long, the user cannot quickly find the location of the agreement. At this time, the AI ​​assistant takes the initiative to help and prompts the user "Did you encounter any problems when going to [Pay]? You can first read the following agreement and click [Agree] to continue to pay", and provides a quick navigation button for the agreement content in the chat interface to help users quickly complete the operation.

[0022] 5. Data Flow and Security The system ensures that data interaction between the business operating system and the AI ​​assistant is secure and reliable. Specific measures include: Data encryption: All transmitted data uses encryption protocols to prevent data from being intercepted or tampered with during transmission; Permission control: Verify the legitimacy of the plug-in through the authentication mechanism, and limit the scope of data access according to the permission configuration to ensure data security; Data isolation: Data not related to business operations (such as plug-in authentication information, configuration information, etc.) is strictly isolated from data related to business operations (such as user operation scenario information) to prevent data leakage; Logging: The system logs all data interactions and operational behaviors to facilitate subsequent audits and problem tracking.

Claims

1. A plug-in AI assistant system, characterized in that: include: Scene capture and intelligent monitoring module, problem prediction and answering module, output feedback module, active interactive interface, the modules are integrated into a plug-in type; The scene capture and intelligent monitoring module is integrated with the business operating system through the API interface, receives and processes the data information transmitted by the business system, and performs user behavior tracking and data collection; The question prediction and answering module receives the behavior parameter information from the scene capture and intelligent monitoring modules, and analyzes in real time whether the user's current operation hits the pre-trained scene. If so, it generates answer rules based on the knowledge base content; Output feedback module, set trigger conditions, when the conditions are met, actively send prompt information and inquiry words, based on the problem prediction results, retrieve relevant answer information from the knowledge base, convert the answer into text and operation buttons, to help users quickly complete the operation; Active interactive interface guides users to seek help and avoids disturbing users who don’t need help.

2. A plug-in AI assistant system according to claim 1, characterized in that: Each module of the system acts as an independent plug-in and interacts with the business operating system through a standardized API interface. The plug-in management system is responsible for loading, unloading and parameter configuration of the plug-in.

3. A plug-in AI assistant system according to claim 1, characterized in that: The scene capture and intelligent monitoring module captures and monitors the user's operating behavior and system status information in real time.

4. A plug-in AI assistant system according to claim 1, characterized in that: The problem prediction and answering module uses machine learning algorithms and pre-trained models to predict problems that users may encounter and generate targeted solutions based on user historical behavior data, system status information, system logs and external knowledge bases.

5. The plug-in AI assistant system according to claim 1, characterized in that: The output feedback module triggering conditions include the operation time exceeding a threshold, the appearance of a specified type of error prompt, and the user repeating the same operation multiple times.

6. A plug-in AI assistant system according to claim 1, characterized in that: The active interaction interface reminds users through "AI assistant entrance highlight + prompt", and they can click the entrance to enter the human-computer chat interface.

7. A method for using a plug-in AI assistant system, characterized in that: include: S1, data interaction between the business operating system and the AI ​​assistant system; S2, data flow and processing between modules of the AI ​​assistant system; S21, the business operating system collects data related to user operation behavior according to the previous definition, and transmits the data to the scene capture and intelligent monitoring module in a decentralized form; S22, after receiving the information, the scene capture and intelligent monitoring module arranges the field information according to the preset scene rules, performs data cleaning, removes the field information that is invalid for scene recognition, and generates a standardized array; S23, the standardized array is submitted to the question prediction and question answering module, and the scene recognition is performed through the AI ​​model. If no scene is hit by the recognition, it is determined that the user does not need assistance; if the scene is hit, the obtained scene rule is submitted to the output feedback module; S24, the output feedback module matches the processing result preset in the system for the scene through the scene information, and feeds back the information to the user via the active interaction interface; S25, after the user collects the system prompt information on the active interaction interface, the user completes the service processing according to the guidance and the quick service processing function; S26, the user's response information on the active interaction interface is transmitted to the business operating system in real time.

Images