Intelligent interaction method, apparatus, device, medium, and product

Abstract

Embodiments of the present application disclose intelligent interaction methods, devices, equipment, media and products, which can be applied to the technical field of data processing. The method comprises: displaying a business page, and displaying an intelligent interaction control for triggering intelligent interaction on the business page; in response to a triggering operation on the intelligent interaction control, displaying an intelligent interaction page for the business page; the intelligent interaction page comprises an instruction input area; in response to a target instruction for the business page sent through the instruction input area, executing an intelligent interaction task indicated by the target instruction through at least one agent, and displaying a task execution progress corresponding to the at least one agent in a progress display area; when the intelligent interaction task is completed, displaying a business-related page after page processing on the business page in the progress display area. The embodiments of the present application can help to improve the convenience of intelligent interaction for the business page.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to intelligent interaction methods, devices, equipment, media, and products. Background Technology

[0002] Currently, during intelligent interaction, if a user needs to initiate a command for a business page, they can send text and a screenshot of the page to the intelligent interaction system to send the task command. The intelligent interaction system can then analyze the sent task command and execute the corresponding task, and then output the task execution result to the user in the form of a dialogue.

[0003] However, the inventors discovered in practice that sending task instructions for business pages via text and screenshots requires inputting a significant amount of text to ensure accuracy, resulting in high communication costs during the intelligent interaction process. Furthermore, after the intelligent interaction system outputs the result of the instruction, if the task requires adjustments to the corresponding business system, the user must manually perform the operation on the business page, detaching from the intelligent interaction interface, making the intelligent interaction process for business pages cumbersome. Therefore, improving the convenience of intelligent interaction for business pages is a pressing issue. Summary of the Invention

[0004] This application provides an intelligent interaction method, device, equipment, medium, and product that can directly trigger intelligent interaction with intelligent objects based on a business page. After sending a target instruction, the task execution progress of each intelligent agent can be displayed when at least one intelligent agent performs the task. Furthermore, when completing a task on a business page, the adjustments on the business page can be directly applied without manually performing operations on the business page, thus enabling better business processing and improving the convenience of intelligent interaction with business pages.

[0005] On one hand, embodiments of this application disclose an intelligent interaction method, the method comprising: Display the business page, and display the intelligent interaction control on the business page to trigger intelligent interaction; In response to a trigger operation on the smart interactive control, a smart interactive page for the business page is displayed; the smart interactive page includes a command input area; In response to a target instruction for a business page sent through the instruction input area, at least one intelligent agent executes the intelligent interactive task indicated by the target instruction, and displays the task execution progress corresponding to each of the at least one intelligent agent in the progress display area; the at least one intelligent agent is determined based on the target instruction and the page context data of the business page; the at least one intelligent agent includes an intelligent agent for page processing of the business page; When a smart interaction task is completed, the business-related page after the business page has been processed is displayed in the progress display area.

[0006] On one hand, embodiments of this application disclose an intelligent interaction method, the method comprising: The system receives intelligent interaction requests triggered by a terminal device on a business page, obtains page context data of the business page through multiple information sources, and returns page data for the intelligent interaction page to the terminal device for displaying the intelligent interaction page on the terminal device; the intelligent interaction page includes a command input area for command entry. Upon receiving a target instruction for a business page sent through the instruction input area, at least one intelligent agent is determined based on the intelligent collaborative processing strategy and the target instruction; the at least one intelligent agent includes an intelligent agent for performing page processing on the business page; The system executes the intelligent interactive task indicated by the target instruction through at least one intelligent agent and returns the task execution progress corresponding to each of the at least one intelligent agent to the terminal device, so as to display the task execution progress corresponding to each of the at least one intelligent agent in the progress display area. When the intelligent interaction task is completed, the page data of the business-related page after the page processing of the business page is returned to the terminal device so that the business-related page can be displayed in the progress display area.

[0007] On one hand, embodiments of this application disclose an intelligent interaction method, the method comprising: Display the target session page, which includes a session input area; In response to sending target session content through the session input area, the system executes instructions corresponding to the first session content based on the first agent and instructions corresponding to the second session content based on the second agent; wherein, the target session content includes first identification information of the first agent and first session content for the first agent, and second identification information of the second agent and second session content for the second agent; The conversation input area triggers the display of an agent list after the target character is entered. The target agent is selected from the agent list and the target identification information of the target agent is entered in the conversation input area. The agent list includes at least two agents, and the target agent is any one of the at least two agents.

[0008] On one hand, embodiments of this application disclose an intelligent interactive device, the device comprising: The business page display module is used to display the business page and display intelligent interactive controls on the business page to trigger intelligent interaction; The intelligent interactive page module is used to respond to the triggered operation of the intelligent interactive control and display the intelligent interactive page for the business page; the intelligent interactive page includes a command input area; The task execution module is used to respond to a target instruction for a business page sent through the instruction input area, execute the intelligent interactive task indicated by the target instruction through at least one intelligent agent, and display the task execution progress of each of the at least one intelligent agent in the progress display area; the at least one intelligent agent is determined based on the target instruction and the page context data of the business page; the at least one intelligent agent includes an intelligent agent for page processing of the business page; The task completion module is used to display the business-related pages after page processing on the business page in the progress display area when an intelligent interactive task is completed.

[0009] On one hand, embodiments of this application disclose an intelligent interactive device, the device comprising: The intelligent interaction request module is used to receive intelligent interaction requests triggered by the terminal device based on the business page, obtain the page context data of the business page through multiple information sources, and return the page data for the intelligent interaction page to the terminal device so as to display the intelligent interaction page on the terminal device; the intelligent interaction page includes a command input area for command input; The agent determination module is used to determine at least one agent based on the intelligent collaborative processing strategy and the target instruction when receiving a target instruction for a business page sent through the instruction input area; the at least one agent includes an agent for page processing of the business page. The task processing module is used to execute the intelligent interactive task indicated by the target instruction through at least one intelligent agent, and return the task execution progress corresponding to at least one intelligent agent to the terminal device so as to display the task execution progress corresponding to at least one intelligent agent in the progress display area. The feedback module is used to return the page data of the business-related pages after page processing of the business page to the terminal device when the intelligent interaction task is completed, so as to display the business-related pages in the progress display area.

[0010] On one hand, embodiments of this application disclose an intelligent interactive device, the device comprising: The conversation page display module is used to display the target conversation page, which includes a conversation input area; The instruction execution module is used to respond to sending target session content through the session input area, execute instructions corresponding to the first session content based on the first agent, and execute instructions corresponding to the second session content based on the second agent; wherein, the target session content includes first identification information of the first agent and first session content for the first agent, and second identification information of the second agent and second session content for the second agent; The conversation input area triggers the display of an agent list after the target character is entered. The target agent is selected from the agent list and the target identification information of the target agent is entered in the conversation input area. The agent list includes at least two agents, and the target agent is any one of the at least two agents.

[0011] One embodiment of this application provides a computer device, including a memory and a processor; the memory is connected to the processor, the memory is used to store a computer program, and the processor is used to invoke the computer program so that the computer device performs the method provided in any of the above embodiments.

[0012] One aspect of this application provides a computer-readable storage medium storing a computer program adapted to be loaded and executed by a processor, so that a computer device having a processor performs the method provided in any of the above aspects.

[0013] One aspect of this application provides a computer program product, which includes a computer program stored in a computer-readable storage medium. The computer program is adapted to be read and executed by a processor so that a computer device having a processor performs the method provided in any of the above aspects.

[0014] Using the embodiments of this application, a business page can be displayed, and intelligent interactive controls for triggering intelligent interaction can be displayed on the business page. Then, in response to the triggering operation of the intelligent interactive controls, an intelligent interactive page for the business page is displayed. The intelligent interactive page includes an instruction input area. Then, in response to a target instruction for the business page sent through the instruction input area, at least one intelligent agent executes the intelligent interactive task indicated by the target instruction, and the task execution progress corresponding to each of the at least one intelligent agent is displayed in the progress display area. The at least one intelligent agent is determined based on the target instruction and the page context data of the business page, and includes an intelligent agent for page processing of the business page. Furthermore, upon completion of the intelligent interactive task, a business-related page after page processing of the business page is displayed in the progress display area. Based on this, intelligent interaction with intelligent objects is directly triggered based on the business page. After sending the target instruction, the task execution progress corresponding to each intelligent agent can be displayed when executing the task through at least one intelligent agent. Furthermore, when completing the task on the business page, adjustments on the business page can be directly applied without manually performing operations on the business page, thus improving the convenience of intelligent interaction with the business page. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the structure of a dialogue system provided in an embodiment of this application; Figure 2 This is a schematic diagram of a scenario for an intelligent interaction method provided in an embodiment of this application; Figure 3 This is a flowchart illustrating an intelligent interaction method provided in an embodiment of this application; Figure 4 This is a schematic diagram illustrating the effect of an intelligent interactive system provided in an embodiment of this application; Figure 5 This is a flowchart illustrating an instruction initiation process provided in an embodiment of this application; Figure 6 This is a flowchart illustrating a candidate skill triggering process provided in an embodiment of this application; Figure 7 This is a flowchart illustrating an intelligent interaction method provided in an embodiment of this application; Figure 8 This is a schematic diagram of a process for determining an intelligent agent provided in an embodiment of this application; Figure 9 This is a timing interaction diagram of an intelligent interaction method provided in an embodiment of this application; Figure 10 This is a flowchart illustrating an intelligent interaction method provided in an embodiment of this application; Figure 11a This is a schematic diagram of a process for entering conversation content provided in an embodiment of this application; Figure 11b This is a schematic diagram of a process for entering conversation content provided in an embodiment of this application; Figure 12 This is a schematic diagram of the structure of an intelligent interactive device provided in an embodiment of this application; Figure 13 This is a schematic diagram of the structure of an intelligent interactive device provided in an embodiment of this application; Figure 14 This is a schematic diagram of the structure of an intelligent interactive device provided in an embodiment of this application; Figure 15 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0018] The following describes some technical terms used in the embodiments of this application: An intelligent agent (also called an intelligent object) is a system capable of perceiving its environment, making decisions, and executing actions to achieve specific goals. It possesses autonomy, adaptability, and interactivity. It can make judgments and decisions based on its learned knowledge and algorithms, thereby performing actions to influence the environment or achieve predetermined goals. This can be built based on a large language model (LLM model) and can also be called an intelligent object. For example, in the embodiments of this application, an intelligent object can be used to receive instructions sent by a user and execute the intelligent interactive tasks indicated by the instructions.

[0019] "In response to" indicates the state in which a corresponding event occurs or a condition is met. The timing of subsequent actions performed in response to this event or condition is not necessarily strongly correlated with the time when the event occurs or the condition is met. For example, in some cases, subsequent actions may be performed immediately when the event occurs or the condition is met; while in other cases, subsequent actions may be performed some time after the event occurs or the condition is met.

[0020] Triggered operations refer to actions performed on information (such as controls) provided by a computer device. These operations send corresponding instructions to the computer device, triggering it to execute the next task. The next task triggered by different triggered operations on different information can all be pre-programmed. These triggered operations can be performed manually by the user; for example, clicks, double-clicks, long presses, and swipes performed by the user on content displayed on the computer screen are all triggered operations. In some cases, these triggered operations can also be executed by the computer device based on a pre-defined program.

[0021] Published content: This refers to user-submitted content information on the content platform, which can include text, images, audio, and video, and can be displayed in formats such as notes, articles, and video files. The specific content information included in published content can be adjusted for different scenarios, and the content information in the published content can be determined by the publisher. Published content is stored on the server and distributed to users browsing the content platform through a content distribution algorithm deployed on the server.

[0022] Please see Figure 1 , Figure 1 This is a schematic diagram of the structure of a dialogue system provided in an embodiment of this application. Figure 1 As shown, this dialogue system may include terminal devices (such as device 11a, device 12a, and device 13a) and server 100a. It is understood that... Figure 1 The number of terminal devices and servers shown is merely illustrative; any number of terminal devices and servers can be used as needed. Terminal devices (such as devices 11a, 12a, and 13a) can communicate with servers via a network (i.e., through a medium that provides a communication link via wired, wireless communication links, or fiber optic cables, etc.) to transmit data.

[0023] It is understood that a client or webpage may run on the terminal device (such as device 12a), and this client or webpage may be a program that provides local services to the user. Server 100a may be the server corresponding to the client or webpage, and server 100a may run programs for providing resources, service data, and other services to the client or webpage. It is understood that an account may be logged in on the client or webpage. For example, the client or webpage running on the terminal device may be the client or webpage corresponding to a business platform such as a social application, e-commerce, or data analysis platform. Furthermore, on the client or webpage, the user can trigger interaction with the intelligent interaction system based on the business platform, enabling the user to interact intelligently with intelligent objects based on the business page.

[0024] It is understood that terminal devices (such as device 11a) may include, but are not limited to, mobile phones, computers, smart voice interaction devices, smart home appliances, vehicle terminals, aircraft, smart speakers, etc., without limitation. The server 100a may be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms, without limitation.

[0025] Please see Figure 2 , Figure 2 This is a schematic diagram of a scenario for an intelligent interaction method provided in an embodiment of this application. For example... Figure 2 As shown, page 211a can be a business page, which can be used to display analyzed data, such as a page for displaying relevant data of a certain product or brand obtained from the analysis. This business page can include some filter options, such as filter option 1 (e.g. Figure 2 As shown in 202a), filter option 2, filter option 3, etc., the filter options here can be filtered by time, brand, product, etc. Furthermore, multiple business objects can be displayed on this page, such as object A, object B, object C, object D, object E, object F, etc. Here, object A (e.g., Figure 2 Taking 203a as an example, object A can also include its sub-objects, such as object a1, object a2, object a3, etc., to achieve multi-level display of page objects. What each object specifically represents depends on the actual content of the business page. Page 211a is only an exemplary view, and its specific display structure and content can be determined based on the actual situation, which is not limited here. In addition, intelligent interaction controls (i.e., control 201a) for triggering intelligent interaction with the intelligent interaction system can be displayed on page 211a.

[0026] Further, in response to the triggering operation of control 201a, page 212a is displayed (step S21a). Page 212a can be an intelligent interactive page, which may include a thumbnail window (window 204a) for displaying a business page in a thumbnail format. The business page can also be displayed in thumbnail format on page 212a. Furthermore, a command input area (area 205a) can be displayed on page 212a for inputting commands by referring to the content displayed in the thumbnail window. Further, in response to command Z1 sent through area 205a, the task execution progress corresponding to at least one intelligent agent is displayed on area 221a (step S22a). Command Z1 can be a target command entered and sent in area 205a. After sending command Z1, the display effect of the intelligent interactive page can be seen on page 213a. This intelligent interactive page may include a dialogue area (area 222a), in which a dialogue message (message 224a) carrying the sent command Z1 can be displayed. Furthermore, the intelligent interaction page may include a progress display area (i.e., area 221a). Within this task display area, the overall task progress of the intelligent interaction task (or simply task) corresponding to instruction Z1 can be displayed (as shown in progress 206a in Figure 2). Additionally, the task execution progress of at least one intelligent agent for the intelligent interaction task can be displayed. This at least one intelligent agent can be determined based on the target instruction and the page context data of the business page. The specific determination method is not elaborated here. For example, this at least one intelligent agent may include agent 1 and agent 2. The task execution progress corresponding to agent 1 can be seen in progress 207a, and the task execution progress corresponding to agent 2 can be seen in progress 208a. As can be seen, when displaying the task execution progress, a progress bar (and / or progress percentage) can be displayed, as well as the currently executing step, such as agent 1 performing step 1 and agent 2 performing step 2.

[0027] Furthermore, upon completion of the task, page 223a is displayed in area 221a (step S23a). Page 223a can be a relevant page generated after completing the intelligent interaction task. It can be a page obtained by processing page 223a, or a page related to the business page generated based on page processing of page 223a; this is not limited here. Page processing can be the process of performing CRUD operations on page objects on the business page. Based on this, this application enables direct CRUD operations on business pages through intelligent interaction, eliminating the need for manual deployment of CRUD operations based on the execution results of intelligent interaction, thereby improving the convenience of intelligent interaction processes for business pages. Additionally, after task completion, the task execution results output by the intelligent interaction system (such as...) can be displayed in the dialogue area. Figure 2 The result (as shown in 209a) informs the current user of the result of the task corresponding to the executed instruction.

[0028] It should be noted that all user data collected in this application is collected with the user's consent and authorization, and the collection, use, and processing of relevant user data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, this application may display prompts before and during the collection of user data to inform the user that their data is being collected. This ensures that the application only begins the steps for collecting user data after receiving confirmation from the user regarding the prompt; otherwise (i.e., without receiving confirmation from the user), the steps for collecting user data are terminated, meaning that user data is not collected.

[0029] It is understood that the above scenarios are merely examples and do not constitute a limitation on the application scenarios of the technical solutions provided in the embodiments of this application. The technical solutions of this application can also be applied to other scenarios. For example, as those skilled in the art will know, with the evolution of system architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0030] Based on the above description, this application proposes an intelligent interaction method. Please refer to [link to relevant documentation]. Figure 3 , Figure 3 This is a flowchart illustrating an intelligent interaction method provided in an embodiment of this application. The method can be executed by the aforementioned computer device, such as a terminal device (as described above). Figure 1 (Device 11a in the example). The intelligent interaction method may include the following steps S101-S104.

[0031] S101. Display the business page and show the intelligent interaction control on the business page to trigger intelligent interaction.

[0032] The business page can be a page used to carry out specific business operations, or a page provided by a content platform for providing data analysis, data querying, and other business operations. In other words, this business page can be a page within a business system (also known as a business platform). For example, this business page could be used to analyze user data for a merchant or brand, providing information such as user profiles, behavioral preferences, and consumption characteristics. Alternatively, it could be used to analyze the operational status of a brand, showcasing its comprehensive operational performance from multiple dimensions, including brand assets, user awareness, market performance, and competitor comparisons. Or, it could be used to analyze published content on a content platform, statistically analyzing interaction data, content tags, and dissemination effects to help account operators select high-quality content and optimize posting strategies.

[0033] The business page may include page objects. These page objects can refer to any object presented on the business page. For example, on a business page used to analyze user data for brand A, the page objects may include brand information for brand A, information describing key data metrics for brand A (such as search volume, post-search exposure penetration rate, etc.), information on popular products and / or standardized product units (and SPUs) under brand A, information on products and / or standardized product units (and SPUs) currently being promoted under brand A, and modules for filtering the displayed content; these are not limited here.

[0034] It is understandable that the page objects on different page rows can be different, depending on the actual business page; no specific restrictions are placed on the page objects here. It is also understandable that on a business page, a page object can contain multiple levels of objects. For example, a business page can include at least one first-level page object, and under that first-level page object, there can be at least one second-level page object, and so on, until the smallest level of page object is reached, thus allowing multiple levels of page objects to be displayed on the business page.

[0035] Intelligent interaction can refer to the process of interaction based on artificial intelligence technology, that is, the ability to interact with an intelligent interaction system. An intelligent interaction system can be an interaction system that relies on artificial intelligence technology. It can accept user input instructions and obtain corresponding responses through the processing of intelligent objects, thereby enabling natural and efficient interaction between people and devices.

[0036] The intelligent interactive control can be a control used to trigger interaction with the intelligent interactive system. It is understood that the intelligent interactive control can be directly displayed on the business page, or it can be displayed on the extended function page in response to a trigger operation on the extended function entry point of the business page; this is not limited here. The extended function entry point can be an entry point used to trigger the display of extended functions. The extended function page can be used to display controls corresponding to the extended functions of the business page. For example, the controls corresponding to the extended functions can include the aforementioned intelligent interactive control, a sharing control for sharing the business page, etc., and this is not limited here.

[0037] It is understood that the intelligent interaction system in this application can be a multi-agent collaborative system, in which various types of intelligent objects can collaborate to complete intelligent interaction tasks. The intelligent interaction system may include a controlling intelligent agent, a capability-matching intelligent agent, and an expert intelligent agent.

[0038] This section illustrates intelligent objects in an intelligent interaction system using diagrams. Please refer to [link / reference]. Figure 4 , Figure 4 This is a schematic diagram illustrating the effect of an intelligent interactive system provided in an embodiment of this application. For example... Figure 4 As shown, an intelligent interaction system can include a master intelligent agent, a capability-matching intelligent agent, and an expert intelligent agent. Among them, the master intelligent agent (such as...) Figure 4 As shown in 401a, this can be the main control agent for task processing, also known as the Leader Agent or the master control agent. This master control agent is specifically responsible for receiving user task descriptions (such as target instructions), matching expert agents based on context data from the business page, and decomposing and orchestrating intelligent interaction tasks. It can be used to schedule various expert agents. Among these, the capability matching agent (such as...) Figure 4 As shown in section 402a, an expert agent (also known as an HR Agent or capability-matching intelligent object) can be used to determine the expert agent that matches the intelligent interaction task. The capability-matching intelligent agent can be used to perform capability-level matching in the existing expert agent library when no matching match is found in the preset expert agent library. If the matching fails or the confidence level is insufficient, template injection is triggered to generate a new expert agent. Here, an expert agent can be a dedicated agent that receives sub-tasks assigned by the master agent and independently performs analysis or operations; it can also be called a Worker Agent or expert intelligent object. It is understood that in an intelligent interaction system, the number of expert agents is not limited; for example, it can include... Figure 4 The expert agent 1 shown (e.g.) Figure 4 As shown in 403a), expert agent 2 (such as...) Figure 4 As shown in 404a), expert agent 3 (such as...) Figure 4 (as shown in 405a in the document). It is understood that, for ease of description, all types of intelligent agents in the intelligent interaction system may be referred to as intelligent objects.

[0039] S102. In response to a trigger operation on the smart interactive control, display the smart interactive page for the business page; the smart interactive page includes a command input area.

[0040] The intelligent interaction page can be a page used to interact with the intelligent interaction system. The triggering operation for the intelligent interaction control can be referred to as the intelligent interaction triggering operation for the business page.

[0041] The intelligent interaction page includes a command input area for inputting commands, and may also include a thumbnail window for displaying business pages in a compressed format.

[0042] The thumbnail window can be used to display a minimized version of a business page. It's understood that the business page displayed in this thumbnail window can be the original business page, and business operations can be performed on this page. Based on this, it's possible to adjust the business page at any time on the intelligent interaction page, and then perform intelligent interactions based on the adjusted page, without having to exit the business page, adjust it, and then re-trigger intelligent interactions on the adjusted page. This improves the flexibility and convenience of the intelligent interaction process for business pages.

[0043] Optionally, the business page displayed in the thumbnail window can be an image of the business page itself. In this case, business operations are not available on the business page displayed in the thumbnail window. The page image can be a screenshot of the business page. Based on this, the business page displayed in the thumbnail window can be loaded quickly when the intelligent interactive page is displayed.

[0044] The instruction input area is for entering instructions for the business page. It is understood that the instructions entered in this area can be text, images, or voice, etc., without limitation. It is also understood that when entering instructions, the business page in the thumbnail window can be referenced. That is, when entering instructions, it is not necessary to provide a complete description of the content on the business page; instead, instructions can be directly referenced from the content on the business page. For example, one can directly specify to process a specific object on the business page. This is because the intelligent interaction system can directly obtain relevant data from the business page in the background and then quickly map the object described in the instruction to the specific object on the business page. This allows for accurate intelligent interaction targeting the business page, solving the problems of redundant expression and semantic ambiguity caused by existing intelligent interaction systems relying solely on text descriptions and being disconnected from the business page context.

[0045] In one possible implementation, the present application embodiment may further include the following steps: 1. Displaying a thumbnail window on the intelligent interactive page for displaying a business page in a thumbnail format; the business page in the thumbnail window is used as a reference for instruction input; 2. In response to a page adjustment operation performed on the business page in the thumbnail window, displaying the business page after the page adjustment operation in the thumbnail window.

[0046] The page adjustment operation can be used to modify a business page. For example, it could involve clicking an entry point or control on the business page, adjusting filter options, etc. There are no specific limitations here. It's understood that after performing a page adjustment operation, a thumbnail window will display the adjusted business page. Based on this, users can directly process the business page on the intelligent interaction page without needing to return to the original business page to perform the adjustment and see the improved display effect.

[0047] Understandably, when displaying the adjusted business page in a thumbnail window, it's possible to obtain the page context data of the adjusted business page. This page context data can then be used to initiate commands for intelligent interaction on the adjusted business page. Based on this, simple adjustments to the business page can be quickly implemented on the intelligent interaction page without returning to the original business page and re-triggering the intelligent interaction controls, thus improving the convenience of intelligent interaction on the business page.

[0048] Optionally, in this application, the page adjustment operation may also include scaling the business page in the thumbnail window, which shrinks or enlarges the view of the business page in the thumbnail window so that users can more easily view the content displayed on the business page and avoid situations where some text is too small to see clearly or the whole page is difficult to see.

[0049] In one possible implementation, in response to a trigger operation on the smart interactive control, the embodiments of this application further include the following steps: 1. In response to a trigger operation on the smart interactive control, sending a smart interaction request for the business page to the server, so that when the server receives the smart interaction request, it obtains multiple signal data through multiple information sources; each information source is used to obtain the corresponding signal data; 2. Page context data for the business page constructed based on the multiple signal data.

[0050] Among them, the intelligent interaction request can be used to trigger a request for intelligent interaction. The intelligent interaction request can carry a page link of the business page, and then the server can determine the business page based on the intelligent interaction request.

[0051] The information source can be the source from which information is obtained. It is understood that in this application, there can be multiple information sources for obtaining page context data, so that multiple dimensions of signal data can be obtained to construct page context data.

[0052] The signal data acquired by each signal source may include any of the following: page links of the business page, page data of the business page, relevant search content of the business page, entity filtering information displayed on the business page, at least one candidate skill, and historical operation data of the target user.

[0053] In other words, page context data includes any of the following: page links of the business page, page data of the business page, relevant search content of the business page, entity filtering information displayed on the business page, at least one candidate skill, and historical operation data of the target user.

[0054] Here, "page link" can refer to a link to a business page or a route identifier. A business page link can be a unified access address pointing to a specific web resource, commonly known as a URL. A route identifier is a custom identifier used within the front-end application to map page components. It does not directly correspond to external resource addresses but is used in Single Page Applications (SPAs) to match the corresponding page components. It can be used to achieve page switching without page refresh, and it only takes effect within the application, not being directly exposed to external network access.

[0055] The page data for a business page can be data within the business page itself. This page data can include both visible content and backend business data (i.e., invisible backend data). Visible content refers to the content rendered and presented to the user by the business page (such as text, control states, selected filter items, etc.), which can be obtained through Document Object Model (DOM) parsing. DOM parsing converts HTML or XML documents into a manipulable tree-like DOM object structure, enabling operations such as reading and modifying document content. Backend business data refers to the structured business data that the business page relies on during runtime, which can be obtained through CLI (command line / interface call) methods.

[0056] The relevant search content for a business page can refer to content related to the current business page retrieved from the product knowledge base through search enhancement (RAG). Optionally, this relevant search content can be retrieved based on page links or the functional module to which the business page belongs. The product knowledge base is a structured intelligent information collection used to centrally store and manage all product-related knowledge. For example, if the business page is an analysis page for brand A, then at least search content for brand A can be retrieved. Based on this, the business page can be used to supplement the intelligent interaction system's understanding of the semantic boundaries and functional scope of the business page.

[0057] The entity filtering information displayed on the business page can refer to the set of entities formed by the filtering operation actively triggered by the user on the business page. For example, on a page analyzing products of a certain brand, if the filter displays product SP1 under brand A, then the entity filtering information displayed on the business page after filtering could be the specific brand and product selected. It is understandable that this type of signal originates from the user's active interaction and can serve as a highly reliable core reference signal for system or human decision-making.

[0058] Candidate skills refer to skills recommended and displayed on the current business page, also known as recommended skills. These skills are derived from the content displayed on the current business page. These candidate skills are triggerable on the intelligent interaction page. Here, "skills" can refer to reusable, standardized function packages within the intelligent interaction system. These can be processes and rules for specific tasks packaged into independent modules, allowing for direct invocation when needed without relearning each time. Examples include pre-packaged standardized capabilities like exporting worksheets or generating PPTs. Based on this, users can quickly trigger a task based on candidate skills without manually entering commands, improving the efficiency of intelligent interaction on business pages.

[0059] Target users refer to users associated with the webpage or client corresponding to the business page. Historical operation data can be the operation data of target users accumulated based on webpages or clients related to the business page. It can reflect the target users' operation preferences, identity information, and other information, and can be used to generate results that are more in line with user habits.

[0060] It is understood that the page context data may also include other types of related data, which are not limited here. The methods for executing intelligent interaction tasks and determining task execution progress based on the target instruction and the page context data of the business page can be found in subsequent descriptions, and will not be elaborated upon here.

[0061] It is understood that this application can collect multiple heterogeneous signal sources in a unified manner and fuse them according to a confidence-based hierarchical rule that prioritizes explicit operation signals (such as entity filtering information displayed on the business page), followed by page content (such as page links and page data of the business page) and document signals (i.e., relevant search content of the business page), then skill template candidate signals (i.e., at least one candidate skill), and finally user profile signals (i.e., historical operation data of the target user). This constructs unified structured page context data, which solves the problem that existing systems rely solely on single text input or single page routing mapping and cannot comprehensively utilize multi-dimensional information of the page to assist in task understanding, thus helping to improve the convenience during intelligent interaction.

[0062] S103. In response to a target instruction for a business page sent through the instruction input area, at least one intelligent agent executes the intelligent interactive task indicated by the target instruction, and displays the task execution progress corresponding to each of the at least one intelligent agent in the progress display area; the at least one intelligent agent is determined based on the target instruction and the page context data of the business page; the at least one intelligent agent includes an intelligent agent for performing page processing on the business page.

[0063] The target instruction can be an instruction initiated against a business page. It can be understood that this target instruction can be sent with reference to the business page displayed in a thumbnail window on the intelligent interaction page. Because of the reference to the business page displayed in the thumbnail window, the target instruction does not need to repeat the relevant content of the business page, and can use abbreviations or alternative names to make the target instruction more concise.

[0064] Optionally, in response to a target command sent to the business page through the command input area, a dialogue area and a progress display area can be displayed on the intelligent interaction page. The dialogue area can display the dialogue messages between the current user and the intelligent interaction system, such as a dialogue message sent by the user carrying the target command. The progress display area can be used to display the task execution progress.

[0065] Optionally, displaying the dialogue area and progress display area on the intelligent interaction page can hide the display of the thumbnail window and candidate skills, so as to avoid displaying too many elements on the intelligent interaction page and focus on the execution process of the intelligent interaction task.

[0066] Optionally, after hiding the thumbnail window, a window viewing control for viewing the thumbnail window can be displayed on the smart interaction page. Then, in response to a trigger operation on the window viewing control, the thumbnail window can be displayed on the smart interaction page. Based on this, users can bring up the thumbnail window at any time to view the business page, and thus initiate further instructions based on the business page.

[0067] Please see Figure 5 , Figure 5 This is a flowchart illustrating an instruction initiation process provided in an embodiment of this application. For example... Figure 5 As shown, page 511a can be an intelligent interactive page, which may include a command input area (i.e., area 501a). Then, command Z1 is entered in area 501a (step S51a). Command Z1 can be input via keyboard or voice, without limitation. Furthermore, the entered command Z1 can be displayed on area 501a, as shown in command 502a in Figure 5. Further, in response to a trigger operation on control 503a, message 504a is sent (step S52a). Message 504a can be a dialogue message carrying command Z1. It is understood that the display effect on the intelligent interactive page at this time can be seen in page 513a, which may include a dialogue area (i.e., area 506a) where message 504a can be displayed, and may also include a progress display area (i.e., area 505a) where the user can view the task execution progress corresponding to each intelligent agent.

[0068] The intelligent interaction task can be the task instructed by the target instruction. This intelligent interaction task can be the task obtained by parsing the target instruction; the specific parsing process will be described in subsequent related descriptions and will not be repeated here.

[0069] It is understandable that at least one subtask can be parsed for an intelligent interaction task. The method for determining at least one subtask will be described later and will not be repeated here. A subtask refers to the subtasks divided within the intelligent interaction task. Different subtasks can have sequential or parallel dependencies. A sequential dependency means that two subtasks are executed sequentially, i.e., one subtask is executed before the other. A parallel dependency means that two subtasks can be executed in parallel, i.e., there is no sequential order between them. For example, for a page used to analyze published content, if the target instruction is to analyze the published content, the intelligent interaction task of analyzing published content can be broken down into subtasks such as note data statistics, user comment sentiment analysis, and content tag extraction. After executing each subtask separately, the final result of the intelligent interaction task can be obtained.

[0070] Understandably, after determining at least one subtask included in the intelligent interaction task, a subtask list can be generated. Furthermore, the execution method for each subtask can be marked in the subtask list. The execution method of the subtask can include whether to generate inline processing operations on the business page, as well as the parallel / serial dependencies between the subtasks. Then, based on the subtask list, each subtask is assigned to each intelligent agent.

[0071] It is understandable that different subtasks can be assigned to at least one corresponding agent in an intelligent agent, so that the corresponding subtask can be executed by the agent.

[0072] Optionally, when executing a corresponding subtask through an agent, a corresponding intelligent execution instance can be created based on the agent to actually execute the subtask through the intelligent execution instance.

[0073] In this context, an intelligent execution instance can be an instance used to execute intelligent interaction tasks, also known as an agent instance. Essentially, an intelligent interaction system can identify at least one agent to execute intelligent interaction tasks, and then create intelligent execution instances corresponding to each agent. Simply put, an intelligent execution instance is a lightweight derivative of an intelligent object in a specific scenario, while an agent is a native entity with complete and independent capabilities. An intelligent execution instance can be an entity unit generated based on a unified agent definition, possessing its own independent operating environment. That is, an intelligent execution instance can be an independent execution instance generated from the same underlying model and the same role definition, targeting different sub-task perspectives. It can be used for multi-perspective parallel analysis under the same capability and responsibility. Furthermore, an intelligent execution instance can focus on the single-perspective execution of the corresponding sub-task without considering the overall task execution. It also has a short lifespan, starting and ending with the sub-task, and only allocates resources required for executing the sub-task, making its processing relatively lightweight.

[0074] It is understood that, in the embodiments of this application, a corresponding agent workspace can be allocated to each intelligent execution instance. The agent workspace can refer to the file storage area allocated to each intelligent execution instance. Each intelligent execution instance can write the intermediate outputs and execution results during the task execution process into its workspace in the form of files, which serves as a carrier for the intermediate state of interaction between the main control agent in the intelligent interaction system, replacing the reliance on a centralized database for state sharing.

[0075] This means that each intelligent execution instance can be used to execute subtasks within an intelligent interaction task. When executing a subtask, the progress of that subtask can be determined. This progress refers to the progress of the intelligent agent (specifically, an intelligent execution instance) in executing the corresponding subtask. Furthermore, this progress can be displayed on the intelligent interaction page, allowing users to better understand the current task's progress and better manage its execution.

[0076] The progress display area can be used to display progress. It is understood that in this application, the progress display area can show the task execution progress of each agent separately, rather than displaying a total task progress. This allows for a clearer display of the task execution progress of each agent, enabling the individual observation of the progress of all agents. This avoids collaborative chaos caused by information asymmetry, ensuring complete control over the overall task's pace. Furthermore, if an agent experiences a lag or error, it can be identified immediately, allowing users to quickly trigger fault recovery or task reassignment, significantly reducing the overall failure risk of intelligent interactive tasks and lowering the collaborative costs when multiple agents are working together.

[0077] The page context data can be relevant data relied upon for intelligent interaction based on the business page. For an introduction to page context data, please refer to the above description; it will not be repeated here. It is understood that this page context data can include relevant data from the business page. Compared to analysis relying solely on entered commands, this application can also rely on relevant data from the business page when processing commands, thereby reducing the communication cost when users input commands and improving the efficiency of intelligent interaction on the business page. This page context data can be determined by multi-dimensional page-related data, and it can be obtained from multiple signal sources.

[0078] In one possible implementation, in response to a target instruction for a business page sent through an instruction input area, executing the intelligent interactive task indicated by the target instruction through at least one intelligent agent may include the following steps: in response to the target instruction for a business page sent through the instruction input area, sending an instruction sending request carrying the target instruction to a server, so that when the server receives the instruction sending request, it determines at least one intelligent agent based on the intelligent collaborative processing strategy and the target instruction, creates an intelligent execution instance corresponding to each intelligent agent, and when it is determined to be at least one subtask in the intelligent interactive task, assigns a corresponding subtask to each intelligent execution instance, and executes the corresponding subtask through at least one intelligent execution instance, so as to realize the execution of the intelligent interactive task indicated by the target instruction through at least one intelligent agent.

[0079] Herein, the instruction sending request may be a request for sending a target instruction, and the instruction sending request may carry the target instruction. Further, the intelligent interaction system may parse the target instruction from the instruction sending request, and determine at least one expert agent as the at least one agent based on the target instruction and an intelligent cooperative processing strategy, where for the introduction of the expert agent, reference may be made to the relevant description above, and details are not described herein again. Herein, for each agent, a corresponding intelligent execution instance may be respectively generated to obtain at least one intelligent execution instance, and a subtask corresponding to the target instruction is executed by the at least one intelligent execution instance.

[0080] In a possible implementation, the intelligent interaction task includes at least one subtask, and each subtask is configured to be assigned to a corresponding agent in the at least one agent for execution; then, displaying the respective task execution progress of the at least one agent in the progress display area may include the following steps: receiving the task execution progress of each agent executing the corresponding subtask, and respectively displaying the respective task execution progress of the at least one agent in the progress display area.

[0081] It can be understood that for the introduction of the subtask, reference may be made to the relevant description above, and details are not described herein again.

[0082] It can be understood that each subtask may be assigned by the master agent in the intelligent interaction system to the corresponding agent of the at least one screened expert agent for processing the intelligent interaction task for execution, and then the respective task execution progress of each agent may be returned to the terminal device in real time.

[0083] Optionally, the task execution progress may include a step currently being executed, and may further include a percentage of the currently completed process in the total completion process (i.e., a progress percentage). Based on this, a user can more intuitively understand the details of the current task execution progress, so as to better grasp the task execution progress and conveniently take over the operation of the agent at any time.

[0084] It can be understood that the at least one agent includes an execution instance for performing page processing on a page object in a service page.

[0085] Herein, the page processing may refer to a process of performing internal linkage processing on the page object corresponding to the service page. Optionally, the page processing includes one or more of the following: adding a page object on the service page, modifying or querying a page object on the service page, and deleting a page object on the service page, which may be briefly referred to as add, delete, modify and query operations.

[0086] In essence, inline operations on objects involve calling built-in operation interfaces within the business system corresponding to the business page to adjust the page object. These built-in operation interfaces can refer to interfaces or tools within the business system that can process business pages; for example, they could be CLI tools or API interfaces. A CLI tool is a tool that interacts with a program through text commands. The user inputs specified commands via the keyboard, and the program receives and executes the corresponding operation. An API interface, or Application Programming Interface, is a set of predefined interaction conventions and rules, whose core function is to enable communication and capability reuse between different software and programs.

[0087] It is understood that this application can also provide an inline operation channel to enable the invocation of platform-built-in skills (such as CLI tools or API interfaces) to perform CRUD operations on business data objects, and to map the execution process of the skill invocation to the progress display area in real time. This allows intelligent objects to directly drive the backend to complete the actual query, addition, modification, and deletion operations (i.e., page processing) of page objects by calling the platform's existing built-in operation interfaces (such as CLI tools or API interfaces) corresponding to the business page. The backend data change notification mechanism automatically refreshes the frontend business page to display the latest results, while the intermediate execution state of each skill invocation is mapped to the intelligent interactive page in real time. Optionally, since the execution status is displayed when the skill is invoked to perform page processing, the user can take over control at any time during the execution process to pause or restart the corresponding execution process. Based on this, this application can overcome the technical limitations of existing read-only and offline interaction between intelligent agents and business systems without simulating page element interaction or relying on the specific frontend implementation structure of the business page, fundamentally solving the problem that intelligent agent analysis results cannot be automatically implemented into the business system.

[0088] In one possible implementation, the present application embodiment may further include the following steps: 1. On the intelligent interaction page, display at least one candidate skill for the business page; the at least one candidate skill is selected from the skill library based on the business page; 2. In response to a trigger operation for a target candidate skill among the at least one candidate skills, display the skill instruction template corresponding to the target candidate skill in the instruction input area; 3. In response to a parameter input operation for the skill instruction template, display an instruction containing the target parameter entered in the parameter input operation in the instruction input area.

[0089] The descriptions of the candidate skills can be found in the above descriptions, and will not be repeated here.

[0090] The skill library can refer to a collection of resources in an intelligent interaction system used to centrally store various skills. The at least one candidate skill is selected from the skill library based on the business page, which means that skills that can be applied on the business page can be selected.

[0091] Optionally, obtaining at least one candidate skill can be achieved by parsing the business scenario of the current business page, then filtering at least one skill from the skill library that matches the business scenario, and using the matched skill as at least one candidate skill. Here, the business scenario can refer to the application scenario of the business page, such as a brand needs analysis scenario or a product transaction analysis scenario, etc., without limitation, thus allowing for the matching of skills suitable for the business scenario.

[0092] The target candidate skill can be any of the at least one candidate skill.

[0093] The skill instruction template serves as a template for the instructions sent to the intelligent interaction system when executing a target candidate skill. It's understood that the skill instruction template can contain pre-configured template text, but it also includes adjustable parameters. Users can modify these parameters to issue instructions tailored to their specific needs based on the candidate skill. For example, if a skill is used to statistically analyze commodity transaction data within a specific time range, the corresponding skill instruction template could include a modifiable time range parameter, allowing users to manually edit the time range before issuing the instruction.

[0094] The parameter entry operation can be used to input parameters from a skill command template. For example, you can click on the parameter item in the skill command template, and then enter the command parameters according to your needs. The command input area will then display the command containing the target parameters entered in the parameter entry operation. The target parameters can be the parameters entered in the parameter entry operation.

[0095] Optionally, in this application, the candidate skill can be a feature-rich application template skill, which can be a fully packaged, practical business solution. For example, the application template skill could be a skill for finding competitor teams or analyzing product trends, and it could be a skill integrating multiple processes that users can directly access and use with one click. Alternatively, the candidate skill can be a reusable single-capability unit skill, such as knowledge Q&A or the ability to process different file formats. It is more of a low-level tool capability used to provide standardized operational support. Optionally, in this application, at least one recommended application template skill and at least one unit skill can be determined separately and then displayed on the intelligent interaction page, thereby improving the display of various types of candidate skills, helping users discover more skill uses, and improving the convenience of intelligent interaction.

[0096] For example, see Figure 6 , Figure 6 This is a flowchart illustrating a candidate skill triggering process provided in an embodiment of this application. For example... Figure 6 As shown, page 611a can be an intelligent interactive page, which can include at least one candidate skill. For example, the at least one candidate skill can be an application template skill (such as...). Figure 6 As shown in Template 1 or Template 2), it can also be used for unit skills (such as... Figure 6(As shown in Skill 1 or Skill 2). Taking Trigger Template 1 (i.e., Skill 601a) as an example, in response to a trigger operation for Skill 601a, Template 603a is displayed in area 602a (step S61a). After responding to a trigger operation for Skill 601a, a dialogue area and a progress display area can be displayed on the intelligent interaction page, as shown on page 612a. Page 612a may include area 602a, which can be a command input area. In this command input area, the skill command template (i.e., Template 603a) corresponding to Skill 601a can be displayed. Template 603a may include the parameter items required for Skill 601a, such as the time range shown in parameter 604a. Furthermore, in response to a parameter input operation for parameter 604a, the command after the parameter input is displayed (step S62a). The parameter input operation can input the parameter "Time T1". Based on this, the input parameter instruction can be "Please count the transaction data within time T1". Further, in response to the trigger operation of control 606a, message 607a is sent (step S63a). This message 607a can be a dialogue message carrying instruction 605a. Therefore, the intelligent interaction system can execute the task indicated by instruction 605a and simultaneously display the task execution progress in area 608a while executing the intelligent interaction task. The specific display method of the task execution progress can be found in the relevant description above and will not be repeated here. Based on this, instructions can be initiated by triggering candidate skills, reducing the amount of manual input required and improving the convenience of initiating instructions.

[0097] S104. When the intelligent interaction task is completed, the business-related page after the page processing of the business page is displayed in the progress display area.

[0098] For details on page processing, please refer to the relevant descriptions above; they will not be repeated here.

[0099] Optionally, when completing an intelligent interaction task, the system can also display the task execution result generated by the intelligent interaction system for the target instruction in the dialogue area of ​​the intelligent interaction page. This task execution result is used to provide feedback to the user in text form.

[0100] The business-related page can be a page used to process page objects on the business page. It can be understood that this business-related page can be displayed in a page window within the progress display area of ​​the intelligent interaction page, thus allowing the inline-processed intelligent interaction page to be displayed directly as a page.

[0101] Optionally, the business-related page can be the business page itself, or it can be a related page actually triggered based on the business page; there is no limitation here.

[0102] In one possible implementation, page processing of a business page is accomplished through a built-in operation interface in the business system corresponding to the business page; page processing includes one or more of the following: adding a page object to the business page, modifying or querying a page object on the business page, and deleting a page object on the business page; business-related pages include one or more of the following: the business page obtained after adjusting the business page, and the page queried after processing the business page.

[0103] The adjusted business page can be the result of adding, deleting, or modifying content on the original business page. Furthermore, after adjustment, the business page can be re-rendered and displayed in the progress display area. The original business page can also be re-rendered and displayed. For example, if a target instruction instructs data to be filtered on a business page according to specific conditions, then after completing the task indicated by that instruction, the business page with filtered data can be directly displayed in the progress display area.

[0104] The page retrieved after processing the business page can be a different page from the business page, and this business-related page can be triggered and viewed based on the business page. For example, in some scenarios, if the target instruction is to create a promotional audience that matches brand A on the business page, then after executing the intelligent interaction task indicated by the instruction, a page for viewing the created promotional audience can be directly displayed in the progress display area.

[0105] Optionally, when displaying business-related pages in the progress display area, the task execution results output by the intelligent interaction system may also include page links to the business-related pages, allowing users to access the corresponding business-related pages through these links.

[0106] It is understandable that in this application, since the intelligent objects' operations on the business page are all completed by calling the existing built-in operation interfaces of the business platform, rather than simulating the user interaction behavior of page elements, the intelligent interaction system does not need to be aware of or depend on the specific front-end implementation structure of the business page. It only needs to ensure that the identified operation interface is consistent with the business data layer of the business system, which has strong implementation decoupling and portability. Furthermore, the user can trigger takeover at any time during the execution of the inline operation (such as triggering a control to indicate that the task is paused). After responding to the takeover operation, the intelligent interaction system pauses the automatic execution of the current inline operation sequence and returns control of the business page to the user, who can then manually complete the subsequent operations on the business page. Optionally, after the user completes the manual operation, they can return control to the intelligent agent through interface operations (such as triggering a control to indicate that the task is continued). The intelligent interaction system can then resume the automatic execution of the inline operation sequence or directly write the results of the manually executed operations into the intelligent agent's workspace.

[0107] Using the embodiments of this application, a business page containing page objects can be displayed, and an intelligent interaction control for triggering intelligent interaction can be displayed on the business page. Then, in response to the triggering operation of the intelligent interaction control, an intelligent interaction page for the business page is displayed. The intelligent interaction page includes a thumbnail window for displaying the business page in a thumbnail format and an instruction input area for inputting instructions by referring to the business page in the thumbnail window. Furthermore, in response to the target instruction for the business page sent through the instruction input area, at least one intelligent agent executes the intelligent interaction task indicated by the target instruction, and the task execution progress corresponding to the at least one intelligent agent is displayed in the progress display area. The at least one intelligent agent is determined based on the target instruction and the page context data of the business page. The at least one intelligent agent includes an execution instance for performing page processing on the page objects in the business page. Further, when the intelligent interaction task is completed, the business-related page after performing page processing on the page objects in the business page is displayed in the progress display area. Based on this, intelligent interaction with smart objects can be triggered directly from the business page. When entering commands, commands can be initiated based on the business page shown in the thumbnail window of the reference smart interaction page, which helps to initiate commands to the business page more conveniently without having to manually send page screenshots for interaction. Then, the task execution progress of each instance is displayed separately when the task is executed through at least one smart agent. Furthermore, after processing the business page, adjustments can be directly applied to the business page without having to manually perform operations on the business page again, which can better facilitate business processing and thus improve the convenience of intelligent interaction with business pages.

[0108] Based on the above description, this application proposes an intelligent interaction method. Please refer to [link to relevant documentation]. Figure 7 , Figure 7 This is a flowchart illustrating an intelligent interaction method provided in an embodiment of this application. The method can be executed by the aforementioned computer device, such as a server (as described above). Figure 1 The server 100a in the diagram can be a server for a business system and / or an intelligent interaction system. The intelligent interaction method may include the following steps S201-S204.

[0109] S201. Receive a smart interaction request triggered by a terminal device on a business page, obtain page context data of the business page through multiple information sources, and return page data for the smart interaction page to the terminal device so as to display the smart interaction page on the terminal device; the smart interaction page includes a command input area for command entry.

[0110] The method for initiating a smart interaction request can be found in the relevant descriptions above, and will not be repeated here.

[0111] For an explanation of how to obtain the page context data of the business page, please refer to the relevant descriptions above, which will not be repeated here.

[0112] The method for displaying the intelligent interactive page can be found in the relevant descriptions above, and will not be repeated here.

[0113] S202. Upon receiving a target instruction for a business page sent through the instruction input area, at least one intelligent agent is determined based on the intelligent collaborative processing strategy and the target instruction; the at least one intelligent agent includes an intelligent agent for performing page processing on the business page.

[0114] The intelligent collaborative processing strategy can be a strategy for determining at least one agent for an intelligent interaction task. This intelligent collaborative strategy can be a strategy that coordinates multiple types of agents in the aforementioned intelligent interaction system to determine the agent performing the task. Specifically, it can be a strategy for determining at least one agent for the intelligent interaction task through the aforementioned master agent, capability-matching agent, and expert agent.

[0115] In one possible implementation, the page context data includes at least the relevant search content of the business page and the page data of the business page. Therefore, determining at least one intelligent agent based on the intelligent collaborative processing strategy and the target instruction may include the following steps: 1. Based on the page data and relevant search content of the business page, optimize the instruction text corresponding to the target instruction to obtain the task instruction text, and determine the intelligent interaction task based on the task instruction text; 2. Search for a set of expert intelligent agents matching the task instruction text from a preset expert intelligent agent set library through the master intelligent agent; 3. If a target expert intelligent agent set matching the task instruction text is found, use at least one expert intelligent agent in the target expert intelligent agent set as at least one intelligent agent for the intelligent interaction task indicated by the target instruction.

[0116] The instruction text refers to the text within the target instruction, which can be used to instruct what operation to perform on the business page. For details regarding the page data and related search content for the business page, please refer to the above descriptions; they will not be repeated here.

[0117] Text optimization can be a process of filling in and restoring referential content and omitted expressions in the instruction text. The task instruction text can be the optimized text representing the actual determined intelligent interaction task.

[0118] For example, if the instruction text contains unclear terms such as "this brand," "current target audience," or "filtered time period," its specific meaning on the business page can be determined based on the page data and related search content of the business page, and then replaced with the actual specific content. For instance, if the instruction text is "this brand," but the page data of the business page determines that "this brand" refers to brand A, then "this brand" in the instruction text can be replaced with brand A.

[0119] Optionally, during text optimization, the task instruction text can be restored by layering the data according to the confidence levels of three types of signals: explicit entity references (i.e., explicitEntityRefs), template entity references (i.e., templateEntityRefs), and user profile context (i.e., userProfileContext). Explicit entity references refer to entities clearly marked and unambiguous in the user input, such as a user directly stating a complete order number. Template entity references can be entities from preset dialogue templates or historical templates, such as the system's preset list of available devices or fixed-time templates. User profile context can be set with entities extracted from the user's historical profile and long-term behavioral habits, such as the user's frequently used delivery addresses or travel preferences.

[0120] The preset expert agent set library is a database of preset expert agent sets in the intelligent interaction system. An expert agent set can be a collection of multiple agents that can work together to complete a task, also known as an expert team. When searching for an expert agent set that matches the task instruction text from the preset expert agent set library, it can be done through semantic matching based on the task instruction text and the relevant descriptive information of each expert agent set in the preset expert agent set library. Therefore, in this application, when matching expert teams, the task execution instruction used is optimized text, resulting in more accurate expert team matching and avoiding inaccuracies.

[0121] Optionally, when searching for a set of expert agents matching the task instruction text from a pre-set expert agent set library by the master intelligent agent, at least one candidate expert agent set can be initially selected from the pre-set expert agent set library based on the page data and relevant search content of the business page. Then, the expert agent set matching the task instruction text can be searched from the at least one candidate expert agent set. The initial selection of at least one candidate expert agent set from the pre-set expert agent set library can filter out expert agents that the current business page completely lacks the necessary processing permissions and capabilities, thereby reducing the computational load required for subsequent expert agent set matching and improving the efficiency of expert agent set matching.

[0122] Optionally, the target set of expert agents can be a set used for matching with expert agents. Once the target set of expert agents is found, the agents in the target set can be directly used as at least one expert agent for the intelligent interaction task indicated by the target instruction, to execute the intelligent interaction task.

[0123] In one possible implementation, the embodiments of this application may further include the following steps: if no target set of expert agents matching the task instruction text is found, the capability matching agent filters at least one expert agent matching the intelligent interaction task from the expert agent library based on the page data of the business page, and the filtered at least one expert agent is used as at least one agent.

[0124] For an introduction to capability-matching agents, please refer to the relevant descriptions above, which will not be repeated here.

[0125] Among them, the expert agent library can refer to the database of expert agents in the intelligent interaction system, which may include multiple expert agents.

[0126] Optionally, when selecting at least one expert agent from the expert agent library that matches the intelligent interaction task through the capability matching agent, the intelligent interaction system can generate a list of capability requirements for the intelligent interaction task, and then match an expert agent from the expert agent library that can provide each capability in the corresponding capability requirement list. If an expert agent that can provide the corresponding capability is successfully matched, that expert agent object can be used as the expert agent to perform the intelligent interaction task, and an agent can be created for it.

[0127] Optionally, if no expert agent capable of providing the corresponding capabilities is successfully matched, a general template agent can be used. The required capabilities, role definitions, and task instructions for the general template agent are then determined to generate a new agent. After the generated agent completes the intelligent interaction task, the results can be verified and evaluated. If the evaluation indicates good task performance, the agent is designated as an expert agent; otherwise, if the evaluation indicates poor task performance, the agent is discarded.

[0128] It is understood that in this application, when matching expert agents based on capability-matching agents, the matching can be at the capability level, that is, determining whether the capabilities of the expert agent can meet the requirements of the intelligent interaction task. Furthermore, the capability-matching agent can rely on page data from the business page for matching, and the role definitions and task instructions for injecting the required capabilities into the general template agent can also be generated based on the page data from the business page. This allows the accuracy of capability matching and agent generation by the capability-matching agent to be superior to existing methods that rely solely on semantic matching based on instruction text, thus helping to improve the accuracy of agent matching and generation.

[0129] It is understandable that in the multi-level intelligent agent matching intelligent collaborative processing strategy adopted in this application, the aforementioned page context data with multi-dimensional data can be used as an auxiliary judgment basis for the team matching stage (i.e., the master intelligent agent matching the expert intelligent agent set) and a supplementary judgment basis for the capability matching stage (i.e., the capability matching intelligent agent matching the expert intelligent agent). At the same time, it can also serve as the basis for optimizing the instruction text. This solves the problem of expression redundancy and semantic ambiguity caused by the existing multi-agent system team matching relying solely on text description and being disconnected from the business page context. As a result, the accuracy and efficiency of determining the intelligent agent used to perform intelligent interaction tasks can be improved.

[0130] In one possible implementation, executing an intelligent interactive task indicated by a target instruction through at least one intelligent agent may include the following steps: 1. Determining at least one subtask contained in the intelligent interactive task based on the task instruction text, and generating a subtask list containing at least one subtask; the subtask list is marked with the task dependency relationship and page processing mark corresponding to each subtask; 2. Assigning each subtask to the corresponding intelligent agent in the at least one intelligent agent according to the subtask list through the master intelligent agent, and executing the corresponding subtask through the at least one intelligent agent.

[0131] For details regarding subtasks and subtask lists, please refer to the aforementioned descriptions; they will not be repeated here. Furthermore, the master control agent can allocate and schedule tasks according to the subtask list. For example, it can use the master control agent to sequentially call agents to execute tasks based on the dependencies of subtasks in the subtask list. It can also determine whether page processing is required based on page processing markers in the subtask list. These page processing markers indicate whether a corresponding subtask requires page processing. Essentially, if a subtask has a page processing marker, it indicates that the subtask requires page processing, and the agent assigned to execute that subtask is the execution instance used to process the page objects in the business page.

[0132] For example, see Figure 8 , Figure 8 This is a schematic diagram of a process for determining an intelligent agent provided in an embodiment of this application. Figure 8 As shown, firstly, the instruction text in the target instruction can be obtained (such as...). Figure 8 As shown in S801a), then based on the page data of the business page and the relevant search content of the business page, the text of the instruction corresponding to the target instruction is optimized to obtain the task instruction text (such as S801a). Figure 8 (S802a shown). The master intelligent agent searches for a set of expert intelligent agents (e.g., S802a) that matches the task instruction text from a preset expert intelligent agent set library. Figure 8 As shown in S803a), then, it is determined whether it has been found (e.g. Figure 8 (S804a shown). If no set of expert agents matching the task instruction text is found, at least one expert agent matching the intelligent interaction task is selected from the expert agent library through the capability matching agent (e.g., ...). Figure 8 As shown in S805a), then create corresponding intelligent execution instances (e.g., for each of the at least one matched expert agent) Figure 8 (S806a shown). If a set of expert agents matching the task instruction text is found, then a corresponding intelligent execution instance is created for at least one expert agent in the matched target expert agent set (e.g., ...). Figure 8 (S807a shown). Further, a subtask list containing at least one subtask is generated based on the task instruction text (e.g., ...). Figure 8 As shown in S808a); the master intelligent agent assigns each subtask to the corresponding intelligent agent's intelligent execution instance (e.g., according to the subtask list) based on the subtask list. Figure 8 (S809a shown).

[0133] S203. The intelligent interactive task indicated by the target instruction is executed by at least one intelligent agent, and the task execution progress corresponding to each of the at least one intelligent agent is returned to the terminal device so as to display the task execution progress corresponding to each of the at least one intelligent agent in the progress display area.

[0134] The method for displaying task execution progress can be found in the relevant descriptions above, and will not be repeated here.

[0135] S204. When the intelligent interaction task is completed, the page data of the business-related page after the page processing of the business page is returned to the terminal device so as to display the business-related page in the progress display area.

[0136] The method for displaying business-related pages can be found in the relevant descriptions above, and will not be repeated here.

[0137] Optionally, when performing intelligent interaction tasks, if the built-in operation interface call corresponding to page processing is completed, the agent used to perform page processing can write the output file of this page processing process to a specified path in its agent workspace. The output file may include, but is not limited to, the input parameters, execution results, and intermediate output content of the built-in operation interface call. Specifically, it can be written in the form of a structured file to record the complete credentials of this built-in operation interface call. Furthermore, the agent can send a task completion report to the master agent. The report content may include a summary of the subtask execution results obtained by executing the corresponding subtask and the storage path of the above-mentioned output file in the agent workspace.

[0138] Furthermore, the master control agent (or other components in the intelligent interaction system responsible for updating the context) can read the output file based on the file path of the output file in the task completion report to obtain the impact of this content operation interface call on the page objects on the corresponding business page. Combined with the latest DOM / CLI content and visible entity filtering status of the affected page that has been re-collected, the updated context data is generated. The difference (diff) between the original page context data and the output file in the agent's workspace is a comparison between the output file and the results of the re-collection of the business page.

[0139] Optionally, for subtasks that do not require page processing, the agent can write the task execution result to the agent's workspace in the form of a file after the subtask is completed, and submit it to the master agent in the form of a task completion report. The task completion report may include a summary of the subtask's task execution result and the file path of the output file. It is understood that since the subtask does not involve page processing, the update of the page data in the page context data will not be triggered at this time.

[0140] Understandably, since the intermediate state interactions between various agents in the intelligent interaction system are all achieved by writing / reading files in the agent's workspace, rather than relying on centralized database transactions or distributed lock mechanisms for state sharing and concurrency control, this application can be implemented in an environment without a centralized database. It has low deployment dependency and high module independence, significantly reducing the overall deployment dependency threshold of the system and improving the independence between functional modules. It effectively avoids the single point of failure of centralized databases, the scheduling overhead and deadlock risk caused by distributed locks.

[0141] It is understandable that the master intelligent object (the component responsible for updating the context in the intelligent interaction system) generates the updated context data by processing the object content of a certain intelligent agent. The updated context data can be saved as the page context data of the current dialogue with the intelligent interaction system to replace the original page context data in the dialogue.

[0142] Optionally, if there are other unfinished sub-tasks in the current intelligent interaction task, the master intelligent agent can integrate the execution results and operation changes of all intelligent agents to generate the task execution result of this round of intelligent interaction task and present it to the user.

[0143] Understandably, on the current intelligent interaction page, if the user continues to input new task instructions within the same dialogue (such as asking follow-up questions about the result of the previous task, or new requirements related to the result of the previous task), the system will use the updated page context data (rather than the old page context data before the start of this round of tasks) as the page context data for the new instruction to execute the intelligent interaction task for that new instruction. The specific task execution process can be found in the above-described flow of executing intelligent interaction tasks, and will not be repeated here. Since the updated context data reflects the latest impact of this round of operations on the page objects of the business page, the user does not need to repeatedly describe the latest state of the business page in subsequent rounds. The intelligent interaction system can correctly understand and process subsequent tasks based on the latest state, which helps improve the convenience of intelligent interaction.

[0144] Understandably, by updating this context data, a cyclical mechanism of "perception-execution-writeback-preservation" can be implemented on the business page. The page context data serves as the input for team matching and task orchestration in intelligent interaction tasks, and can also be updated after page processing. This new page context data then becomes the input for subsequent rounds of team matching and task orchestration within the same session, ensuring that multiple rounds of continuous interaction within the same session are based on the latest business state and avoiding misunderstandings caused by context lag. Based on this, the problem of existing systems where the execution results of expert agents are disconnected from the business page context, leading to subsequent interactions within the same session failing to perceive the latest business state and requiring users to repeatedly describe changed information is solved.

[0145] Using the embodiments of this application, intelligent interaction with smart objects can be directly triggered based on the business page. When entering instructions, instructions can be initiated based on the business page shown in the thumbnail window of the reference smart interaction page, which helps to initiate instructions to the business page more conveniently without having to manually send page screenshots for interaction. Then, the task execution progress corresponding to each instance is displayed according to the task being executed by at least one smart agent. Furthermore, after processing the business page, adjustments can be directly applied to the business page without having to manually perform operations on the business page, which can better perform business processing and thus help improve the convenience of intelligent interaction with the business page.

[0146] Please see Figure 9 , Figure 9 This is a timing interaction diagram of an intelligent interaction method provided in an embodiment of this application. The intelligent interaction method involves a terminal device, a business system (specifically, a server of the business system), and an intelligent interaction system (specifically, a server of the intelligent interaction system). The server of the business system and the server of the intelligent interaction system can be the same or different; this is not limited here. The intelligent interaction method may include the following steps: S301. The terminal device displays a service page and displays intelligent interaction controls on the service page to trigger intelligent interaction. S302, The terminal device responds to the trigger operation on the smart interactive control by generating a smart interactive request; S303, The terminal device sends an intelligent interaction request to the business system; S304. The business system obtains the page context data of the business page through multiple information sources and determines the page data for the intelligent interactive page. S305. The business system returns page data of the intelligent interactive page to the terminal device; S306. Terminal devices display page data based on intelligent interactive pages; S307. The terminal device responds to the target instruction for the business page sent through the instruction input area and generates an instruction sending request; S308, The terminal device sends instructions and requests to the intelligent interactive system; S309. The intelligent interaction system determines at least one intelligent agent based on intelligent collaborative processing strategies and target instructions; the at least one intelligent agent includes an intelligent agent for page processing of business pages; S310. The intelligent interaction system executes the intelligent interaction task indicated by the target instruction through at least one intelligent agent, and determines the task execution progress corresponding to each of the at least one intelligent agent. S311. The intelligent interaction system returns the task execution progress of at least one intelligent agent to the terminal device. S312. The terminal device displays the task execution progress of at least one intelligent agent in the progress display area; S313. When the intelligent interaction system completes the intelligent interaction task, it determines the page data of the business-related pages after the business page has been processed. S314. The intelligent interactive system returns page data of the business-related page to the terminal device; S315. The terminal device displays the business-related page in the progress display area.

[0147] Using the embodiments of this application, intelligent interaction with smart objects can be directly triggered based on the business page. When entering instructions, instructions can be initiated based on the business page shown in the thumbnail window of the reference smart interaction page, which helps to initiate instructions to the business page more conveniently without having to manually send page screenshots for interaction. Then, the task execution progress corresponding to each instance is displayed according to the task being executed by at least one smart agent. Furthermore, after processing the business page, adjustments can be directly applied to the business page without having to manually perform operations on the business page, which can better perform business processing and thus help improve the convenience of intelligent interaction with the business page.

[0148] Please see Figure 10 , Figure 10 This is a flowchart illustrating an intelligent interaction method provided in an embodiment of this application. The method can be executed by the aforementioned computer device, such as a terminal device (as described above). Figure 1 (Device 11a in the example). The intelligent interaction method may include the following steps S401-S402.

[0149] S401. Display the target session page, which includes a session input area.

[0150] The target session page can include a page used for intelligent dialogue with the intelligent interaction system. This target session page can be the intelligent interaction page triggered by the aforementioned business page, or it can be any session page used for dialogue with the intelligent interaction system. It can be triggered by a session entry point on a page different from the business page. Dialogue with the intelligent interaction system is equivalent to engaging in intelligent interaction.

[0151] The conversation input area can be used to input the content of the conversation with the intelligent interaction system. It is understood that if the target conversation page is the aforementioned intelligent interaction page for business purposes, then the conversation input area on that page can be the command input area within the aforementioned intelligent interaction page. If the target conversation page is any conversation page for engaging with the intelligent interaction system, then the display method of the conversation input area on that page can refer to the command input area within the aforementioned intelligent interaction page.

[0152] S402, in response to sending target session content through the session input area, execute instructions corresponding to the first session content based on the first agent, and execute instructions corresponding to the second session content based on the second agent; wherein, the target session content includes first identification information of the first agent and first session content for the first agent, and second identification information of the second agent and second session content for the second agent; wherein, after the target character is entered, the session input area triggers the display of an agent list, and the target agent is selected from the agent list and the target identification information of the target agent is entered in the session input area; the agent list includes at least two agents, and the target agent is any one of the at least two agents.

[0153] The target session content can be session content used to instruct multiple intelligent agents in an intelligent interaction system to execute corresponding instructions.

[0154] It is understood that the multiple intelligent agents here can be two or more. This example uses two intelligent agents as an illustration; for instance, these multiple intelligent agents can be a first intelligent agent and a second intelligent agent. Optionally, the first intelligent agent and the second intelligent agent can be expert intelligent agents (i.e., expert agents) in an intelligent interaction system.

[0155] The first identification information can be information used to identify the first intelligent agent, for example, the first identification information can be the name of the first intelligent agent. The second identification information can be information used to identify the second intelligent agent, for example, the second identification information can be the name of the second intelligent agent.

[0156] The first session content can be the content of instructions used to instruct the first intelligent agent to execute. The second session content can be the content of instructions used to instruct the second intelligent agent to execute.

[0157] It is understandable that when executing instructions corresponding to the content of the first session based on the first intelligent agent, the first interactive task can be determined through the instructions corresponding to the content of the first session, and then the corresponding first interactive task can be executed through the first intelligent agent.

[0158] Optionally, when executing instructions corresponding to the first session content based on the first intelligent agent, the intelligent agent corresponding to the first intelligent agent can be created by the intelligent interaction system, and then the instructions corresponding to the first session content can be executed through the intelligent agent corresponding to the first intelligent agent.

[0159] Similarly, the method of executing instructions corresponding to the second session content based on the second agent can be found in the above description of executing instructions corresponding to the first session content based on the first agent, and will not be repeated here.

[0160] Optionally, when initiating target session content to the intelligent interaction system, a dialogue area and a progress display area can be displayed on the target session page. Furthermore, the target session content can be displayed in the dialogue area, and the task execution progress corresponding to the first intelligent agent and the task execution progress corresponding to the second intelligent agent can be displayed in the progress display area. Here, the task execution progress can refer to the progress of executing the instructions corresponding to the session content (such as the first session content or the second session content).

[0161] It is understandable that if the target session content includes the session content of at least two agents, then when displaying the task execution progress, the task execution progress of each of the at least two agents can be displayed separately. The method for displaying the task execution progress of the at least two agents can be found in the description above regarding the display of the task execution progress of the corresponding agent in the progress display area of ​​the intelligent interaction page, and will not be repeated here.

[0162] It is understandable that after completing the instructions corresponding to the first and second intelligent agents respectively, the results of the instruction execution of the first and second intelligent agents can be summarized to output a conversation response to the target conversation content, which can then be displayed in the dialogue area of ​​the target conversation page.

[0163] The target character can be a character used to trigger the selection of the agent; for example, the target character can be "@".

[0164] The list of intelligent agents can be selected intelligent agents in the intelligent interaction system. The list of intelligent agents can include intelligent agent information of at least two intelligent agents. The intelligent agent information refers to the relevant information of the intelligent agents. For example, the intelligent agent information can include the intelligent agent's avatar, name, function description, etc., which are not limited here.

[0165] Optionally, the agents in this agent list can be determined based on the current user's historical interaction agents and the agents recommended by the intelligent interaction system. Historical interaction agents can be agents used by the current user. Recommended agents can be agents recommended by the intelligent interaction system, and these recommended agents can be determined based on the number of times each user uses agents within a preset time range, for example, agents that have been used a preset number of times.

[0166] The target agent can be an agent from the agent list. The target identification information can be the identification information of the target agent, such as the name of the target agent. It can be understood that the target agent can be the first agent and the second agent mentioned above. That is to say, the method of entering the first identification information and conversation content for the first agent and the second identification information and conversation content for the second agent in the conversation input area can be found in the relevant description of entering the target identification information and conversation content for the target agent.

[0167] This section provides a flowchart illustrating the process of entering conversational content targeting the agent. Please refer to [link / reference]. Figure 11a , Figure 11a This is a schematic diagram illustrating a process for entering conversation content, as provided in an embodiment of this application. Figure 11aAs shown, page 1111a can be a target session page, and this page 1111a can include a session input area (i.e., area 1101a). Furthermore, in response to the input of character 1102a in area 1101a, list 1103a is displayed (step S111a). Character 1102a can be the target character mentioned above, for example, the target character is "@". List 1103a can be a list of agents, which can include multiple agents, such as agent 1, agent 2, and agent 3. Further, in response to the selection operation of agent 1, information 1104a of agent 1 is input in area 1101a (step S112a). Information 1104a can be the identification information corresponding to agent 1 (such as first identification information). Further, session content N1 for agent 1 is entered (step S113a). The session content N1 (i.e., content 1105a) can be session content (such as the first session content) used to instruct agent 1 to execute instructions. Furthermore, in response to the sending operation of the session content displayed in area 1101a, the session content is sent to the intelligent interaction system so that the intelligent interaction system can execute the instructions corresponding to session content N1 based on agent 1.

[0168] For further details, please see Figure 11b , Figure 11b This is a schematic diagram illustrating a process for entering conversation content, as provided in an embodiment of this application. Figure 11b As shown, page 1111b can be a target session page, and page 1111b can include a session input area (i.e., area 1101b). Area 1101b can include data that has been processed through the above-mentioned... Figure 11aThe conversation content for agent 1 (i.e., content 1102b) is entered in the following manner. Then, in response to the character 1103b being entered in area 1101b, list 1104b is displayed (step S111b). Character 1103b can be the target character mentioned above, for example, the target character is "@". List 1104b can be a list of agents, which can include multiple agents, such as agent 1, agent 2, and agent 3. Further, in response to the selection operation of agent 2, information 1105b of agent 2 is entered in area 1101b (step S112b). Information 1105b can be the identification information corresponding to agent 2 (such as second identification information). Further, conversation content N2 for agent 2 is entered (step S113b). This conversation content N2 (i.e., content 1106b) can be conversation content used to instruct agent 2 to execute instructions (such as second conversation content). Furthermore, in response to the sending operation of the session content displayed in area 1101b, the session content is sent to the intelligent interaction system so that the intelligent interaction system can execute the instruction corresponding to session content N1 based on agent 1, and execute the instruction corresponding to session content N2 based on agent 2.

[0169] Using the embodiments of this application, intelligent agents can be selected by inputting characters on the conversation page. Furthermore, multiple intelligent agents can be selected in one conversation, and conversation content can be input for each intelligent agent to execute the corresponding instructions of the conversation content. Thus, multiple intelligent agents can be triggered in one operation, and different instructions can be entered for different intelligent agents, thereby improving the flexibility and convenience of intelligent agent triggering.

[0170] Please see Figure 12 , Figure 12 This is a schematic diagram of the structure of an intelligent interactive device provided in an embodiment of this application. Figure 12 As shown, the intelligent interactive device 1 can be a computer program (including program code) running on a computer device (such as the aforementioned terminal device, such as device 11a). For example, the intelligent interactive device 1 is an application software. It is understood that the intelligent interactive device 1 can be used to execute the corresponding steps in the data processing method provided in the embodiments of this application. Figure 12 As shown, the intelligent interactive device 1 may include: a business page display module 11, an intelligent interactive page module 12, a task execution module 13, and a task completion module 14; The business page display module 11 is used to display a business page and display intelligent interactive controls on the business page for triggering intelligent interaction; The intelligent interactive page module 12 is used to display an intelligent interactive page for the business page in response to a trigger operation on the intelligent interactive control; the intelligent interactive page includes an instruction input area; The task execution module 13 is configured to respond to a target instruction sent through the instruction input area for the business page, execute the intelligent interactive task indicated by the target instruction through at least one intelligent agent, and display the task execution progress corresponding to the at least one intelligent agent in the progress display area; the at least one intelligent agent is determined based on the target instruction and the page context data of the business page; the at least one intelligent agent includes an intelligent agent for page processing of the business page; The task completion module 14 is used to display the business-related page after page processing of the business page on the progress display area when the intelligent interactive task is completed.

[0171] In one implementation, the intelligent interactive page module 12 is also used for: On the intelligent interaction page, at least one candidate skill is displayed for the business page; the at least one candidate skill is selected from the skill library based on the business page. In response to a trigger operation targeting a target candidate skill among the at least one candidate skills, a skill instruction template corresponding to the target candidate skill is displayed in the instruction input area; In response to a parameter entry operation for a skill instruction template, an instruction containing the target parameters entered in the parameter entry operation is displayed in the instruction input area.

[0172] In one implementation, the intelligent interactive page module 12 is also used for: A thumbnail window is displayed on the intelligent interaction page to show a simplified version of the business page; the business page in the thumbnail window is used as a reference for command input. In response to a page adjustment operation performed on a business page in the thumbnail window, the business page after the page adjustment operation is performed is displayed in the thumbnail window.

[0173] In one implementation, page processing of the business page is accomplished through a built-in operation interface in the business system corresponding to the business page. The page processing includes one or more of the following: adding page objects to the business page, modifying or querying page objects on the business page, and deleting page objects on the business page; The business-related pages include one or more of the following: business pages obtained after adjusting the business pages, and pages retrieved after processing the business pages.

[0174] In one implementation, the page context data includes any of the following: page links of the business page, page data of the business page, relevant search content of the business page, entity filtering information displayed on the business page, at least one candidate skill, and historical operation data of the target user; In one implementation, the intelligent interactive page module 12 is also used for: In response to a trigger operation on the smart interactive control, a smart interaction request for the business page is sent to the server, so that when the server receives the smart interaction request, it obtains multiple signal data through multiple information sources; each information source is used to obtain the corresponding signal data. Page context data for the business page is constructed based on the multiple signal data.

[0175] In one implementation, task execution module 13 is further used for: In response to a target instruction for the business page sent through the instruction input area, an instruction sending request carrying the target instruction is sent to the server. Upon receiving the instruction sending request, the server determines at least one intelligent agent based on the intelligent collaborative processing strategy and the target instruction, creates an intelligent execution instance corresponding to each intelligent agent, and assigns a corresponding sub-task to each intelligent execution instance when it is determined to be at least one sub-task in the intelligent interaction task. The server then executes the corresponding sub-task through the at least one intelligent execution instance to realize the execution of the intelligent interaction task indicated by the target instruction through at least one intelligent agent.

[0176] In one implementation, the intelligent interaction task includes at least one subtask, each subtask being assigned to a corresponding intelligent agent among the at least one intelligent agents for execution; Task execution module 13 is also used for: The task execution progress of each agent in executing the corresponding subtask is received, and the task execution progress of each of the at least one agent is displayed in the progress display area.

[0177] Please see Figure 13 , Figure 13 This is a schematic diagram of the structure of an intelligent interactive device provided in an embodiment of this application. Figure 13 As shown, the intelligent interactive device 2 can be a computer program (including program code) running on a computer device (such as the server mentioned above), for example, the intelligent interactive device 2 is an application software; it is understood that the intelligent interactive device 2 can be used to execute the corresponding steps in the data processing method provided in the embodiments of this application. Figure 13As shown, the intelligent interaction device 2 may include: an intelligent interaction request module 21, an intelligent agent determination module 22, a task processing module 23, and a completion feedback module 24; The intelligent interaction request module 21 is used to receive intelligent interaction requests triggered by the terminal device based on the business page, obtain the page context data of the business page through multiple information sources, and return the page data for the intelligent interaction page to the terminal device so as to display the intelligent interaction page on the terminal device; the intelligent interaction page includes an instruction input area for instruction input; The agent determination module 22 is used to determine at least one agent based on the intelligent collaborative processing strategy and the target instruction when receiving a target instruction for the business page sent through the instruction input area; the at least one agent includes an agent for page processing of the business page. Task processing module 23 is used to execute the intelligent interactive task indicated by the target instruction through the at least one intelligent agent, and return the task execution progress corresponding to the at least one intelligent agent to the terminal device, so as to display the task execution progress corresponding to the at least one intelligent agent in the progress display area; The feedback module 24 is used to return page data of the business-related page after page processing of the business page to the terminal device when the intelligent interaction task is completed, so as to display the business-related page on the progress display area.

[0178] In one implementation, the page context data includes at least the relevant search content of the business page and the page data of the business page; The agent determination module 22 is also used for: Based on the page data of the business page and the relevant search content of the business page, the text of the instruction text corresponding to the target instruction is optimized to obtain the task instruction text, and the intelligent interactive task is determined based on the task instruction text; The master control agent searches for a set of expert agents that match the task instruction text from a preset set of expert agents. If a set of target expert agents that matches the task instruction text is found, at least one expert agent in the set of target expert agents is used as at least one agent for the intelligent interactive task indicated by the target instruction.

[0179] In one implementation, the agent determination module 22 is further configured to: If no target set of expert agents matching the task instruction text is found, the capability matching agent filters at least one expert agent matching the intelligent interaction task from the expert agent library based on the page data of the business page, and the filtered at least one expert agent is taken as the at least one agent.

[0180] In one implementation, the task processing module 23 is further used for: Based on the task instruction text, at least one subtask is determined to be included in the intelligent interaction task, and a subtask list containing the at least one subtask is generated; the subtask list is marked with the task dependency relationship and page processing mark corresponding to each subtask. The master control agent assigns each subtask to the corresponding agent in the at least one agent according to the subtask list, and the at least one agent executes the corresponding subtask.

[0181] Please see Figure 14 , Figure 14 This is a schematic diagram of the structure of an intelligent interactive device provided in an embodiment of this application. Figure 14 As shown, the intelligent interactive device 3 can be a computer program (including program code) running on a computer device (such as the aforementioned terminal device, such as device 11a). For example, the intelligent interactive device 3 is an application software. It is understood that the intelligent interactive device 3 can be used to execute the corresponding steps in the data processing method provided in the embodiments of this application. Figure 14 As shown, the intelligent interactive device 3 may include: a conversation page display module 31 and an instruction execution module 32; The conversation page display module 31 is used to display the target conversation page, which includes a conversation input area; The instruction execution module 32 is used to respond to sending target session content through the session input area, execute instructions corresponding to the first session content based on the first intelligent agent, and execute instructions corresponding to the second session content based on the second intelligent agent; wherein, the target session content includes first identification information of the first intelligent agent and first session content for the first intelligent agent, and second identification information of the second intelligent agent and second session content for the second intelligent agent; The conversation input area triggers the display of an agent list after the target character is entered. The target agent is selected from the agent list and the target identification information of the target agent is entered in the conversation input area. The agent list includes at least two agents, and the target agent is any one of the at least two agents.

[0182] Please see Figure 15 , Figure 15This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. For example, this computer device can be the terminal device or server described above. Figure 15 As shown, the computer device 1000 may include a processor 1001, a network interface 1004, and a memory 1005. Furthermore, the computer device 1000 may also include a user interface 1003 and at least one communication bus 1002. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface). The memory 1005 may be high-speed RAM or non-volatile memory, such as at least one disk storage device. Optionally, the memory 1005 may also be at least one storage device located remotely from the processor 1001. Figure 15 As shown, the memory 1005, which is a computer-readable storage medium, may include an operating system, a network communication module, a user interface module, and a device control application.

[0183] In such Figure 15 In the computer device 1000 shown, the network interface 1004 provides network communication functionality; the user interface 1003 is mainly used to provide an input interface for the user; and the processor 1001 can be used to call the device control application stored in the memory 1005 to execute the data processing method described in any of the corresponding embodiments above, which will not be repeated here. Furthermore, the beneficial effects of using the same method will also not be repeated.

[0184] Furthermore, it should be noted that this application also provides a computer-readable storage medium storing a computer program adapted to be loaded and executed by a processor, so that a computer device having a processor performs the methods described in the preceding embodiments. Therefore, these will not be repeated here. Additionally, the beneficial effects of using the same method will also not be repeated. For technical details not disclosed in the embodiments of the computer-readable storage medium involved in this application, please refer to the description of the method embodiments of this application.

[0185] The aforementioned computer-readable storage medium can be the apparatus provided in any of the foregoing embodiments or the internal storage unit of the aforementioned computer device, such as the hard disk or memory of the computer device. The computer-readable storage medium can also be an external storage device of the computer device, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., provided on the computer device. Furthermore, the computer-readable storage medium may include both internal storage units and external storage devices of the computer device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0186] Furthermore, it should be noted that this application also provides a computer program product, which includes a computer program stored in a computer-readable storage medium. The computer program is adapted to be read and executed by a processor, so that a computer device having a processor performs the method provided in any of the corresponding embodiments described above. Additionally, the beneficial effects of using the same method will not be repeated here. For technical details not disclosed in the computer program product or computer program embodiments involved in this application, please refer to the description of the method embodiments of this application.

[0187] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An intelligent interaction method, characterized in that, The method includes: Display the business page, and display intelligent interaction controls on the business page for triggering intelligent interaction; In response to a trigger operation on the smart interactive control, a smart interactive page for the business page is displayed; the smart interactive page includes an instruction input area; In response to a target instruction for the business page sent through the instruction input area, at least one intelligent agent executes the intelligent interactive task indicated by the target instruction, and displays the task execution progress corresponding to each of the at least one intelligent agent in the progress display area; the at least one intelligent agent is determined based on the target instruction and the page context data of the business page; the at least one intelligent agent includes an intelligent agent for performing page processing on the business page; When the intelligent interactive task is completed, the business-related page after page processing of the business page is displayed on the progress display area.

2. The method according to claim 1, characterized in that, The method further includes: On the intelligent interaction page, at least one candidate skill is displayed for the business page; the at least one candidate skill is selected from the skill library based on the business page. In response to a trigger operation targeting a target candidate skill among the at least one candidate skills, a skill instruction template corresponding to the target candidate skill is displayed in the instruction input area; In response to a parameter entry operation for a skill instruction template, an instruction containing the target parameters entered in the parameter entry operation is displayed in the instruction input area.

3. The method according to claim 1, characterized in that, The method further includes: A thumbnail window is displayed on the intelligent interaction page to show a simplified version of the business page; the business page in the thumbnail window is used as a reference for command input. In response to a page adjustment operation performed on a business page in the thumbnail window, the business page after the page adjustment operation is performed is displayed in the thumbnail window.

4. The method according to claim 1, characterized in that, Page processing of the business page is accomplished through the built-in operation interface in the business system corresponding to the business page; The page processing includes one or more of the following: adding page objects to the business page, modifying or querying page objects on the business page, and deleting page objects on the business page; The business-related pages include one or more of the following: business pages obtained after adjusting the business pages, and pages retrieved after processing the business pages.

5. The method according to claim 1, characterized in that, The page context data includes any of the following: page links of the business page, page data of the business page, relevant search content of the business page, entity filtering information displayed on the business page, at least one candidate skill, and historical operation data of the target user.

6. The method according to claim 1, characterized in that, In response to a trigger operation on the smart interactive control, the method further includes: In response to a trigger operation on the smart interactive control, a smart interaction request for the business page is sent to the server, so that when the server receives the smart interaction request, it obtains multiple signal data through multiple information sources; each information source is used to obtain the corresponding signal data. Page context data for the business page is constructed based on the multiple signal data.

7. The method according to claim 1, characterized in that, The step of responding to a target instruction for the service page sent through the instruction input area, and executing the intelligent interaction task indicated by the target instruction through at least one intelligent agent, includes: In response to a target instruction for the business page sent through the instruction input area, an instruction sending request carrying the target instruction is sent to the server. Upon receiving the instruction sending request, the server determines at least one intelligent agent based on the intelligent collaborative processing strategy and the target instruction, creates an intelligent execution instance corresponding to each intelligent agent, and assigns a corresponding sub-task to each intelligent execution instance when it is determined to be at least one sub-task in the intelligent interaction task. The server then executes the corresponding sub-task through the at least one intelligent execution instance to realize the execution of the intelligent interaction task indicated by the target instruction through at least one intelligent agent.

8. The method according to claim 1, characterized in that, The intelligent interaction task includes at least one subtask, and each subtask is assigned to be executed by the corresponding intelligent agent among the at least one intelligent agents. Displaying the task execution progress of each of the at least one agent in the progress display area includes: The task execution progress of each agent in executing the corresponding subtask is received, and the task execution progress of each of the at least one agent is displayed in the progress display area.

9. An intelligent interaction method, characterized in that, The method includes: The system receives a smart interaction request triggered by a terminal device on a business page, obtains the page context data of the business page through multiple information sources, and returns page data for the smart interaction page to the terminal device so as to display the smart interaction page on the terminal device; the smart interaction page includes a command input area for command input. Upon receiving a target instruction for the business page sent through the instruction input area, at least one intelligent agent is determined based on the intelligent collaborative processing strategy and the target instruction; the at least one intelligent agent includes an intelligent agent for performing page processing on the business page. The at least one intelligent agent executes the intelligent interactive task indicated by the target instruction and returns the task execution progress corresponding to the at least one intelligent agent to the terminal device, so as to display the task execution progress corresponding to the at least one intelligent agent in the progress display area; Upon completion of the intelligent interaction task, the page data of the business-related page after page processing of the business page is returned to the terminal device so as to display the business-related page in the progress display area.

10. The method according to claim 9, characterized in that, The page context data includes at least the relevant search content of the business page and the page data of the business page; The determination of at least one intelligent agent based on the intelligent cooperative processing strategy and the target instruction includes: Based on the page data of the business page and the relevant search content of the business page, the text of the instruction text corresponding to the target instruction is optimized to obtain the task instruction text, and the intelligent interactive task is determined based on the task instruction text; The master control agent searches for a set of expert agents that match the task instruction text from a preset set of expert agents. If a set of target expert agents that matches the task instruction text is found, at least one expert agent in the set of target expert agents is used as at least one agent for the intelligent interactive task indicated by the target instruction.

11. The method according to claim 10, characterized in that, The method further includes: If no target set of expert agents matching the task instruction text is found, the capability matching agent filters at least one expert agent matching the intelligent interaction task from the expert agent library based on the page data of the business page, and the filtered at least one expert agent is taken as the at least one agent.

12. The method according to claim 10, characterized in that, The execution of the intelligent interaction task indicated by the target instruction through the at least one intelligent agent includes: Based on the task instruction text, at least one subtask is determined to be included in the intelligent interaction task, and a subtask list containing the at least one subtask is generated; the subtask list is marked with the task dependency relationship and page processing mark corresponding to each subtask. The master control agent assigns each subtask to the corresponding agent in the at least one agent according to the subtask list, and the at least one agent executes the corresponding subtask.

13. An intelligent interaction method, characterized in that, The method includes: Display the target session page, which includes a session input area; In response to sending target session content through the session input area, an instruction corresponding to the first session content is executed based on a first agent, and an instruction corresponding to the second session content is executed based on a second agent; wherein, the target session content includes first identification information of the first agent and the first session content for the first agent, and second identification information of the second agent and the second session content for the second agent; The session input area triggers the display of an agent list after the target character is entered. The target agent is selected from the agent list and its target identifier information is entered in the session input area. The agent list includes at least two agents, and the target agent is any one of the at least two agents.

14. An intelligent interactive device, characterized in that, The device includes: A business page display module is used to display a business page and to display intelligent interactive controls on the business page for triggering intelligent interaction. The intelligent interactive page module is used to display an intelligent interactive page for the business page in response to a trigger operation on the intelligent interactive control; the intelligent interactive page includes an instruction input area; A task execution module is configured to respond to a target instruction sent through the instruction input area for the business page, execute the intelligent interactive task indicated by the target instruction through at least one intelligent agent, and display the task execution progress corresponding to the at least one intelligent agent in the progress display area; the at least one intelligent agent is determined based on the target instruction and the page context data of the business page; the at least one intelligent agent includes an intelligent agent for page processing of the business page; The task completion module is used to display the business-related page after page processing of the business page on the progress display area when the intelligent interactive task is completed.

15. An intelligent interactive device, characterized in that, The device includes: The intelligent interaction request module is used to receive intelligent interaction requests triggered by the terminal device based on the business page, obtain the page context data of the business page through multiple information sources, and return page data for the intelligent interaction page to the terminal device so as to display the intelligent interaction page on the terminal device; the intelligent interaction page includes a command input area for command input; The agent determination module is used to determine at least one agent based on the intelligent collaborative processing strategy and the target instruction when receiving a target instruction for the business page sent through the instruction input area; the at least one agent includes an agent for performing page processing on the business page. The task processing module is used to execute the intelligent interactive task indicated by the target instruction through the at least one intelligent agent, and return the task execution progress corresponding to the at least one intelligent agent to the terminal device, so as to display the task execution progress corresponding to the at least one intelligent agent in the progress display area; The feedback module is used to return page data of the business-related page after page processing of the business page to the terminal device when the intelligent interaction task is completed, so as to display the business-related page on the progress display area.

16. A computer device, characterized in that, Including memory and processor; The memory is connected to the processor, the memory is used to store computer programs, and the processor is used to invoke the computer programs so that the computer device performs the method according to any one of claims 1-13.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program adapted to be loaded and executed by a processor to cause a computer device having the processor to perform the method of any one of claims 1-13.

18. A computer program product, characterized in that, The computer program product includes a computer program stored in a computer-readable storage medium, the computer program being adapted to be read and executed by a processor to cause a computer device having the processor to perform the method of any one of claims 1-13.

Images