Cooperative office method and device based on intelligent agent and medium

Abstract

The invention relates to the field of artificial intelligence, and particularly discloses an agent-based collaborative office method and device and a medium, and the method comprises the steps: receiving an initial interaction content sent by a user through a polymorphic interaction interface; performing context supplement on the initial interaction content to obtain a target interaction task; inputting the target interaction task into a preset model to obtain an interaction task result; and displaying the interaction task result by using a visual component based on the result type of the interaction task result. According to different software types, interaction with the user is carried out through the polymorphic interaction interface, and the multi-scene interaction requirement can be met. Meanwhile, by performing context supplement on the initial interaction content sent by the user, the intention recognition capability of the large model on the interaction content can be improved, so that the accuracy of the output content of the model is improved, and the use experience of the user is improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, specifically to a collaborative office method, device, and medium based on intelligent agents. Background Technology

[0002] With the advent of the era of large-scale models, more and more software supports enhancing the intelligent capabilities of systems by connecting to large-scale models. In the process of intelligent transformation, the user's interactive experience is crucial. A unified intelligent interactive experience can quickly attract users' attention and better cultivate users' habits of using intelligent capabilities.

[0003] When integrating intelligent capabilities into software, the system's intelligent capabilities may fail to accurately identify the user's intent due to the often overly simplistic nature of the questions asked, resulting in incorrect responses. Furthermore, traditional systems may require extensive interface modifications to integrate intelligent capabilities, necessitating the provision of diverse intelligent interactive experiences across different functional scenarios. Summary of the Invention

[0004] To address the aforementioned issues, this application proposes a collaborative office method, device, and medium based on intelligent agents, wherein the method includes: The system receives initial interactive content from the user through a multi-form interactive interface; it supplements the context of the initial interactive content to obtain the target interactive task; it inputs the target interactive task into a preset model to obtain the interactive task result; and it displays the interactive task result using a visualization component based on the result type of the interactive task result.

[0005] In one example, the multi-form interactive interface includes at least one of a full-screen interactive interface, a floating interactive interface, an embedded interactive interface, a thumbnail interactive interface, and a silent interactive interface.

[0006] In one example, before receiving the user's interactive content through the multi-form interactive interface, the method further includes: obtaining the user's predefined custom prompt words; displaying the custom prompt words to the user; after displaying the interactive task result using a visualization component based on the result type of the interactive task result, the method further includes: determining that the evaluation value of the interactive task result is lower than a preset threshold; replacing the custom prompt words with pre-stored standard prompt words and displaying them to the user.

[0007] In one example, the step of supplementing the initial interaction content with context to obtain the target interaction task specifically includes: obtaining system-side stored information and obtaining business background parameters through an open interface; the system-side stored information includes at least one of user attributes, system configuration information, and current functional modules; and fusing the system-side stored information and the business background parameters with the initial interaction content to obtain the interaction task.

[0008] In one example, the step of fusing the system-side stored information and the business background parameters with the initial interaction content to obtain the interaction task specifically includes: fusing the system-side stored information and the business background parameters with the initial interaction content to obtain a first intermediate result; filtering the first intermediate result to remove confidential questions, unauthorized questions, and illegal questions contained therein to obtain a second intermediate result; correcting typos and ambiguous expressions in the second intermediate result to obtain a third intermediate result; and generating a standard question expression corresponding to the third intermediate result through semantic mapping to obtain the interaction task.

[0009] In one example, after displaying the interactive task result using a visualization component, the method further includes: deleting non-text content from the interactive task result to obtain a text result; generating follow-up prompts based on the text result and historical context; and displaying the follow-up prompts to the user.

[0010] In one example, after displaying the interactive task result using a visualization component based on the result type of the interactive task result, the method further includes: determining the interactive task result and the corresponding initial interactive content; saving the interactive task result and the corresponding initial interactive content to a distributed database according to a preset classification rule; the preset classification rule includes at least one of classification by time period and classification by functional module.

[0011] In one example, the class structure in the database includes at least one of an assistant class structure, an assistant dialogue record structure, and an assistant dialogue record message structure; the assistant class structure includes at least one of an assistant primary key, assistant ID, assistant question distribution rules, whether the assistant is enabled, frequently asked questions, and an assistant onboarding page; the assistant dialogue record structure includes at least one of an assistant primary key, client identifier, dialogue name, dialogue start time, dialogue end time, the menu where the question is located, and user; the assistant dialogue record message structure includes at least one of an attribute name, dialogue record primary key, whether the question is clarified, whether it is a system message, message content, message context information, message generation time, and the associated previous message.

[0012] This application also provides a collaborative office device based on intelligent agents, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform: receiving initial interaction content sent by a user through a multi-form interactive interface; supplementing the initial interaction content with context to obtain a target interaction task; inputting the target interaction task into a preset model to obtain an interaction task result; and displaying the interaction task result using a visualization component based on the result type of the interaction task result.

[0013] This application also provides a non-volatile computer storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured to: receive initial interactive content sent by a user through a multi-form interactive interface; supplement the initial interactive content with context to obtain a target interactive task; input the target interactive task into a preset model to obtain an interactive task result; and display the interactive task result using a visualization component based on the result type of the interactive task result.

[0014] The method proposed in this application offers the following advantages: It allows for interaction with users through multi-form interactive interfaces, catering to diverse interaction needs across various scenarios, depending on the software type. Furthermore, by supplementing the initial user-sent interaction content with contextual information, it enhances the large model's ability to recognize the intent behind the interaction, thereby improving the accuracy of the model's output and ultimately enhancing the user experience. Attached Figure Description

[0015] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating a collaborative office method based on intelligent agents, as described in an embodiment of this application. Figure 2 This is a schematic diagram of the interface of a newly created intelligent agent in an embodiment of this application; Figure 3 This is a schematic diagram of the interface for defining components when creating a new intelligent agent according to an embodiment of this application; Figure 4 This is a schematic diagram of the basic information maintenance interface for a newly created intelligent agent in an embodiment of this application; Figure 5 This is a schematic diagram of a newly created intelligent agent's binding menu information maintenance interface in an embodiment of this application; Figure 6This is a schematic diagram of the access interface of a newly created intelligent agent to an intelligent application in an embodiment of this application; Figure 7 This is a schematic diagram of a guide card configuration interface for creating a new intelligent agent in an embodiment of this application; Figure 8 This is a schematic diagram of a full-screen interactive interface of an intelligent agent in an embodiment of this application; Figure 9 This is a schematic diagram of a floating interactive interface of an intelligent agent in an embodiment of this application; Figure 10 This is a schematic diagram of the structure of a collaborative office device based on an intelligent agent, as described in an embodiment of this application. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0017] The technical solutions provided by the various embodiments of this application are described in detail below with reference to the accompanying drawings.

[0018] Figure 1 This diagram illustrates a process flow for a collaborative work method based on intelligent agents, provided in one or more embodiments of this specification. This method can be applied to various business domains, such as internet finance, e-commerce, instant messaging, gaming, and government services. The process can be executed by intelligent agents pre-installed within the system, and certain input parameters or intermediate results can be manually adjusted to improve accuracy.

[0019] The analysis method involved in the embodiments of this application can be implemented by a terminal device or a server, and this application does not impose any special limitations on it. For ease of understanding and description, the following embodiments are all described in detail using an intelligent agent as an example. It should be noted that the server can be a single device or a system composed of multiple devices, that is, a distributed server, and this application does not impose any specific limitations on it.

[0020] When building this intelligent agent, after the user completes the standard intelligent agent configuration (interfacing with the large model, improving prompt words, and supplementing self-developed interface tools), the intelligent agent can be described a second time (including keyword and description information maintenance, all of which are custom configurations used for hierarchical configuration loading), such as... Figure 2 As shown.

[0021] Then, when defining components, developers can first import open-source components or develop custom components, package them into a JavaScript script, and load the component onto the installation disk. Then, in the component management section, the component is named and its path is specified, making it easy for the assistant to call it at runtime, such as... Figure 3 As shown.

[0022] When defining a smart application assistant, you need to create a new assistant, enter the assistant number, name, description, and other information. The enabled status is enabled by default. Figure 4 As shown. Then select the menu bound to the assistant, mainly choosing which menus can display the assistant's floating form, such as... Figure 5 As shown. Then, it connects to intelligent applications to maintain track which intelligent applications can forward questions asked to the assistant, such as... Figure 6 As shown. Finally, you need to configure the onboarding cards, which are used to display the operation guidance components in the assistant's full-screen page, such as... Figure 7 As shown.

[0023] like Figure 1 As shown in the figure, this application provides a collaborative office method based on intelligent agents, including: S101: Receives initial interactive content from the user through a multi-form interactive interface.

[0024] First, users send initial interaction content to the intelligent agent through a multi-form interactive interface in the intelligent application. The initial interaction content here refers to the question that the user inputs that requires assistance, such as asking the intelligent agent what the schedule is for today, or asking the intelligent agent to generate a business report based on business data.

[0025] The multi-form interactive interfaces include full-screen interactive interfaces, floating interactive interfaces, embedded interactive interfaces, thumbnail interactive interfaces, and silent interactive interfaces. Full-screen interactive interfaces are primarily used for complex data visualization, and their presentation forms include... Figure 8 As shown, users can enter initial interaction content in the input box located below the full interactive interface. The embedded interactive interface is mainly used for real-time data interaction with the business menu. Users can enter initial interaction content in the input box located below the embedded interactive interface and view current tasks or business data on most of the left-hand pages. The embedded interactive page can be moved by dragging. Dynamic adaptation for mobile devices supports thumbnail voice interaction and loading of the system robot in a silent state. By setting up various interactive interfaces, it can adapt to various work scenarios, thereby meeting the interaction needs of multiple scenarios.

[0026] S102: Supplement the initial interaction content with context to obtain the target interaction task.

[0027] Because user questions are often too simple, such as "Help me generate a report" or "How do I follow this process?", the intelligent agent may not be able to accurately identify the user's intent, leading to an incorrect response. To avoid this, after receiving the initial user interaction, the agent can supplement the context to obtain the target interaction task.

[0028] During context supplementation, it is necessary to obtain system-side stored information and acquire business background parameters through open interfaces. The system-side stored information includes user attributes, system configuration information, and information about the current functional modules. This system-side stored information and business background parameters are then integrated with the initial interaction content to obtain the interaction task.

[0029] User attributes can include data such as gender, job title, and organization (e.g., user profile data), while system configurations can include data such as time zone (for time-related responses) and language preference (for switching between Chinese and English). The current functional module refers to the specific function the user is using (e.g., "contract generation," "data analysis," etc.). By retrieving information stored on the system side, basic contextual issues can be resolved. For example, when a user asks "What are my plans for today?", the system can automatically distinguish between "personal schedule" and "team meeting" based on the time zone and job title.

[0030] By connecting to business systems via APIs to obtain business background parameters, we can supplement in-depth domain information. Typical business background parameters include the business year (for financial issues, it's necessary to distinguish between fiscal years 2023 and 2024), organization type (to differentiate between hospitals and internet companies regarding the same terminology), and process stage (such as the "inquiry" or "contract signing" stage in a sales process). By obtaining these business background parameters, when a user asks "What is the current KPI progress?", we can automatically link it to their department's annual target data, thereby accurately identifying the user's intent and reducing the need for users to repeat themselves. This eliminates the need for users to repeatedly state "I am a designer, please answer using design terminology" each time they ask a question.

[0031] In one embodiment, to reduce input noise in the model, after context supplementation, a built-in sensitive word filtering layer in the agent can be used to intercept confidential, unauthorized, and illegal questions in real time; a knowledge retrieval service layer can automatically correct typos and ambiguous expressions; and finally, a standard question expression can be generated through semantic mapping to reduce input noise in the large model. Specifically, the agent first integrates system-stored information and business background parameters with the initial interaction content to obtain a first intermediate result. Then, the first intermediate result is filtered to remove confidential, unauthorized, and illegal questions, resulting in a second intermediate result. Next, typos and ambiguous expressions in the second intermediate result are corrected to obtain a third intermediate result. Finally, a standard question expression corresponding to the third intermediate result is generated through semantic mapping to obtain the interaction task.

[0032] S103: Input the target interaction task into the preset model to obtain the interaction task result.

[0033] After obtaining the target interaction task, it can be input into a pre-bound model of the agent. The output result corresponding to the target interaction task is determined by the preset model and serves as the interaction task result. There can be one or more preset models. When multiple preset models exist simultaneously (e.g., both ChatGPT3.0 and ChatGPT4.0), the appropriate model can be selected based on the agent's current computing power and specific task requirements. For example, when the agent's server has low remaining computing power and the user has high requirements for timely feedback, a smaller and faster preset model can be used for calculation. Furthermore, since different models have different capabilities in processing different types of data, multiple preset models can be used simultaneously to output interaction task results, allowing the user to refer to multiple interaction task results.

[0034] S104: Based on the result type of the interactive task result, use a visualization component to display the interactive task result.

[0035] After obtaining the results of the interactive task, the results can be displayed using corresponding visualization components based on the result type and content. Complex data rendering such as tables and bar charts can be achieved through the Echarts chart component built into the agent. Standardized component development interfaces can also be provided to support secondary development of custom components and rapid integration.

[0036] In one embodiment, to facilitate user questioning, prompts can be displayed to the user during interactions with various interactive interfaces. This helps the user better organize their thoughts and improves the model's recognition capabilities. Users can define their own prompts; for example, they can generate prompts for frequently asked questions, such as "Help me generate a daily report based on today's work." These user-generated prompts can be stored in the intelligent application. When a user asks a question, if the application recognizes a pre-defined prompt, it can display the user-defined prompt. If the prompt matches the user's question, the user can directly select the prompt to ask their question. It should be noted that prompts can represent not only complete tasks but also types of people; for example, a prompt could be "Please help me translate the following content." When the model recognizes such a prompt, it can better understand the user's intent, thereby improving the accuracy of the model's output.

[0037] If a user is dissatisfied with the model's output corresponding to a custom prompt word, for example, by clicking a negative review to give feedback to the agent, then the evaluation value of the interactive task result is considered to be lower than a preset threshold. In this case, the custom prompt word displayed in the prompt word area can be replaced with a pre-stored standard prompt word and displayed to the user.

[0038] Besides users directly clicking the negative review button and giving a low satisfaction rating, the agent can also independently judge the interactive task results output by the model. For example, if the custom prompt contains "Please generate a table," but the displayed interactive task result is plain text, then the custom prompt is considered to have a poor match with the model. Furthermore, the evaluation value of the interactive task result can be determined by the number of user modification iterations. When the number of user modification iterations exceeds a preset threshold, the evaluation value of the interactive task result is considered to be below the preset threshold.

[0039] Typical scenarios for error handling include: format failure scenarios (e.g., a custom prompt word comparing the configurations of the iPhone 15 and Samsung S24 using a Markdown table, but the model returns a plain text list without generating a table), semantic deviation scenarios (e.g., a custom prompt word providing advice in a warm and soothing tone as a depression psychologist, but the model fails to adapt to the emotional requirements and only outputs mechanical clinical terminology), and logical error scenarios (e.g., a custom prompt word listing five potential Nobel Prize winners for 2025, requiring an analysis of their academic achievements, but the model fabricates non-existent scholars or awards, such as including "2025 Nobel Prize in Mathematics").

[0040] The aforementioned pre-stored standard prompts can be stored in the computer device's storage. When a user asks a question, the intelligent application can select the pre-stored standard prompts from the storage device. Alternatively, the intelligent application can also obtain the pre-stored standard prompts from other external devices. For example, the pre-stored standard prompts can be stored in the cloud, and the intelligent application can retrieve them from the cloud when a user asks a question. This embodiment does not limit the method of obtaining the pre-stored standard prompts.

[0041] In one embodiment, in a conversational chain-like follow-up questioning scenario, to improve the quality of user follow-up questions, supplementary information can be generated based on the historical question-and-answer context. The intelligent compression module automatically removes non-text content such as charts and attachments, maintaining semantic integrity while reducing the consumption of large model tokens. Specifically, the agent deletes non-text content from the interaction task results to obtain text results, and generates follow-up question prompts based on the text results and historical context, finally displaying the follow-up question prompts to the user.

[0042] In one embodiment, to provide data support to the agent, it is necessary to save the question content and the interaction task results output by the model. This requires determining the interaction task results and the corresponding initial interaction content, and saving these results and initial interaction content to a distributed database according to preset classification rules. The preset classification rules include at least one of classification by time period and classification by functional module. By saving the interaction task results and the corresponding initial interaction content, users can review content and model responses from a short period (e.g., three days ago), avoiding the need for users to re-ask questions if they forget the model's output.

[0043] The aforementioned databases, especially the utility class structure in relational databases, primarily include assistant class structures, assistant dialogue record structures, and assistant dialogue record message structures. The assistant class structure includes the assistant primary key, assistant ID, assistant question distribution rules, whether the assistant is enabled, frequently asked questions, and the assistant onboarding page. The assistant dialogue record structure includes the assistant primary key, client identifier, dialogue name, dialogue start time, dialogue end time, the menu where the question is located, and the user. The assistant dialogue record message structure includes attribute names, dialogue record primary key, whether the question is clarified, whether it is a system message, message content, message context information, message generation time, and the associated previous message.

[0044] Specifically, the helper class structure table is as follows:

[0045] The assistant's dialogue log structure table is as follows:

[0046] The assistant's conversation log message structure table is as follows:

[0047] The agent-based collaborative office method provided in this application can solve the technical problems of high reconstruction cost, single interaction form, poor compatibility of prompt words and low conversation efficiency when traditional business systems access large models.

[0048] like Figure 10 As shown, this application embodiment also provides a collaborative office device based on intelligent agents, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to: The system receives initial interactive content from the user through a multi-form interactive interface; it supplements the context of the initial interactive content to obtain the target interactive task; it inputs the target interactive task into a preset model to obtain the interactive task result; and it displays the interactive task result using a visualization component based on the result type of the interactive task result.

[0049] This application embodiment also provides a non-volatile computer storage medium storing computer-executable instructions, wherein the computer-executable instructions are configured as follows: The system receives initial interactive content from the user through a multi-form interactive interface; it supplements the context of the initial interactive content to obtain the target interactive task; it inputs the target interactive task into a preset model to obtain the interactive task result; and it displays the interactive task result using a visualization component based on the result type of the interactive task result.

[0050] The various embodiments in this application are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the device and medium embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the description of the method embodiments.

[0051] The devices and media provided in this application are one-to-one with the methods. Therefore, the devices and media also have similar beneficial technical effects as their corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the devices and media will not be repeated here.

[0052] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0053] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0054] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0055] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0056] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0057] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0058] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0059] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0060] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A collaborative office method based on intelligent agents, characterized in that, include: Receive initial interactive content from users through a multi-form interactive interface; The initial interaction content is supplemented with context to obtain the target interaction task; The target interaction task is input into a preset model to obtain the interaction task result; Based on the result type of the interactive task, the results of the interactive task are displayed using visualization components.

2. The method according to claim 1, characterized in that, The multi-form interactive interface includes at least one of the following: full-screen interactive interface, floating interactive interface, embedded interactive interface, thumbnail interactive interface, and silent interactive interface.

3. The method according to claim 2, characterized in that, Before receiving user-sent interactive content through the multi-form interactive interface, the method further includes: Retrieve the user-defined custom prompt words; Display the custom prompts to the user; After displaying the interactive task results using a visualization component, the method further includes: The evaluation value of the interactive task result is determined to be lower than a preset threshold; Replace the custom prompt with a pre-stored standard prompt and display it to the user.

4. The method according to claim 1, characterized in that, The step of supplementing the initial interaction content with context to obtain the target interaction task specifically includes: Obtain system-side storage information and retrieve business background parameters through open interfaces; The system-side stored information includes at least one of user attributes, system configuration information, and current functional modules. The system-side stored information and the business background parameters are integrated with the initial interaction content to obtain the interaction task.

5. The method according to claim 4, characterized in that, The step of fusing the system-side stored information and the business background parameters with the initial interaction content to obtain the interaction task specifically includes: The system-side stored information and the business background parameters are fused with the initial interaction content to obtain a first intermediate result; The first intermediate result is filtered to remove classified questions, unauthorized questions, and illegal questions contained in the first intermediate result, to obtain the second intermediate result; The typos and ambiguous expressions in the second intermediate result are corrected to obtain the third intermediate result; The standard question expression corresponding to the third intermediate result is generated through semantic mapping to obtain the interactive task.

6. The method according to claim 1, characterized in that, After displaying the results of the interactive task using visualization components, the method further includes: Remove non-text content from the results of the interactive task to obtain text results; Based on the text results and historical context, prompts for follow-up questions are generated; The prompts for further questions are displayed to the user.

7. The method according to claim 1, characterized in that, After displaying the interactive task results using a visualization component, the method further includes: Determine the result of the interactive task and the corresponding initial interactive content; The results of the interactive tasks and the corresponding initial interactive content are saved to a distributed database according to preset classification rules. The preset classification rules include at least one of classification by time period and classification by functional module.

8. The method according to claim 7, characterized in that, The class structure in the database includes at least one of the following: assistant class structure, assistant dialogue record structure, and assistant dialogue record message structure. The assistant class structure includes at least one of the following: assistant primary key, assistant number, assistant question distribution rules, whether the assistant is enabled, frequently asked questions, and assistant onboarding page; The assistant dialogue record structure includes at least one of the following: assistant primary key, client identifier, dialogue name, dialogue start time, dialogue end time, menu where the question is located, and user; The assistant dialogue record message structure includes at least one of the following: attribute name, dialogue record primary key, whether it is a question clarification, whether it is a system message, message content, message context information, message generation time, and the associated previous message.

9. A collaborative office device based on intelligent agents, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform: Receive initial interactive content from users through a multi-form interactive interface; The initial interaction content is supplemented with context to obtain the target interaction task; The target interaction task is input into a preset model to obtain the interaction task result; Based on the result type of the interactive task, the results of the interactive task are displayed using visualization components.

10. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer-executable instructions are set as follows: Receive initial interactive content from users through a multi-form interactive interface; The initial interaction content is supplemented with context to obtain the target interaction task; The target interaction task is input into a preset model to obtain the interaction task result; Based on the result type of the interactive task, the results of the interactive task are displayed using visualization components.

Images