Digital assistant use method and computing device

Abstract

This application relates to the field of computing, and in particular provides a method for using a digital assistant and a computing device. The method for using a digital assistant includes: a first digital assistant receiving a first instruction triggered by a first user, the first instruction including instruction content provided by the first user; planning at least one work task to be performed based on the instruction content in the first instruction; acquiring personalized skills in the first user's personal space; and using the personalized skills to perform the at least one work task.

Description

Technical Field

[0001] This application relates to the field of computing, and in particular to a method for using a digital assistant and a computing device. Background Technology

[0002] In the field of artificial intelligence, an agent is a software entity with autonomous perception, decision-making, and execution capabilities. Its core value lies in replacing humans to complete automated and intelligent tasks in specific scenarios. As the application of agents becomes more widespread, especially in enterprise management scenarios, the demand for digital assistants is becoming increasingly diversified.

[0003] Digital assistants, in an enterprise environment, are virtual assistants based on large models, configured for each employee. They possess the ability to understand, plan, execute, remember, and interact with tasks, and can replace employees in completing daily office tasks. Currently, digital assistants only have simple task execution capabilities, such as data querying and report generation, and are highly dependent on human intervention, lacking strong collaborative capabilities. Summary of the Invention

[0004] This application provides a method for using a digital assistant and a computing device. The digital assistant can autonomously plan the work tasks to be performed and help users efficiently complete various personalized work tasks by reusing personalized skills in their personal space.

[0005] According to a first aspect of the embodiments of this application, a method for using a digital assistant is provided, including: The first digital assistant receives a first instruction triggered by the first user, the first instruction including the instruction content provided by the first user; Based on the instructions in the first instruction, plan at least one work task that needs to be performed; To acquire personalized skills from the first user's personal space; Utilize your individual skills to perform at least one work task.

[0006] In this embodiment, the first digital assistant of the first user can receive a first instruction triggered by the first user, and then autonomously plan one or more work tasks to be executed according to the instruction content in the first instruction. It can also retrieve personalized skills from the user's personal space to execute each task, realizing personalized and automated closed-loop processing from instruction to task execution. This not only accurately responds to the user's customized needs and makes the task execution fit the user's exclusive skills, avoiding the task and skill adaptation deviation caused by generalized processing, but also eliminates the need for manual task splitting and selection tools, greatly improving the efficiency and accuracy of task processing. At the same time, by reusing personalized skills from the personal space, the digital assistant's adaptability to different user scenarios is enhanced, helping users to efficiently complete various personalized work tasks.

[0007] In conjunction with the first aspect, in some implementations of the first aspect, personalized skills are utilized to perform at least one work task, including: Based on the task requirements of each work task, select the target skills needed to perform each work task from the personalized skills; Utilize the target skills required for each task to perform that task.

[0008] In this embodiment, the corresponding target skill can be accurately matched from the user's personalized skills according to the specific needs of each task, and the task can be executed. This achieves precise adaptation and efficient execution of skills and tasks, allowing each task to be efficiently advanced through the adapted personalized skills. This ensures the accuracy and adaptability of task execution, and eliminates the need for manual skill selection and task allocation. It significantly improves the efficiency of multi-task parallel processing. At the same time, the application of personalized skills can make task execution more in line with the user's personalized task execution scenario, further enhancing the work capabilities and effectiveness of digital assistants for different users.

[0009] In conjunction with the first aspect, in some implementations of the first aspect, personalized skills include at least one of the following: The digital assistant of the second user, where the second user refers to a user associated with the first user; The first user's personal skills refer to the skills that the first user possesses to perform specific technical services; The first user's personal tools refer to plugins / software set up by the first user that can be directly invoked; The first user's personal knowledge base refers to the knowledge base formed by the first user's knowledge files.

[0010] In this embodiment, by setting the personalized skills of the first user to include a digital assistant of the second user associated with the first user, the first user's personal skills, directly accessible personal tools, and a dedicated personal knowledge base, a comprehensive and customized support is provided for the digital assistant's work capabilities. This makes the skills cover a wider range of scenarios, not only avoiding the limitations of a single skill and greatly improving the adaptability and implementation capability of the digital assistant in dealing with complex tasks, but also making task execution more in line with the user's dedicated resource configuration and usage habits, effectively improving the personalized service value of the digital assistant and the reuse efficiency of work skills.

[0011] In conjunction with the first aspect, some implementations of the first aspect also include: If the target skill for any work task includes at least one of the following: a digital assistant, personal skills, personal tools, and personal knowledge base required to perform the work task, then the work task is performed using at least one of the following: a digital assistant, personal skills, personal tools, and personal knowledge base.

[0012] In this embodiment of the application, by setting target skills for work tasks, namely, at least one of digital assistants, personal skills, personal tools, and personal knowledge bases, flexible adaptation of executable skills corresponding to each work task is achieved, realizing efficient connection between task execution and diverse personalized skills, ensuring the smooth execution of various work tasks, while simplifying the judgment logic of skill matching and improving the flexibility and efficiency of task execution.

[0013] In conjunction with the first aspect, some implementations of the first aspect also include: In response to a selection action performed by the first user on at least one of the digital assistants, skill libraries, tool libraries, and knowledge bases contained in their team space, the first user's personalized skills are obtained.

[0014] In this embodiment, the first user selects personalized skills from the digital assistant, skill library, tool library, and knowledge base included in the team space, achieving team-level sharing of work skills and realizing a capability-sharing model of "one person develops, everyone reuses." Customized work skills ensure efficient linkage between personalized skills and team resources, allowing the digital assistant's capabilities to not only meet the user's specific needs but also reuse high-quality team resources, further enhancing the flexibility of skill configuration and the adaptability of task execution.

[0015] In conjunction with the first aspect, in certain implementations of the first aspect, based on the instructions in the first instruction, at least one task to be executed is planned, including: Identify the user intent of the first user from the content of the first instruction to obtain the target task corresponding to the user intent; Extract task information related to the target task from the first user's personal memory; Based on the task information related to the target task, task planning is performed on the target task to obtain at least one corresponding work task.

[0016] In this embodiment, the user's intent is accurately identified from the instruction content contained in the user's command to determine the target task. Combined with the first user's personal memory, task information related to the target task is extracted. Based on this task information, task planning is carried out, achieving personalized and precise breakdown of work tasks. The task planning process has a higher degree of alignment with the user's historical behavior and memory, avoiding deviations caused by generalized planning, providing precise support for task breakdown, ensuring the integrity and adaptability of the task flow, improving the efficiency and accuracy of task planning, and effectively enhancing the personalized service capabilities and task execution effects of the digital assistant.

[0017] In conjunction with the first aspect, in some implementations of the first aspect, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task, including: Generate enhanced prompts based on task information related to the target task; Based on the enhanced prompts, construct a task framework corresponding to the target task. The task framework includes: a strategic layer, at least one tactical layer in the strategic layer, and at least one execution layer in each tactical layer. Based on the framework structure corresponding to the strategic layer, generate strategic information for the strategic layer; Based on the preset task structure, the task is decomposed at each tactical layer to obtain the atomic tasks and / or composite tasks corresponding to each tactical layer. Based on the preset execution structure, the atomic tasks and / or composite tasks of the corresponding tactical layer are decomposed into steps at the execution layer to obtain the steps to be executed for each atomic task and the steps to be executed for the composite task. Each atomic task and each composite task whose steps need to be performed is identified as at least one working task corresponding to the target task.

[0018] In this embodiment, enhanced prompts are generated based on target task-related information. A hierarchical task framework, comprising strategic, tactical, and execution layers, is then constructed based on these enhanced prompts. This enables structured, hierarchical decomposition and standardized management of target tasks, resulting in clear task hierarchy and logical order. Specific execution tasks are further generated from atomic and composite tasks at the tactical layer. This refines high-level objectives into directly executable units, significantly improving the completeness, logic, and executability of task planning. Simultaneously, it allows for the top-down, layered implementation of complex objectives, effectively reducing task execution complexity and enhancing the planning efficiency and execution reliability of the digital assistant for complex tasks.

[0019] In conjunction with the first aspect, in some implementations of the first aspect, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task, further including: Determine the current planning status corresponding to the first instruction, which is either a user-initiated request to adjust the plan, a re-planning process after task execution failure, or the first planning attempt. When the planning status is initial planning, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task. When the user actively requests to adjust the planning status, the task adjustment information provided by the first user is collected, the task information related to the target task is updated according to the task adjustment information, and the target task is re-planned according to the updated task information to obtain at least one working task corresponding to the target task. When the planning status is "task execution failed and replanning", collect relevant information about the previous task execution failure; based on the relevant information about the previous task execution failure, update the task information related to the target task, and replan the target task based on the updated task information to obtain at least one working task corresponding to the target task.

[0020] In this embodiment, the planning state corresponding to the first instruction is determined, such as user-initiated adjustment, replanning after execution failure, or initial planning. The corresponding plan is then executed according to different planning states, achieving precise control over the entire lifecycle of task planning. This enables accurate identification and differentiated processing of planning states across multiple scenarios, significantly improving the flexibility, reliability, and adaptability of task planning, while simultaneously meeting the needs for automated execution and collaborative manual intervention.

[0021] In conjunction with the first aspect, some implementations of the first aspect, in the process of task planning for the target task based on task information related to the target task, also include: Based on the task planning progress, a progress bar is generated to indicate the execution status of the task plan. And / or, monitor abnormal information in real time during the task planning and execution process. When a preset error message is detected, trigger the planning status to be set to task execution failure and replanning to start the task planning repair process. And / or, through a preset manual intervention interface, collect user confirmation or modification instructions on error repair results or task planning.

[0022] In this embodiment, a visual progress bar is generated based on the task planning progress, which can intuitively and in real time prompt the task execution status, improving the user's perception and control of the task progress; abnormal information during the execution process is monitored in real time, and when preset error information is detected, the task execution failure replanning state is automatically triggered and the repair process is started, which can effectively improve the fault tolerance and self-healing ability of task execution; through a preset manual intervention interface, the user's confirmation and modification instructions on error repair results or task planning are collected, realizing the organic combination of automated planning and manual decision-making, which not only ensures the efficiency of task execution, but also enhances the flexibility and controllability of the system, thereby improving the overall reliability, adaptability and user experience of task planning.

[0023] In conjunction with the first aspect, in some implementations of the first aspect, after obtaining at least one working task corresponding to the target task, the method further includes: Output the target task and at least one corresponding work task for the first user to view and confirm; In response to the confirmation operation triggered by the first user on the target task and at least one work task, at least one work task is obtained; Alternatively, in response to a modification operation triggered by a first user on a target task or at least one work task, obtain the target task or at least one work task modified by the first user.

[0024] In this embodiment, a target task and at least one corresponding work task are output for user review and confirmation. After reviewing the target task and each work task, the user can confirm or modify them. This means the system can respond to the first user's confirmation or modification operation, thus clarifying the final target task and at least one work task. Through user interaction, the first user can clearly understand the task execution logic, thereby adjusting the task content according to actual needs. This ensures a precise match between the work task and the user's actual needs. Simultaneously, the user interaction process reduces the probability of task execution failure or errors, improving the accuracy and effectiveness of task implementation.

[0025] In conjunction with the first aspect, in some implementations of the first aspect, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task, further including: If the first user has the permission to execute the target task, the target task is planned based on the task information related to the target task, and at least one work task corresponding to the target task is obtained.

[0026] In this embodiment, user execution permission verification is added to the task planning stage. Only when the first user has the permission to execute the target task will the target task be broken down into specific work tasks based on relevant task information. This achieves precise linkage between permission control and task planning. It can avoid task execution obstruction or illegal operations due to insufficient permissions, ensuring the security and compliance of task execution, ensuring the accurate implementation of task planning within the scope of permissions, and making the work task breakdown both in line with user needs and historical information. This reduces subsequent task execution failures or errors caused by permission issues, and improves the smoothness and overall efficiency of task execution.

[0027] In conjunction with the first aspect, some implementations of the first aspect also include: If missing information is detected during the execution of at least one task, a supplementary prompt message is output, prompting the first user to supplement the missing information. Collect missing information supplemented by the first user, and use the missing information and personalized skills to perform at least one job task.

[0028] In this embodiment, during the execution of at least one task, missing information is detected in real time. Supplementary prompts are output to guide the first user to complete the missing information. The task execution continues based on the supplemented information and personalized skills, achieving dynamic completion and execution of tasks, ensuring the completeness and accuracy of task completion. This allows for resolving missing information issues during task execution, avoiding restarting the task execution process, and improving the fault tolerance and success rate of task processing.

[0029] According to a second aspect of the embodiments of this application, a digital assistant usage device is provided, including: a transceiver unit and a processing unit.

[0030] The transceiver unit is used to receive a first instruction triggered by a first user, the first instruction including the instruction content provided by the first user.

[0031] The processing unit is used to plan at least one work task to be executed according to the instructions in the first instruction; obtain personalized skills in the personal space of the first user; and use the personalized skills to execute at least one work task.

[0032] According to a third aspect of the embodiments of this application, a computing device is provided, including: a memory and a processor; the memory is used to store a computer program; the processor is used to execute the computer program to implement any of the above-described digital assistant usage methods.

[0033] According to a fourth aspect of the embodiments of this application, a communication device is provided, including a transceiver and a processor, wherein the transceiver is used to receive or send data, and the processor is used to execute any of the digital assistant usage methods of the embodiments of this application.

[0034] According to a fifth aspect of the embodiments of this application, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, it implements a method of using a digital assistant as described above.

[0035] According to a sixth aspect of the present application, a computer product is provided, comprising: a computer program that, when executed by a processor, implements the steps of any digital assistant usage method.

[0036] It should be understood that both the foregoing general description and the following detailed description are exemplary and intended to provide further illustration of the claimed technology. Attached Figure Description

[0037] The above and other objects, features, and advantages of the embodiments of this application will become more apparent from the more detailed description of the embodiments in conjunction with the accompanying drawings. The accompanying drawings are used to provide a further understanding of the embodiments of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the embodiments of this application and do not constitute a limitation thereof. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0038] Figure 1 The illustration shows an example of an application scenario of a digital assistant according to an embodiment of this application; Figure 2 The figure shows a flowchart of a digital assistant usage method according to an embodiment of this application; Figure 3 The figure shows another flowchart of a digital assistant usage method according to an embodiment of this application; Figure 4 The figure shows another example flowchart of a digital assistant usage method according to an embodiment of this application; Figure 5 The figure shows a schematic diagram of a digital assistant user device according to an embodiment of this application; Figure 6 The figure shows a hardware block diagram of a computing device according to an embodiment of this application. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of this application more apparent, exemplary embodiments according to the embodiments of this application will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the embodiments of this application, and not all embodiments of the embodiments of this application. It should be understood that the embodiments of this application are not limited to the exemplary embodiments described herein.

[0040] The technical solutions of this application embodiment can be applied to the field of digital assistant applications. For digital assistants, they can autonomously plan the work tasks to be performed and help users efficiently complete various personalized work tasks by reusing personalized skills in their personal space.

[0041] The technical solutions of the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0042] like Figure 1 The diagram shown is an example of an application scenario for a digital assistant provided in an embodiment of this application. This application scenario may include: a first digital assistant 10 and the first digital assistant's personalized skills 20 and personal memory 30.

[0043] The first digital assistant 10 can receive a first instruction triggered by a first user, the first instruction including instruction content provided by the first user; plan at least one work task to be performed according to the instruction content in the first instruction; obtain personalized skills 20 in the first user's personal space; and use the personalized skills 20 to perform at least one work task.

[0044] In one possible design, personalized skill 20 may include at least one of the following: The digital assistant 20a of the second user refers to the user related to the first user. The digital assistant can refer to an intelligent agent or AI (Artificial Intelligence) model with corresponding processing capabilities. The first user's personal skills 20b, personal skills refer to the skills possessed by the first user to perform specific technical services; The first user's personal tools 20c, personal tools refer to plugins / software set up by the first user that can be directly called; The first user's personal knowledge base 20d refers to the knowledge base formed by the first user's knowledge files.

[0045] In another possible design, personal memory 30 may include at least one of the following: Expert experience information 30a, common sense memory information 30b, historical experience information 30c.

[0046] The first user's personalized skills can be derived from their team space 40, which may contain digital assistants, skills, tools, and knowledge bases. The first user can select at least one of the digital assistants, skills, tools, and knowledge bases contained in the team space 40 to build their personalized skills.

[0047] like Figure 2 The diagram shown is a flowchart of a digital assistant usage method provided in an embodiment of this application. The digital assistant usage method includes the following steps: S201, The first digital assistant receives a first instruction triggered by the first user, the first instruction including the instruction content provided by the first user.

[0048] In this context, a digital assistant can refer to an intelligent agent or AI model with corresponding processing capabilities. For example, a digital assistant can assist users in performing relevant work tasks, that is, it can execute any of the digital assistant usage methods provided in the embodiments of this application.

[0049] Optionally, the first instruction can refer to an instruction triggered by the first user. The first instruction can be any type of instruction, such as text, video, voice, or image. The instruction content can refer to the text or features corresponding to the first instruction. If the first instruction is text, the text in the first instruction can be directly identified as the instruction content. If the first instruction is voice, text can be extracted from the voice to obtain the instruction content. If the first instruction is video, text and / or image features can be extracted from the video to obtain the instruction content. If the first instruction is image, image features can be extracted from the image to obtain the instruction content. There are no strict limitations on the methods for extracting text or image features.

[0050] The user terminal can collect a first instruction triggered by a first user and send the first instruction to a first digital assistant running on the server. The first digital assistant can run on the server, and the first digital assistant receiving the first instruction triggered by the first user can mean that the first digital assistant running on the server receives the first instruction triggered by the first user.

[0051] Specifically, the user terminal can collect the text / voice / video / image input by the first user to obtain the first instruction.

[0052] Optionally, before the first digital assistant receives the first instruction triggered by the first user, the method further includes: receiving a login request from the first user, the login request including the first user's login information; performing security verification on the first user's login information to obtain a security verification result; and determining the first digital assistant associated with the first user if the security verification result is successful. Afterwards, the first digital assistant can be controlled to execute the digital assistant usage method of this application embodiment.

[0053] It is understandable that the first user's first digital assistant can be pre-associated. After registering, the first user can initiate a request to create a digital assistant. The server can respond to this request, generate a first digital assistant for the first user, and establish an association between the first user's user information and the employee ID of the first digital assistant. Of course, determining the first digital assistant associated with the first user can include: querying the employee ID associated with the first user's user information and identifying the digital assistant corresponding to that employee ID as the first digital assistant.

[0054] In one possible design, the server can be a cloud server, such as a company's private cloud or a server for an internal work system. An internal work system could be, for example, an internal OA (Office Automation) system, an ERP (Enterprise Resource Planning) system, or a code platform.

[0055] S202. Based on the instructions in the first instruction, plan at least one work task that needs to be executed.

[0056] Optionally, a work task can refer to a specific, measurable task that the first digital assistant needs to perform, or a problem that needs to be solved. Specifically, it can include the following components: task input, task output, task objective, and task evaluation method.

[0057] The input to a task can be, for example, text, video, audio, or images. The task objective can specifically refer to the tools needed during task execution and the specific tasks to be performed. The task output refers to the results returned by the model. The evaluation method can specifically refer to judging the effectiveness of task execution, such as using metrics like accuracy, ROUGE (Recall-Oriented Understudy for Gisting Evaluation), BLEU (Bilingual Evaluation Understudy), and PPL (Perplexity).

[0058] For example, a task may include at least one of the following: a tool-type task, a natural language understanding task, a natural language generation task, a knowledge-type task, or a reasoning task.

[0059] For example, tool-type tasks can be tasks that require the use of tools, such as recording audio, recording work meetings, retrieving work minutes, and sending emails. Natural language understanding tasks can be at least one of the following: text classification, entity recognition, relation extraction, and text matching. Natural language generation tasks can be at least one of the following: text summarization (long text to short text), text translation, dialogue generation (chat, customer service), content creation, code generation / completion, and rewriting. Knowledge-type tasks can be at least one of the following: question answering, knowledge base question answering, information extraction, and instruction following. Reasoning tasks can include at least one of the following: document understanding, logical reasoning, and text enhancement.

[0060] S203, Obtain personalized skills from the first user's personal space.

[0061] Alternatively, personal space can refer to storage space used to store the personalized skills and personal memories of the corresponding user. (Regarding personalized skills and personal memories...)

[0062] Personalized skills may include at least one of the following: The digital assistant of the second user refers to the user associated with the first user. The digital assistant can refer to an intelligent agent or AI model with corresponding processing capabilities.

[0063] The first user's personal skills refer to the skills that the first user possesses to perform specific technical services.

[0064] The first user's personal tools refer to plugins / software that can be directly accessed by the first user.

[0065] The first user's personal knowledge base refers to the knowledge base formed by the first user's knowledge files.

[0066] In this embodiment, the second user is a user related to the first user. This could be a user within the first user's network of relationships, or an expert user accessible to the first user. An expert user can refer to a user with expert status or identification. Expert users can be associated with pre-trained intelligent agents or AI models possessing specialized processing capabilities. Expert users can include internal and external expert users. Internal expert users can refer to expert users belonging to the company. External expert users can refer to expert users not belonging to the company but obtained through third-party channels.

[0067] For example, a second user can be a user belonging to the same group as the first user. The system can query the digital assistant library for the second user's digital assistant and identify that second user's digital assistant as their own. Similarly, a second user can be an expert user. The system can query the digital assistant library for the expert user's intelligent agent or AI model and identify that intelligent agent or AI model as their own digital assistant. The digital assistant library can be used to store multiple digital assistants for multiple users, as well as multiple intelligent agents or AI models for multiple expert users.

[0068] Furthermore, after obtaining the second user's digital assistant, the second user's digital assistant can be stored in the first user's personal space.

[0069] Furthermore, the users in the first user's network of relationships can include: users related to the first user based on the first user's identity information, such as job information or group information, or any one or more other types of information.

[0070] It is understandable that the identity information of the first and second users can be processed with user authorization and through technical means to eliminate the risk of privacy leakage while preserving data availability. Sensitive identity information is desensitized, including but not limited to replacement, concealment, generalization, encryption, synthesis, tokenization, and other methods.

[0071] In this embodiment, personalized skills also include the first user's personal skills. Personal skills can also be obtained from a skill library. The first user can select the skills they need from the skill library and store them in their personal space to form personalized skills. Personal skills may include, for example, skills such as: developing code, generating warranty cards, querying orders, and reporting to experts.

[0072] In this embodiment, personalized skills also include the first user's personal tools. These personal tools can be obtained from a tool library. The first user can select the tools they need from the tool library and store them in their personal space, thus forming personal tools. Personal tools may include, for example, email tools, microcode tools, external tools, tools from the XI toolset, etc.

[0073] In this embodiment, the personalized skill also includes the first user's personal knowledge base. The personal knowledge base may include one or more files, such as development documents, self-test manuals, etc. Of course, the files in the first user's personal knowledge base can be obtained from a public knowledge base. The public knowledge base may include files with public attributes that are accessible to all users.

[0074] In this embodiment, personal memory may include expert experience information, which can support the reuse of strategies for complex tasks. Expert experience information may include, for example, historical processing strategies for various tasks, common problems and solutions, and various development documents. For example, historical processing strategies may include information such as calling "microcode tools" to complete development tasks and calling email tools to send emails.

[0075] In this embodiment, personal memory may include common-sense information. Common-sense information may include, for example, a company address, a supervisor's email address, or A's email address. B's supervisor is C's employee number is Static information such as common reporting formats. Using common sense to memorize information can avoid repeatedly asking basic questions.

[0076] In this embodiment, personal memory may include historical experience information. Historical experience information may include, for example, the user's historical usage records of the digital assistant, as well as the user's suggestions for modification and preference settings for similar tasks. Personalized task customization can be achieved through personal memory.

[0077] S204. Utilize your individual skills to perform at least one work task.

[0078] Optionally, S204 may include: the first digital employee using personalized memory to plan the first user's work tasks and obtain at least one work task.

[0079] A work task can refer to an executable unit obtained by the first digital employee based on user intent, goals, and other information, through task planning, including intent parsing, goal decomposition, dependency identification, and task sequence planning. A work task can specifically include: the executing entity (such as required tools), the executing object (such as input data), the operation content, constraints, and expected results.

[0080] In this embodiment, the work tasks may include, but are not limited to: data analysis tasks, model invocation tasks, tool invocation tasks, decision-making tasks, and result integration tasks.

[0081] In this embodiment, the first digital assistant of the first user can receive a first instruction triggered by the first user, and then autonomously plan one or more work tasks to be executed according to the instruction content in the first instruction. It can also retrieve personalized skills from the user's personal space to execute each task, realizing personalized and automated closed-loop processing from instruction to task execution. This not only accurately responds to the user's customized needs and makes the task execution fit the user's exclusive skills, avoiding the task and skill adaptation deviation caused by generalized processing, but also eliminates the need for manual task splitting and selection tools, greatly improving the efficiency and accuracy of task processing. At the same time, by reusing personalized skills from the personal space, the digital assistant's adaptability to different user scenarios is enhanced, helping users to efficiently complete various personalized work tasks.

[0082] In conjunction with the first aspect, in some implementations of the first aspect, personalized skills are utilized to perform at least one work task, including: Based on the task requirements of each work task, select the target skills needed to perform each work task from the personalized skills; and use the target skills needed to perform each work task to execute each work task.

[0083] Optionally, based on the task requirements of each work task, the target skills required to perform each task can be selected from personalized skills. This includes: if the task requirement is a standardized business process (e.g., generating an insurance card, submitting an expense report), the corresponding skills can be obtained from personalized skills. If the task requirement is a functional operation (e.g., sending an email, querying an order), the corresponding tool interface can be called. If the task requirement is a knowledge-based requirement, the latest policies or technical documents can be retrieved from the knowledge base and injected into the context as reasoning to complete knowledge generation. If the task requirement is to call an expert user's digital assistant, the expert user's digital assistant can be called to complete team-level collaboration.

[0084] Optionally, the entire task scheduling process is based on a globally unified context state to achieve collaborative scheduling management, and the execution results generated by each execution step are stored in the corresponding storage unit. For example, the planning execution results (PLAN_EXECUTE_RESULTS) are written to the planning execution result storage unit in real time, and the tool execution results (TOOL_EXECUTE_RESULTS) are written to the tool execution result storage unit, thereby ensuring the traceability, auditability and recoverability of the entire task execution process.

[0085] In this embodiment, the corresponding target skill can be accurately matched from the user's personalized skills according to the specific needs of each task, and the task can be executed. This achieves precise adaptation and efficient execution of skills and tasks, allowing each task to be efficiently advanced through the adapted personalized skills. This ensures the accuracy and adaptability of task execution, and eliminates the need for manual skill selection and task allocation. It significantly improves the efficiency of multi-task parallel processing. At the same time, the application of personalized skills can make task execution more in line with the user's personalized task execution scenario, further enhancing the work capabilities and effectiveness of digital assistants for different users.

[0086] In conjunction with the first aspect, in some implementations of the first aspect, personalized skills include at least one of the following: The digital assistant of the second user, where the second user refers to a user associated with the first user; The first user's personal skills refer to the skills that the first user possesses to perform specific technical services; The first user's personal tools refer to plugins / software set up by the first user that can be directly invoked; The first user's personal knowledge base refers to the knowledge base formed by the first user's knowledge files.

[0087] In this embodiment, by setting the personalized skills of the first user to include a digital assistant of the second user associated with the first user, the first user's personal skills, directly accessible personal tools, and a dedicated personal knowledge base, a comprehensive and customized support is provided for the digital assistant's work capabilities. This makes the skills cover a wider range of scenarios, not only avoiding the limitations of a single skill and greatly improving the adaptability and implementation capability of the digital assistant in dealing with complex tasks, but also making task execution more in line with the user's dedicated resource configuration and usage habits, effectively improving the personalized service value of the digital assistant and the reuse efficiency of work skills.

[0088] In conjunction with the first aspect, some implementations of the first aspect also include: If the target skill for any work task includes at least one of the following: a digital assistant, personal skills, personal tools, and personal knowledge base required to perform the work task, then the work task is performed using at least one of the following: a digital assistant, personal skills, personal tools, and personal knowledge base.

[0089] Optionally, the digital assistant required to perform the work task may include: a first digital assistant of a first user and / or a second digital assistant of a second user. Personal skills, personal tools, and personal knowledge bases can be referred to the description in the above embodiments, and will not be repeated here.

[0090] For example, if the job task is to send an email to a third user, the target skills for performing this job task include: The first digital assistant needed to perform work tasks; personal tools include email sending tools and third-party email addresses in a personal knowledge base.

[0091] For example, if the job task is: to allow "User B" to reply to routine emails on my behalf this week, then the target skills for performing this job task include: The second digital assistant for user B, required to perform work tasks, includes personal tools such as an email sending tool and a third user's email address within a personal knowledge base.

[0092] Optionally, if the digital assistant required to perform the work task includes a second digital assistant of the second user, then the execution result of the second digital assistant of the second user performing the work task can be obtained, and the execution result of the work task can be fed back to the first digital assistant, which will then store the execution result.

[0093] In this embodiment of the application, by setting target skills for work tasks, namely, at least one of digital assistants, personal skills, personal tools, and personal knowledge bases, flexible adaptation of executable skills corresponding to each work task is achieved, realizing efficient connection between task execution and diverse personalized skills, ensuring the smooth execution of various work tasks, while simplifying the judgment logic of skill matching and improving the flexibility and efficiency of task execution.

[0094] In conjunction with the first aspect, some implementations of the first aspect also include: In response to a selection action performed by the first user on at least one of the digital assistants, skill libraries, tool libraries, and knowledge bases contained in their team space, the first user's personalized skills are obtained.

[0095] Optionally, the digital assistant contained in the team space of the first user may include the user's own first digital assistant or the second digital assistant of the second user in the same team space.

[0096] It is understandable that a team space can refer to a team identifier, a list of team members, or a space ownership configuration that is bound to the user identifier of the first user. Based on the mapping relationship between the user identifier and the team space, the team space that the first user currently belongs to or has access to is determined, so as to achieve accurate matching and ownership confirmation between the first user and the corresponding team space.

[0097] In this embodiment, the first user selects personalized skills from the digital assistant, skill library, tool library, and knowledge base included in the team space, achieving team-level sharing of work skills and realizing a capability-sharing model of "one person develops, everyone reuses." Customized work skills ensure efficient linkage between personalized skills and team resources, allowing the digital assistant's capabilities to not only meet the user's specific needs but also reuse high-quality team resources, further enhancing the flexibility of skill configuration and the adaptability of task execution.

[0098] like Figure 3 The diagram shown is another flowchart of a digital assistant usage method provided in this application embodiment. The digital assistant usage method includes the following steps: S301, The first digital assistant receives a first instruction triggered by the first user, the first instruction including the instruction content provided by the first user.

[0099] S302. Identify the user intent of the first user from the instruction content of the first instruction, so as to obtain the target task corresponding to the user intent.

[0100] S303. Extract task information related to the target task from the first user's personal memory.

[0101] As mentioned above, personal memory can include at least one of the following: expert experience information, common sense memory information, and historical experience information.

[0102] Optionally, S303 may include: extracting task information related to the target task from the first user's expert experience information, common sense memory information, and / or historical experience information.

[0103] Specifically, the intent of the target task can be analyzed, and information matching the intent of the target task can be retrieved from the first user's expert experience information, common sense memory information, and / or historical experience information. The information matching the intent of the target task is identified as task information related to the target task.

[0104] S304. Based on the task information related to the target task, perform task planning for the target task to obtain at least one working task corresponding to the target task.

[0105] Optionally, task information related to the target task can be injected into the task planning stage as part of the contextual prompt.

[0106] S305. Obtain personalized skills from the first user's personal space; personalized skills include at least one of the following: the second user's digital assistant, the first user's personal skills, the first user's personal tools, and the first user's personal knowledge base.

[0107] S306. Utilize personalized skills to perform at least one work task.

[0108] In this embodiment, the user's intent is accurately identified from the instruction content contained in the user's command to determine the target task. Combined with the first user's personal memory, task information related to the target task is extracted. Based on this task information, task planning is carried out, achieving personalized and precise breakdown of work tasks. The task planning process has a higher degree of alignment with the user's historical behavior and memory, avoiding deviations caused by generalized planning, providing precise support for task breakdown, ensuring the integrity and adaptability of the task flow, improving the efficiency and accuracy of task planning, and effectively enhancing the personalized service capabilities and task execution effects of the digital assistant.

[0109] In the task planning stage, a multi-layered planning approach can be used. As an example, based on task information related to the target task, task planning is performed to obtain at least one corresponding work task, including: Generate enhanced prompts based on task information related to the target task; Based on the enhanced prompts, construct a task framework corresponding to the target task. The task framework includes: a strategic layer, at least one tactical layer in the strategic layer, and at least one execution layer in each tactical layer. Based on the framework structure corresponding to the strategic layer, generate strategic information for the strategic layer; Based on the preset task structure, the task is decomposed at each tactical layer to obtain the atomic tasks and / or composite tasks corresponding to each tactical layer. Based on the preset execution structure, the atomic tasks and / or composite tasks of the corresponding tactical layer are decomposed into steps at the execution layer to obtain the steps to be executed for each atomic task and the steps to be executed for the composite task. Each atomic task and each composite task whose steps need to be performed is identified as at least one working task corresponding to the target task.

[0110] Optionally, enhanced prompts can be generated based on task information related to the target task, which may specifically include: Enhanced prompts are generated based on preset enhanced prompt rules. These enhanced prompt rules are represented by the symbol f. Enhanced prompts = f(user question, task information related to the target task) The task information related to the target task may include, for example, tool libraries, historical experience, knowledge bases, SOP templates, etc.

[0111] Optionally, the mission framework may include a strategic layer, a tactical layer, and an execution layer.

[0112] The strategic framework can include: title, thought, and steps.

[0113] A theme can refer to a summary of the target task. An idea can indicate a requirement refactoring specification for the target task. Steps refer to the nested structure or relationship between the tactical and execution layers. For example, the strategic layer can include: Tactical Layer 1 and Tactical Layer 2. Tactical Layer 1 can include Task 1 and Task 2. Tactical Layer 2 can include Task 3 and Task 4.

[0114] It is understandable that the tasks at the tactical level can be the smallest executable tasks, or they can be complex tasks with nested tasks or multiple levels of tasks.

[0115] The tactical layer task structure can include: atomic-level task structures (JobStep) and composite-level task structures (TaskStep). Atomic tasks can be constructed using atomic-level task structures, and composite tasks can be constructed using composite-level task structures.

[0116] An atomic-level task structure may include at least one of the following fields: The Id (identifier) ​​field is used to uniquely identify the current atomic task; The name field is used to identify the name of the execution tool or function; The description field is used to describe the function and purpose of the atomic task; The type field is configured to identify the task type as an atomic task type; The parameters field is used to explicitly declare all tool input parameters. All parameters are explicitly defined through this field and are not passed implicitly. The constraints field is used to store execution constraint information, including at least one of the following: call_limit, output_type, and whether manual review is required_review.

[0117] A composite-level task structure may include at least one of the following fields: The id field is used to uniquely identify the current composite task; The name field is used to identify the workflow name; The description field is used to describe the overall functionality of the composite task; The type field is configured to identify the task type as workflow. The step_by_step field is used to store an ordered set of substeps, which are arranged in the order of execution. The constraints field is used to store the execution constraints of the composite task, including at least one of the following: execution order (execution_order), output type (output_type), template file name (ppt_template), and whether manual review is required (requires_review).

[0118] An execution structure may include: The id field is used to uniquely identify the current sub-step; The paragraph field is used to identify report paragraphs or output chapter titles; The requirement field is used to store the execution requirements and constraints of the sub-step. The tool structure includes a tool name field and a parameters field that explicitly defines query conditions and parameters.

[0119] For example, an example of an atomic task may include the following fields: "id":"step_01"; "name": "platform_email_send_email"; "description":"Send project launch notification email"; "type": "task"; "parameters":{ "recipients": ["li.xx@company.com", "wang.xx@company.com"];"subject":"

Solid State Battery Project

[0120] Atomic tasks may also include "constraints" information, which includes the following fields: "call_limit": 1; "output_type": "text"; "requires_review": false.

[0121] Among them, the parameters are explicitly declared: all tool parameters must be explicitly defined through the parameters field (unlike the implicit passing in traditional solutions).

[0122] Execution constraint annotation: call_limit=1: Forces a single call (resolves issues with batch calls failing in email tools). requires_review=false: Automatically skips the manual review process.

[0123] For example, a composite task could have the following fields: A1, "id": "step_02", "name": "workflow_report_generation", "description": "Main content of generating technical feasibility report", and "type": "workflow", "step_by_step": [{ A11、"id":"sub_02_01","paragraph":"Material stability analysis","requirement": "Requires 2023 laboratory test data","tool":{"name": "retriever_lab_data","parameters":{"query": "Solid electrolyte ionic conductivity","time_range": "2023-01-01~2023-12-31"。

[0124] A12、"id": "sub_02_02","paragraph": "Mass Production Cost Forecast","requirement": "Comparison with Company A's Latest Solution","tool": {"name": "retriever_market_data","parameters":{"query": "Company A's Solid State Battery Cost","source_type": "competitor_report"} A13、"id": "sub_02_03","paragraph": "Integration Conclusion","requirement": "SWOT Analysis Framework Used","tool": {"name": "llm_summary","parameters":{"input_steps": ["sub_02_01", "sub_02_02"],"analysis_framework": "SWOT"}}}], A11, A12, and A13 are all subtasks of task A1.

[0125] A2. "constraints":{"execution_order": "sequential","output_type": "ppt","ppt_template": "tech_report_v3.pptx","requires_review": true} A2 represents the constraints for the composite task.

[0126] Table 1 below shows an example of a task framework corresponding to a target task provided in an embodiment of this application.

[0127]

[0128] In this embodiment, enhanced prompts are generated based on target task-related information. A hierarchical task framework, comprising strategic, tactical, and execution layers, is then constructed based on these enhanced prompts. This enables structured, hierarchical decomposition and standardized management of target tasks, resulting in clear task hierarchy and logical order. Specific execution tasks are further generated from atomic and composite tasks at the tactical layer. This refines high-level objectives into directly executable units, significantly improving the completeness, logic, and executability of task planning. Simultaneously, it allows for the top-down, layered implementation of complex objectives, effectively reducing task execution complexity and enhancing the planning efficiency and execution reliability of the digital assistant for complex tasks.

[0129] In one possible design, before identifying the user's intent from the instruction content of the first instruction to obtain the target task corresponding to the user intent, the following steps are also included: Based on task information related to the target task, task planning is performed on the target task to obtain at least one corresponding work task, which also includes: Determine the current planning status corresponding to the first instruction, which is either a user-initiated request to adjust the plan, a re-planning process after task execution failure, or the first planning attempt. When the planning status is initial planning, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task. When the user actively requests to adjust the planning status, the task adjustment information provided by the first user is collected, the task information related to the target task is updated according to the task adjustment information, and the target task is re-planned according to the updated task information to obtain at least one working task corresponding to the target task. When the planning status is "task execution failed and replanning", collect relevant information about the previous task execution failure; based on the relevant information about the previous task execution failure, update the task information related to the target task, and replan the target task based on the updated task information to obtain at least one working task corresponding to the target task.

[0130] In this embodiment, the planning state corresponding to the first instruction is determined, such as user-initiated adjustment, replanning after execution failure, or initial planning. The corresponding plan is then executed according to different planning states, achieving precise control over the entire lifecycle of task planning. This enables accurate identification and differentiated processing of planning states across multiple scenarios, significantly improving the flexibility, reliability, and adaptability of task planning, while simultaneously meeting the needs for automated execution and collaborative manual intervention.

[0131] As another embodiment, the process of planning the target task based on task information related to the target task further includes: Based on the task planning progress, a progress bar is generated to indicate the execution status of the task plan. And / or, monitor abnormal information in real time during the task planning and execution process. When a preset error message is detected, trigger the planning status to be set to task execution failure and replanning to start the task planning repair process. And / or, through a preset manual intervention interface, collect user confirmation or modification instructions on error repair results or task planning.

[0132] In this embodiment, a visual progress bar is generated based on the task planning progress, which can intuitively and in real time prompt the task execution status, improving the user's perception and control of the task progress; abnormal information during the execution process is monitored in real time, and when preset error information is detected, the task execution failure replanning state is automatically triggered and the repair process is started, which can effectively improve the fault tolerance and self-healing ability of task execution; through a preset manual intervention interface, the user's confirmation and modification instructions on error repair results or task planning are collected, realizing the organic combination of automated planning and manual decision-making, which not only ensures the efficiency of task execution, but also enhances the flexibility and controllability of the system, thereby improving the overall reliability, adaptability and user experience of task planning.

[0133] In conjunction with the first aspect, in some implementations of the first aspect, after obtaining at least one working task corresponding to the target task, the method further includes: Output the target task and at least one corresponding work task for the first user to view and confirm; In response to the confirmation operation triggered by the first user on the target task and at least one work task, at least one work task is obtained; Alternatively, in response to a modification operation triggered by a first user on a target task or at least one work task, obtain the target task or at least one work task modified by the first user.

[0134] Optionally, in response to a modification operation triggered by a first user on a target task or at least one work task, obtaining the modified target task or at least one work task may include: in response to a modification operation triggered by a first user on a target task or at least one work task, obtaining supplementary information input by the first user, modifying the target task or at least one work task based on the supplementary information input by the first user, and obtaining the modified target task or at least one work task.

[0135] Specifically, based on the supplementary information input by the first user, the task content of the target task can be supplemented to obtain a new target task. Then, task planning can be performed on the new target task to obtain at least one corresponding work task. Alternatively, based on the supplementary information input by the first user, the enhanced prompts can be updated. Using the updated enhanced prompts, task planning can be performed on the target task to obtain at least one corresponding work task.

[0136] In this embodiment, a target task and at least one corresponding work task are output for user review and confirmation. After reviewing the target task and each work task, the user can confirm or modify them. This means the system can respond to the first user's confirmation or modification operation, thus clarifying the final target task and at least one work task. Through user interaction, the first user can clearly understand the task execution logic, thereby adjusting the task content according to actual needs. This ensures a precise match between the work task and the user's actual needs. Simultaneously, the user interaction process reduces the probability of task execution failure or errors, improving the accuracy and effectiveness of task implementation.

[0137] In conjunction with the first aspect, in some implementations of the first aspect, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task, further including: If the first user has the permission to execute the target task, the target task is planned based on the task information related to the target task, and at least one work task corresponding to the target task is obtained.

[0138] Optionally, the first user can also be authenticated, so that if the first user is successfully authenticated and has the execution permission for the target task, the target task can be planned according to the task information related to the target task, and at least one work task corresponding to the target task can be obtained.

[0139] To verify the identity of the first user, the user's identity information can be collected in real time, and the user's execution permissions can be securely authenticated using this information. For example, the user's facial image can be collected and compared with the user's facial image in their personal space. If the comparison is successful, it is determined that the first user has the permission to execute the target task.

[0140] To confirm whether the first user has the necessary permissions to execute the target task, the permission requirements for executing the target task can be obtained. If the first user's permission information matches the permission requirements, then the first user has the permission to execute the target task; otherwise, the first user does not have the permission to execute the target task. The permission requirements for executing the target task can be determined based on the specific task information. The first user's permission information can be obtained from the first user's personal space.

[0141] The first user's permission information refers to one or more permissions that the first user possesses. Permission requirements may include, for example, the need for one or more permissions. Permissions can refer to the ability to perform a certain task, such as the permission to send an email, the permission to query certain data, or the permission to access reports, etc.

[0142] Matching the first user's permission information with the permission requirement means that one or more permissions in the first user's permission information, including one or more permissions required by the permission requirement. For example, if a user's first permission information includes the permission to send emails and the permission to access reports, and the permission requirement is the permission to access reports, then the first user's permission information matches that permission requirement.

[0143] If one or more permissions in the first user's permission information do not include all of the required permissions, then the first user's permission information does not match the permission requirements. For example, if a user's first permission information includes the permission to send emails and the permission to access reports, but the permission requirements are the permission to send emails and the permission to query certain data, then the first user's permission information does not match the permission requirements.

[0144] In this embodiment, user execution permission verification is added to the task planning stage. Only when the first user has the permission to execute the target task will the target task be broken down into specific work tasks based on relevant task information. This achieves precise linkage between permission control and task planning. It can avoid task execution obstruction or illegal operations due to insufficient permissions, ensuring the security and compliance of task execution, ensuring the accurate implementation of task planning within the scope of permissions, and making the work task breakdown both in line with user needs and historical information. This reduces subsequent task execution failures or errors caused by permission issues, and improves the smoothness and overall efficiency of task execution.

[0145] In conjunction with the first aspect, some implementations of the first aspect also include: If missing information is detected during the execution of at least one task, a supplementary prompt message is output, prompting the first user to supplement the missing information. Collect missing information supplemented by the first user, and use the missing information and personalized skills to perform at least one job task.

[0146] In other words, during the execution of at least one task, the system continuously monitors the completeness of the information required for the current task. If missing information is detected, supplementary prompts are generated and output to guide the first user in obtaining the necessary information. The system collects the missing information provided by the first user based on the supplementary prompts and integrates this information with pre-configured or learned personalized skills. Based on the supplemented information and personalized skills, the system completes the task information required for execution. Then, based on the task information related to the target task, it performs task planning to obtain at least one corresponding task. This ensures that the information required for task execution is complete, the execution process is coherent, and it meets the user's personalized needs, thereby improving the success rate and accuracy of task execution.

[0147] In this embodiment, during the execution of at least one task, missing information is detected in real time. Supplementary prompts are output to guide the first user to complete the missing information. The task execution continues based on the supplemented information and personalized skills, achieving dynamic completion and execution of tasks, ensuring the completeness and accuracy of task completion. This allows for resolving missing information issues during task execution, avoiding restarting the task execution process, and improving the fault tolerance and success rate of task processing.

[0148] like Figure 4 The diagram shown is another flowchart of a digital assistant usage method provided in this application embodiment. The digital assistant usage method includes the following steps: S401, User Input: The first digital assistant receives a first instruction triggered by the first user, the first instruction including the instruction content provided by the first user.

[0149] S402. Determine user intent: Identify the user intent of the first user from the instruction content of the first instruction to obtain the target task corresponding to the user intent.

[0150] S403. Obtain task-related information: Extract task information related to the target task from the first user's personal memory. Personal memory includes: expert experience information 4031, common sense memory information 4032, and historical experience information 4033.

[0151] S404. Task Planning: Based on task information related to the target task, perform task planning for the target task to obtain at least one working task corresponding to the target task.

[0152] In one possible design, it also includes: if task planning is performed on the target task based on task information related to the target task, and if at least one work task corresponding to the target task is not obtained, the planning fails and returns to execution S401.

[0153] S405, Waiting for User Confirmation Planning: Output the target task and at least one corresponding work task for the first user to view and confirm; in response to the first user's confirmation operation on the target task and at least one work task, obtain at least one work task; or, in response to the first user's modification operation on the target task or at least one work task, obtain the first user's modified target task or at least one work task.

[0154] S406. Task Execution Scheduling: Based on the task requirements of each work task, select the target skills required to execute each work task from the personalized skills; and use the target skills required for each work task to execute each work task.

[0155] Personalized skills include at least one of the following: The second user's digital assistant 4061, where the second user refers to a user related to the first user; the first user's personal skills 4062, where personal skills refer to the skills possessed by the first user to perform specific technical services; the first user's personal tools 4063, where personal tools refer to plugins / software set up by the first user that can be directly invoked; and the first user's personal knowledge base 4064, where personal knowledge base refers to the knowledge base formed by the first user's knowledge files.

[0156] Optionally, if at least one task fails to execute, return to S404 to reschedule the task.

[0157] S407. Waiting for user to supplement information: If information is missing during the execution of at least one task, a supplementary prompt message is output, prompting the first user to supplement the missing information; the missing information supplemented by the first user is collected, and at least one task is executed using the missing information and personalized skills.

[0158] S408. Task execution result: Obtain the execution result of at least one work task.

[0159] As an example, before executing S404 and task planning, the method further includes: S409, permission verification: that is, determining whether the first user has the permission to execute the target task. If so, S404 is executed to plan the target task based on the task information related to the target task, obtaining at least one work task corresponding to the target task. If not, S410, insufficient permission prompt, that is, outputting a permission prompt message to indicate that the first user does not have the permission to execute the target task. Then, S401 is executed.

[0160] The following examples illustrate specific application scenarios of the embodiments of this application.

[0161] Application Scenario 1: A user issues a first instruction to their digital assistant: "Schedule a quarterly debriefing meeting with the East China sales team and generate a draft of the meeting minutes." Upon receiving this instruction, the digital assistant first performs intent recognition. It then uses a relevant intent recognition model to analyze the request as a complex task requiring multi-step planning.

[0162] For example, the system can retrieve the first user's identity information, preferred work hours, and time zone from their personal space. The memory module can then retrieve the member list and frequently used contact information of the "H Zone Sales Team" from the first user's personal memory, as well as query standard agenda templates and common reporting formats for "quarterly reviews" from historical meeting minutes, and utilize the latest sales data from the enterprise knowledge base.

[0163] Before task planning, it is necessary to generate enhanced prompts based on the task information related to the target task, and then execute task planning based on the enhanced prompts.

[0164] During the task planning phase, the first digital assistant can first construct a task framework for the target task, completing the framework construction from the strategic layer, tactical layer to the execution layer, and obtaining at least one work task that needs to be executed: First, call the calendar system to query the available time slots of the first user and the East China team members.

[0165] Second, initiate meeting invitations based on available time and wait for confirmation from key members.

[0166] Third, once the meeting time is confirmed, an online meeting link will be automatically created and synchronized to the calendars of all participants.

[0167] Fourth, pull the sales data for the East China region for this quarter from the BI system and generate visualization charts.

[0168] Fifth, the system automatically generates a draft of meeting minutes based on preset templates and meeting content (such as key indicators, trend analysis, and problem summaries).

[0169] After receiving the above tasks, each task is presented to the user in a structured format with prompts such as: "Confirm execution? If adjustments are needed, please provide your suggestions." Once user confirmation is detected, the task scheduling module begins executing at least one task one by one.

[0170] Furthermore, during the meeting invitation phase, which is the execution of the third task, the first digital assistant determines that a certain participant has no available time slots for the next three days. Therefore, it pauses the process and notifies the first user to provide a new time. After detecting the first user's reply that "next Tuesday morning is acceptable," the first digital assistant can reconfirm the available time slots for each participant and successfully lock the time. Afterward, the first digital assistant can continue with subsequent processes, such as sending meeting invitations to each participant and receiving feedback from them.

[0171] After the meeting invitation is completed, the first digital assistant can proceed to the fourth task, which is to call the BI interface to obtain sales data and generate visualization charts. Then, it will perform the fifth task, which is to automatically generate a draft of the meeting minutes based on a preset template and meeting content (such as key indicators, trend analysis, and a summary of issues), and save the generated draft of the meeting minutes to the first user's designated shared directory.

[0172] Understandably, the status, results, and exception information of each step in the entire task execution process are written to context variables in real time to ensure full traceability. After the task is completed, the system automatically stores the meeting organization process, parameter configuration, user feedback, and typical problem handling methods as "experience" in the memory bank (such as the first user's personal memory bank and the public memory bank) for reuse in similar tasks in the future or for querying related meeting records.

[0173] Application Scenario 2: User 1, unable to handle daily tasks due to a temporary business trip, issues a first instruction to their digital assistant: "Allow User B to view my to-do list and reply to regular emails this week, but not to access financial documents or modify system settings." Upon receiving this instruction, the digital assistant first performs intent recognition to determine the target task. Based on the target task, it extracts relevant task information from the user's personal memory. Then, based on the task information, it analyzes that the request involves granting external access permissions, classifying it as a sensitive operation, and immediately initiates the access control process.

[0174] The first digital assistant first authenticates the first user's identity, that is, confirms that the current operator is the first user. This can be done by verifying the USER_ID (user identifier) ​​in the login context with the authentication mechanism to ensure that the authorized behavior is legitimate.

[0175] Next, the first digital assistant retrieves "User B's" organizational identity information from the personalized skills in the user's personal space, confirming that they are a colleague in the same department and possess basic collaboration permissions, meeting the authorization prerequisites. The system then analyzes the key elements of the instruction: the authorized object is "User B," the time frame is "within this week," and the allowed operations include "viewing to-do items" and "replying to regular emails." Prohibited operations are explicitly limited to "accessing financial documents" and "modifying system settings." Based on this, the first digital assistant generates a structured temporary authorization policy through the permission engine and writes it to the security policy storage area. The policy includes an automatic expiration mechanism to ensure that the authorization automatically expires after seven days. Simultaneously, the first digital assistant establishes a "proxy session" marker in the context. All subsequent operations performed by User B through their second digital assistant will be marked as "proxy operations" and logged, which the first user can view at any time.

[0176] When User B requests a reply to a client's email regarding delivery time, their second digital assistant re-verifies the user's permissions before execution. Confirming the operation falls under the "regular email" category and does not involve sensitive information, the assistant uses email tools to compose and send the email. However, when User B attempts to query "last quarter's expense details," the system detects that the request violates the "financial documents" restriction, immediately halting the operation and returning the message: "You do not have permission to access this information. Please contact the first user for authorization." Throughout the process, the second digital assistant operates with the first user's will at its core, ensuring all proxy actions are executed within pre-defined boundaries. This improves collaboration efficiency while guaranteeing data security. After the task is completed, the authorization policy automatically expires without manual intervention, achieving a secure, flexible, and traceable digital assistant proxy mechanism.

[0177] like Figure 5 The diagram shown is a structural schematic of a digital assistant user device provided in an embodiment of this application. The digital assistant user device 500 includes a transceiver unit 501 and a processing unit 502.

[0178] The transceiver unit 501 is used to receive a first instruction triggered by a first user, the first instruction including instruction content provided by the first user.

[0179] The processing unit 502 is used to plan at least one work task to be executed according to the instruction content in the first instruction; obtain personalized skills in the personal space of the first user; and use the personalized skills to execute at least one work task.

[0180] As one embodiment, the processing unit utilizes personalized skills to perform at least one task, specifically including: Based on the task requirements of each work task, select the target skills needed to perform each work task from the personalized skills; and use the target skills needed to perform each work task to execute each work task.

[0181] As yet another example, personalized skills include at least one of the following: The digital assistant of the second user, where the second user refers to a user associated with the first user; The first user's personal skills refer to the skills that the first user possesses to perform specific technical services; The first user's personal tools refer to plugins / software set up by the first user that can be directly invoked; The first user's personal knowledge base refers to the knowledge base formed by the first user's knowledge files.

[0182] As yet another embodiment, the processing unit is also used for: If the target skill for any work task includes at least one of the following: a digital assistant, personal skills, personal tools, and personal knowledge base required to perform the work task, then the work task is performed using at least one of the following: a digital assistant, personal skills, personal tools, and personal knowledge base.

[0183] As yet another embodiment, the processing unit is also used for: In response to a selection action performed by the first user on at least one of the digital assistants, skill libraries, tool libraries, and knowledge bases contained in their team space, the first user's personalized skills are obtained.

[0184] As another embodiment, the processing unit executes at least one work task planned to be performed according to the instructions in the first instruction, specifically including: Identify the user intent of the first user from the instruction content of the first instruction to obtain the target task corresponding to the user intent; extract task information related to the target task from the first user's personal memory; perform task planning on the target task based on the task information related to the target task to obtain at least one work task corresponding to the target task.

[0185] As another embodiment, the processing unit performs task planning on the target task based on task information related to the target task, and obtains at least one working task corresponding to the target task, specifically including: Based on task information related to the target task, enhanced prompts are generated. Based on these prompts, a task framework corresponding to the target task is constructed. This framework includes a strategic layer, at least one tactical layer within the strategic layer, and at least one execution layer within each tactical layer. Based on the framework structure corresponding to the strategic layer, strategic information for that layer is generated. Based on a preset task structure, tasks are decomposed at each tactical layer to obtain atomic tasks and / or composite tasks corresponding to each tactical layer. Based on a preset execution structure, the atomic tasks and / or composite tasks of the corresponding tactical layers are broken down into steps at the execution layer to obtain the steps required for each atomic task and the steps required for each composite task. Each atomic task and each composite task with known steps to be executed is identified as at least one working task corresponding to the target task.

[0186] As another embodiment, the processing unit performs task planning on the target task based on task information related to the target task, and obtains at least one working task corresponding to the target task, specifically including: Determine the current planning status corresponding to the first instruction, which is either a user-initiated request to adjust the plan, a re-planning process after task execution failure, or the first planning attempt. When the planning status is initial planning, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task. When the user actively requests to adjust the planning status, the task adjustment information provided by the first user is collected, the task information related to the target task is updated according to the task adjustment information, and the target task is re-planned according to the updated task information to obtain at least one working task corresponding to the target task. When the planning status is "task execution failed and replanning", collect relevant information about the previous task execution failure; based on the relevant information about the previous task execution failure, update the task information related to the target task, and replan the target task based on the updated task information to obtain at least one working task corresponding to the target task.

[0187] As another embodiment, the process of planning the target task based on task information related to the target task further includes: Based on the task planning progress, a progress bar is generated to indicate the execution status of the task plan. And / or, monitor abnormal information in real time during the task planning and execution process. When a preset error message is detected, trigger the planning status to be set to task execution failure and replanning to start the task planning repair process. And / or, through a preset manual intervention interface, collect user confirmation or modification instructions on error repair results or task planning.

[0188] As yet another embodiment, the processing unit is also used for: Output the target task and at least one corresponding work task for the first user to view and confirm; In response to the confirmation operation triggered by the first user on the target task and at least one work task, at least one work task is obtained; Alternatively, in response to a modification operation triggered by a first user on a target task or at least one work task, obtain the target task or at least one work task modified by the first user.

[0189] As another embodiment, the processing unit performs task planning on the target task based on task information related to the target task, and obtains at least one working task corresponding to the target task, specifically including: If the first user has the permission to execute the target task, the target task is planned based on the task information related to the target task, and at least one work task corresponding to the target task is obtained.

[0190] As yet another embodiment, the processing unit is also used for: If missing information is detected during the execution of at least one task, a supplementary prompt message is output, prompting the first user to supplement the missing information. Collect missing information supplemented by the first user, and use the missing information and personalized skills to perform at least one job task.

[0191] In the embodiments of this application, Figure 5 The device shown can also be a chip or a chip system, such as a system on chip (SoC) or a baseboard management controller (BMC).

[0192] Figure 6 This is a hardware block diagram of a computing device provided in an embodiment of this application. The computing device 600 according to an embodiment of this application includes at least a memory 601, a processor 602, and a transceiver 603. The memory 601 is used to store computer programs, such as a first digital assistant, and the transceiver 603 is used to interact with a user. The processor 602 is used to execute the computer program, such as the first digital assistant, to implement the digital assistant usage method of any of the above embodiments.

[0193] In addition, the memory 601, processor 602, and transceiver 603 are all electrically connected to the bus 604.

[0194] Furthermore, embodiments of this application also provide a computer-readable storage medium for storing a computer program. When executed by a processor, the computer program implements the digital assistant usage method of any of the preceding embodiments of this application.

[0195] Computer-readable storage media include, but are not limited to, volatile storage media and / or non-volatile storage media. Volatile storage media may include, for example, random access storage media (RAM) and / or cache storage media. Non-volatile storage media may include, for example, read-only storage media (ROM), hard disks, flash memory, optical disks, magnetic disks, etc.

[0196] This application also provides a computer program product, including a computer program / instructions, which, when executed by a processor, implements the digital assistant usage method of any of the preceding embodiments of this application.

[0197] The basic principles of the embodiments of this application have been described above with reference to specific examples. However, it should be noted that the advantages, benefits, and effects mentioned in the embodiments of this application are merely examples and not limitations, and should not be considered as essential features of each embodiment of this application. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the embodiments of this application from necessarily employing the aforementioned specific details.

[0198] The block diagrams of devices, apparatuses, devices, and systems involved in the embodiments of this application are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as “comprising,” “including,” “having,” etc., are open-ended terms meaning “including but not limited to,” and are used interchangeably with them. The terms “or” and “and” as used herein refer to the terms “and / or,” and are used interchangeably with them unless the context explicitly indicates otherwise. The term “such as” as used herein refers to the phrase “such as but not limited to,” and is used interchangeably with it.

[0199] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0200] It should also be noted that in the systems and methods of this application embodiment, each component or step can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions of the embodiments of this application.

[0201] Various changes, substitutions, and modifications can be made to the technology herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of the embodiments of this application is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0202] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use embodiments of this application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other aspects without departing from the scope of embodiments of this application. Therefore, embodiments of this application are not intended to be limited to the aspects shown herein, but rather to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0203] The above description has been given for illustrative and descriptive purposes. Furthermore, this description is not intended to limit the embodiments of this application to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method of use of a digital assistant, the method comprising: include: The first digital assistant receives a first instruction triggered by a first user, the first instruction including the instruction content provided by the first user; Based on the instructions in the first instruction, plan at least one work task that needs to be executed; Obtain personalized skills from the first user's personal space; Utilize the personalized skills to perform the at least one work task.

2. The method of claim 1, wherein, The use of the personalized skills to perform the at least one work task includes: Based on the task requirements of each work task, select the target skills required to perform each work task from the personalized skills; Utilize the target skills required for each task to perform that task.

3. The method of claim 2, wherein, The personalized skills include at least one of the following: The digital assistant of the second user, where the second user refers to a user associated with the first user; The personal skills of the first user refer to the skills possessed by the first user to perform specific technical services; The personal tools of the first user refer to plugins / software set up by the first user that can be directly invoked; The first user's personal knowledge base refers to the knowledge base formed by the first user's knowledge files.

4. The method according to claim 3, characterized in that, Also includes: If the target skill for any work task includes at least one of the following: a digital assistant, personal skills, personal tools, and a personal knowledge base required to perform the work task, then the work task is performed using at least one of the following: a digital assistant, personal skills, personal tools, and a personal knowledge base.

5. The method according to claim 3, characterized in that, Also includes: In response to the first user's selection of at least one of the digital assistant, skill library, tool library, and knowledge base contained in their team space, the first user's personalized skills are obtained.

6. The method according to any one of claims 1-5, characterized in that, The step of planning at least one work task to be executed based on the instruction content in the first instruction includes: Identify the user intent of the first user from the instruction content of the first instruction, so as to obtain the target task corresponding to the user intent; Extract task information related to the target task from the first user's personal memory; Based on the task information related to the target task, task planning is performed on the target task to obtain at least one working task corresponding to the target task.

7. The method according to claim 6, characterized in that, After obtaining at least one work task corresponding to the target task, the method further includes: Output the target task and at least one corresponding work task for the first user to view and confirm; In response to the confirmation operation triggered by the first user for the target task and the at least one work task, the at least one work task is obtained; Alternatively, in response to a modification operation triggered by the first user on the target task or the at least one work task, the modified target task or at least one work task by the first user can be obtained.

8. The method according to claim 6, characterized in that, The step of performing task planning on the target task based on task information related to the target task to obtain at least one work task corresponding to the target task further includes: If the first user has the execution permission to execute the target task, the target task is planned according to the task information related to the target task, and at least one work task corresponding to the target task is obtained.

9. The method according to claim 6, characterized in that, The step of performing task planning on the target task based on task information related to the target task to obtain at least one work task corresponding to the target task includes: Based on the task information related to the target task, generate enhanced prompt information; Based on the enhanced prompt information, a task framework corresponding to the target task is constructed. The task framework includes: a strategic layer, at least one tactical layer in the strategic layer, and at least one execution layer in each tactical layer. Based on the framework structure corresponding to the strategic layer, generate the strategic information for the strategic layer; Based on the preset task structure, the task is decomposed at each tactical layer to obtain the atomic tasks and / or composite tasks corresponding to each tactical layer. Based on the preset execution structure, the atomic tasks and / or composite tasks of the corresponding tactical layer are decomposed into steps at the execution layer to obtain the steps to be executed for each atomic task and the steps to be executed for the composite task. Each atomic task and each composite task whose steps need to be performed is identified as at least one working task corresponding to the target task.

10. The method according to claim 6, characterized in that, The step of performing task planning on the target task based on task information related to the target task to obtain at least one work task corresponding to the target task further includes: Determine the current planning status corresponding to the first instruction, wherein the planning status is that the user actively requests to adjust the plan, the task execution fails and is re-planned, or the initial planning is performed; When the planning status is the initial planning stage, task planning is performed on the target task based on task information related to the target task to obtain at least one working task corresponding to the target task. When the planning status is that the user actively requests to adjust the planning, the task adjustment information provided by the first user is collected, the task information related to the target task is updated according to the task adjustment information, and the target task is re-planned according to the updated task information to obtain at least one working task corresponding to the target task. When the planning status is "task execution failure and replanning", relevant information about the previous task execution failure is collected; based on the relevant information about the previous task execution failure, the task information related to the target task is updated, and based on the updated task information, the target task is replanned to obtain at least one working task corresponding to the target task.

11. The method according to claim 6, characterized in that, The process of planning the target task based on task information related to the target task also includes: Based on the task planning progress, a progress bar is generated, which is used to indicate the execution status of the task plan; And / or, monitor abnormal information in real time during the task planning and execution process. When a preset error message is detected, trigger the planning status to be set to task execution failure and replanning to start the task planning repair process. And / or, through a preset manual intervention interface, collect user confirmation or modification instructions on error repair results or task planning.

12. A computing device, characterized in that, include: A processor and a memory, the memory storing a first digital assistant, the first digital assistant being invoked by the processor to perform the digital assistant usage method according to any one of claims 1-11.

Images