Digital twin agent method and apparatus, electronic device, and storage medium

Abstract

The application relates to a digital avatar agent method and device, electronic equipment and a storage medium. The method comprises: creating a digital avatar instance corresponding to a user account, the digital avatar instance being an agent object of the user account in a digital environment; in response to an interaction message, the digital avatar instance generates a first response within the authorized range of the digital avatar instance; and in the case that the interaction message exceeds the authorized range of the digital avatar instance, the interaction message is pushed to the user account, and a second response is generated based on the instruction returned by the user account. Thus, the continuity of business communication during the absence of the user can be ensured, the agent response can be controlled, and the accuracy and safety of the reply can be ensured through the user cooperation mechanism when the authorized range of the avatar is exceeded.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, and more particularly to a digital avatar proxy method, device, electronic device, and storage medium. Background Technology

[0002] In today's instant messaging-driven office environment, immediacy is central to organizational efficiency. However, when users are on leave, vacationing, or handling urgent external matters, they often fail to respond promptly to business communications, leading to delays in handling project inquiries or routine business needs. This increases waiting costs for colleagues and collaborators, causing a break in the organizational communication chain. To address these issues, some existing solutions attempt to introduce AI (Artificial Intelligence) assistants to help process business communications during user absences.

[0003] However, existing AI assistant solutions have significant shortcomings in practical applications. First, current AI assistants have limited processing capabilities when faced with complex scenarios involving interest coordination or creative decision-making. Furthermore, most current AI systems habitually adopt a strategy of directly refusing to answer when their capabilities exceed their limits. While this strategy avoids the risk of overstepping authority, it comes at the cost of disrupting business communication, further exacerbating the breakdown in communication and collaboration. Second, existing AI assistants generally lack controllable behavioral boundaries, leading to either risks of information leakage or unauthorized promises due to overly broad permissions, or an inability to effectively play a supporting role due to overly narrow permissions. Summary of the Invention

[0004] This application provides a digital avatar agent method, device, electronic device, and storage medium to solve the technical problem in the prior art that AI assistants cannot balance business communication continuity and behavioral boundary controllability when users are away from their posts.

[0005] In a first aspect, this application provides a digital avatar proxy method, the method comprising: Create a digital clone instance corresponding to the user account, which serves as a proxy object for the user account in the digital environment; In response to the interaction message, the digital clone instance generates a first response within the authorized scope of the digital clone instance; In addition, if the interaction message exceeds the authorization scope of the digital clone instance, the interaction message is pushed to the user account, and a second response is generated based on the instruction returned by the user account.

[0006] In one possible implementation, the method further includes: In the case of generating the first response or the second response, source identification information is added, which is used to indicate that the response content was generated by the digital clone instance or was generated with the confirmation of the user account.

[0007] In one possible implementation, the method further includes: In response to a takeover command issued by the user account, the digital clone instance suspends or terminates the currently ongoing response generation process and switches subsequent interactions to be handled by the user account.

[0008] In one possible implementation, the method further includes: After the user account is processed, the proxy state of the digital clone instance is restored, and it continues to respond to subsequent interaction messages.

[0009] In one possible implementation, the method further includes: Based on the activation conditions set by the user account, configure at least one of the following for the digital clone instance: an activation period, a disabling period, or an automatic activation rule.

[0010] In one possible implementation, the method further includes: The digital clone instance prioritizes the received interaction messages and processes them sequentially according to their priority.

[0011] In one possible implementation, the method further includes: The digital clone instance pushes a proxy work report to the user account, so that the user account can provide feedback to the digital clone instance based on the proxy work report; Based on the user account's feedback regarding the digital clone instance, the digital clone instance is optimized.

[0012] In one possible implementation, the method further includes: In response to the detection that preset security constraints are met, the digital clone instance switches from the current working mode to a restricted response mode; wherein, the preset security constraints include at least one of the following: the risk level of the interaction scenario exceeds a threshold, the completeness of the information that the digital clone instance can obtain is lower than a preset standard, and the boundaries of the operation permissions involved in the interaction message are unclear; In the restricted response mode, the output capability of the digital clone instance is limited to performing only at least one of the following operations: outputting a guiding response, outputting a status feedback response, or outputting a rejection response.

[0013] In one possible implementation, creating a digital clone instance uniquely corresponding to a user account includes: Obtain the user account's identity and permission information; The digital clone instance is created based on the identity identifier, and the authorization scope of the digital clone instance is configured according to the permission information.

[0014] In one possible implementation, generating a second response based on the instruction returned by the user account includes: The instructions returned by the user account are subjected to semantic parsing and reconstruction operations. Based on the identity and permission information of the sender of the interaction message, the disclosable information content is extracted to generate a second response.

[0015] In one possible implementation, the digital clone instance generates a first response within the authorized scope of the digital clone instance, including: The digital clone instance performs type identification on the interactive messages; Based on the identification results, at least one intelligent agent is invoked from among the multiple intelligent agents configured within the digital clone instance; wherein the invoked intelligent agent generates a sub-response within its corresponding task processing capability range; The digital clone instance integrates the sub-responses generated by each intelligent agent to generate the first response.

[0016] In one possible implementation, the method further includes: A digital personality model of the user account is constructed based on the user account's private domain data. The digital personality model includes language style features, business processing logic features, and decision-making pattern features. When the digital clone instance generates the first response, it performs style adaptation processing on the response content based on the digital personality model.

[0017] In one possible implementation, the method further includes: In response to the received lifecycle management command, a state change operation is performed on the digital clone instance; The state change operation includes at least one of the following: enabling the digital clone instance, freezing the digital clone instance, and destroying the digital clone instance.

[0018] In one possible implementation, the method further includes: The memory of the digital clone entity is constructed and maintained, and the memory is used to persistently store memory data that can be accessed across multiple interactive sessions.

[0019] Secondly, this application provides a digital clone agent device, the device comprising: The clone creation module is used to create digital clone instances corresponding to user accounts, and the digital clone instances serve as proxy objects for the user accounts in the digital environment; An autonomous response module is used to respond to interactive messages, wherein the digital clone instance generates a first response within the authorized scope of the digital clone instance; The collaborative decision-making module is used to push the interaction message to the user account when the interaction message exceeds the authorization scope of the digital clone instance, and generate a second response based on the instructions returned by the user account.

[0020] Thirdly, this application provides an electronic device, including: a processor and a memory, wherein the processor is configured to execute a digital clone agent program stored in the memory to implement the digital clone agent method described in any one of the first aspects.

[0021] Fourthly, this application provides a storage medium storing one or more programs that can be executed by one or more processors to implement the digital clone agent method described in any one aspect.

[0022] Compared with the prior art, the technical solution provided in this application has the following advantages: The method provided in this application creates a digital clone instance corresponding to a user account as a proxy object for the user account in the digital environment. This allows the digital clone instance to assume the proxy function of business interaction when the user account cannot intervene in real time, thereby ensuring the business continuity and communication accessibility of the user account when it is off-duty or in a low-response state, and avoiding business collaboration interruption or information blockage caused by the user's inability to reply in a timely manner. On this basis, this solution sets an authorization scope, so that the digital clone instance can only generate the first response autonomously within its authorized limits. This not only realizes the immediate automatic response to routine business inquiries, but also constrains the autonomous behavior of the digital clone instance within a controllable boundary, preventing the risk of information leakage or unauthorized commitments due to excessive automation. Furthermore, when the interaction message exceeds the authorization scope of the digital clone instance, this solution does not adopt the direct rejection strategy commonly used by existing AI systems, but pushes the interaction message to the user account, which then intervenes to process it and generates a second response based on the instructions returned by the user account. The aforementioned mechanism, on the one hand, ensures timely responses and efficient business processing within the authorized scope of autonomy, preventing user accounts from being excessively disturbed by high-frequency repetitive queries; on the other hand, when exceeding the authorized scope, it pushes information to the user account and generates a response based on the user's instructions, ensuring the accuracy and security of processing results in complex or sensitive scenarios, and ensuring that every interactive message receives an effective response. Therefore, this solution improves business communication efficiency while clarifying the responsibility boundary between proxy behavior and user behavior through permission constraints on digital avatar instances and the final decision-making intervention of user accounts, providing a reliable technical foundation for the controllable deployment of the avatar proxy mechanism in the organizational environment. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0024] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0026] Figure 1A flowchart illustrating an embodiment of a digital clone proxy method provided in this application; Figure 2 A block diagram illustrating an embodiment of a digital clone agent device provided in this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0028] The following disclosure provides numerous different embodiments or examples for implementing various structures of this application. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of this application. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0029] To address the technical problem of existing AI assistants failing to balance business communication continuity and behavioral boundary controllability during user absence, this application provides a digital avatar proxy method, device, electronic device, and storage medium. This enables continuous business communication and controllable proxy responses during user absence, and ensures the accuracy and security of responses through user collaboration mechanisms when the scope of the avatar's authorization is exceeded.

[0030] Figure 1 This is a flowchart illustrating an embodiment of a digital avatar proxy method provided in this application. Figure 1 As shown, the method includes the following steps: Step 101: Create a digital clone instance corresponding to the user account. The digital clone instance serves as a proxy object for the user account in the digital environment.

[0031] In step 101, a digital avatar instance corresponding to the user account is created. The user account is a unique identifier for the user in the digital office environment, through which the user logs into the system and exercises corresponding operational permissions. The digital avatar instance is a virtual proxy object generated for the user account. It acts as a proxy for the user account in the digital environment, performing communication and transaction processing functions on behalf of the user when the user cannot personally handle interactive tasks. For example, there is a one-to-one correspondence between the proxy object and the user account; that is, one user account corresponds to only one digital avatar instance to ensure identity consistency and traceability of behavior.

[0032] In one embodiment, the specific implementation of creating a digital clone instance corresponding to a user account includes: first, obtaining the user account's identity identifier and permission information, wherein the identity identifier is used to uniquely identify the user's identity, and the permission information is used to characterize the user's role in the organizational structure and the scope of data and operations that can be accessed; then, creating a digital clone instance based on the identity identifier; and finally, configuring the authorization scope of the digital clone instance according to the permission information, so that the digital clone instance can only perform operations within the authorized scope during subsequent operation.

[0033] Step 101 establishes a binding relationship between the digital clone instance and the user account, and assigns initial operational boundaries to the clone instance, thereby providing an identity and permission basis for the subsequent proxy response process.

[0034] Step 102: In response to the interaction message, the digital clone instance generates a first response within the authorized scope of the digital clone instance.

[0035] In step 102, in response to the interaction message, the digital clone instance generates a first response within its authorized scope. The interaction message refers to digital communication content initiated by the sender of the interaction message and directed to the user account, and its forms include, but are not limited to, instant messaging messages, emails, and in-system notifications. The sender of the interaction message refers to other users or systems that have a business interaction relationship with the user account. The authorized scope refers to the boundaries of operations and information access that the digital clone instance is allowed to perform, defining the types of business that the clone instance can handle, the data sources it can access, and the limits of its responses.

[0036] When a digital avatar instance receives an interaction message, it first analyzes the message's content and verifies whether the business type and information scope involved fall within the authorized scope. If the verification passes, the digital avatar instance, within the authorized scope, generates a response corresponding to the interaction message based on a pre-built knowledge system, serving as the first response. The generation of the first response includes, but is not limited to: retrieving knowledge data related to the interaction message within the authorized scope, organizing the response content based on the retrieval results, and outputting it in a style consistent with the user account's expression habits.

[0037] Through step 102, the digital clone instance can automatically handle business interactions within the authorized scope without real-time intervention from the user account, ensuring the immediacy and continuity of business communication.

[0038] In one embodiment, the digital clone instance generates a first response within the authorized scope of the digital clone instance, including: the digital clone instance performing type identification on the interactive message; based on the identification result, calling at least one intelligent agent from multiple intelligent agents configured within the digital clone instance; wherein the called intelligent agent generates a sub-response within its corresponding task processing capability; and the digital clone instance integrating the sub-responses generated by each intelligent agent to generate the first response.

[0039] In this embodiment, after receiving an interaction message, the digital clone instance first performs type identification on the message. Type identification refers to analyzing and classifying the business domain, problem category, or task nature involved in the interaction message to determine the type of capability required to process it. For example, the interaction message may involve different types such as business consultation, data query, document writing, or process collaboration, and different types of problems correspond to different processing logic and capability requirements.

[0040] Based on the type identification result, the digital avatar instance calls at least one agent from its internally configured agents that matches the type of the interaction message. An agent is an independent functional module with specific task processing capabilities, capable of independently completing a specific type of sub-task within its own task processing capabilities. Each agent corresponds to one or more specific task processing capabilities; different agents are functionally independent but logically collaborative. The digital avatar instance is pre-configured with multiple agents with different functions to accommodate diverse interaction message types. When an interaction message involves only a single capability requirement, the digital avatar instance calls the corresponding single agent to perform the processing; when an interaction message involves multiple capability requirements, the digital avatar instance can call multiple agents to collaboratively perform the processing.

[0041] Upon receiving the invocation instruction from the digital avatar instance, the invoked agent processes the interactive message or its related subtasks within its corresponding task processing capabilities and generates corresponding sub-responses. A sub-response is the result obtained by the agent after processing a portion or aspect of the interactive message based on its own capabilities. During the agent's operation, its scope of operation is uniformly controlled by the digital avatar instance; the agent only performs operations within the authorized scope of the digital avatar instance and does not have the authority to independently respond to external requests.

[0042] Each agent returns its generated sub-response to the digital clone instance, which then integrates these sub-responses. Integration includes merging, deduplication, sorting, logical coherence, and semantic refinement of the sub-responses to eliminate redundancy, conflicts, or poor connections between the outputs of different agents, resulting in a logically coherent and complete final response, which is then output as the first response to the sender of the interaction message.

[0043] Through the above methods, digital avatar instances can leverage multiple internally configured agents to collaboratively process complex interactive messages. Different types of tasks are handled by agents specializing in their respective domains, with the final output being uniformly aggregated by the digital avatar instance. This mechanism ensures the professionalism and accuracy of the response content while enabling the reuse and flexible orchestration of capability modules. It avoids the potential for insufficient accuracy or logical inconsistencies that may arise when a single model handles all types of tasks, thus improving the digital avatar instance's ability to handle diverse business inquiries and its response quality. Furthermore, since all agents are built into the digital avatar instance and do not possess independent external identities, the digital avatar instance always presents itself as a single, unified entity, ensuring consistency of identity and clarity of responsibility.

[0044] Step 103: If the interaction message exceeds the authorization scope of the digital clone instance, push the interaction message to the user account and generate a second response based on the instructions returned by the user account.

[0045] In step 103, if the interaction message exceeds the authorization scope of the digital clone instance, the digital clone instance pushes the interaction message to the user account and generates a second response based on the instructions returned by the user account.

[0046] Among these, exceeding the authorized scope includes, but is not limited to, the following situations: the operation requested by the interaction message involves data that the digital clone instance does not have the right to access or business operations that it does not have the right to execute; the interaction message involves business commitments, resource allocations, or decision-making permissions that exceed the preset boundaries of the clone instance; the questions contained in the interaction message exceed the reasoning ability of the clone instance's existing knowledge system.

[0047] When a digital avatar instance determines that an interaction message exceeds the authorized scope, it does not directly generate a response. Instead, it triggers a collaborative decision-making process, pushing the relevant content of the interaction message to the user account to request the user account's intervention. For example, the push method includes, but is not limited to: displaying the content of the interaction message in the user account's client interface and providing a corresponding processing entry point. After receiving the push, the user account can return processing instructions through the client interface. These instructions include, but are not limited to: directly inputting the final processing result, providing processing principles or restrictions, marking it as temporarily not processed, or rejecting processing.

[0048] After receiving the instruction returned by the user account, the digital clone instance generates a response content corresponding to the interaction message based on the instruction, which serves as a second response, and outputs the second response to the sender of the interaction message.

[0049] In step 103, when the interaction message exceeds the authorized scope of the digital clone instance, the executing entity of this application does not directly refuse to answer. Instead, it pushes the interaction message to the user account for processing and generates a second response based on the instructions returned by the user account, outputting it to the sender of the interaction message. In other words, when the clone instance lacks sufficient capabilities or permissions, the executing entity does not simply terminate the interaction. Instead, it promptly returns the decision-making power to the user account, ensuring that each interaction message receives a valid response. This avoids communication interruptions caused by directly refusing to answer, thus guaranteeing the continuity and integrity of business communication. Furthermore, since the final response content is generated based on the instructions returned by the user account, the digital clone instance only acts as an intermediary for information transmission and semantic translation throughout the entire process, without generating response content that exceeds its authorized scope. This ensures the security and reliability of processing results in complex or sensitive scenarios and avoids the risk of unclear responsibility arising from the clone instance's unauthorized operations.

[0050] In one embodiment, generating a second response based on an instruction returned by a user account includes: performing semantic parsing and reconstruction operations on the instruction returned by the user account, and extracting disclosable information content from the instruction based on the sender's identity and permission information of the interaction message to generate a second response.

[0051] In this embodiment, after receiving the instruction returned by the user account, the digital clone instance first performs semantic parsing and reconstruction operations on the instruction. Semantic parsing refers to the semantic analysis and understanding of the original instruction content returned by the user account, identifying the core intent, key information, constraints, and logical relationships contained therein. Since the instructions returned by the user account after receiving the push interaction message may take various forms, such as complete reply text, key processing opinions, brief instructions containing only principles or restrictions, or even a reference to an existing document or data, it is necessary to extract effective information units that can be used to generate a reply from the above different forms of original input through semantic parsing, and to sort out and reorganize the logical relationships between the information units. Reconstruction refers to reorganizing the information units according to the preset reply structure and expression logic based on the results of semantic parsing, transforming them from the original input form into a structured content form suitable for external output, thereby providing a standardized content foundation for subsequent information extraction and reply generation.

[0052] After completing semantic parsing and reconstruction, the digital avatar instance obtains the sender's identity and permission information from the interactive message. Based on this information, it extracts disclosable information from the reconstructed content and generates a second response. The identity identifier uniquely identifies the sender within the organizational structure, while the permission information characterizes the sender's role and accessibility. Since user account-returned instructions may contain information applicable only to specific levels or scopes of recipients, not all information is suitable for disclosure to all senders. Therefore, the reconstructed content needs to be filtered and adapted based on the sender's identity and permission information. Only information matching the sender's permission level is extracted, while information exceeding the sender's permissions is hidden, thus achieving differentiated information control. For example, user account-returned instructions might contain internal progress information accessible only to members of the same project team. For cross-departmental or external senders, the digital avatar instance only extracts the conclusive summary from the instructions to generate a response. For example, the instructions returned by a user account may include multiple alternative solutions. The digital clone instance decides whether to disclose all solutions or only the final adopted solution in the reply, depending on the sender's permission level. Finally, based on the extracted disclosable information, a final reply corresponding to the interaction message is generated and output as a second response to the sender.

[0053] Through the above methods, the digital clone instance, in the process of transforming user account instructions into the final response, not only completes the semantic understanding and structural reorganization of the original instructions, but also performs adaptive filtering of information content based on the recipient's identity and permissions. This mechanism ensures the accurate transmission of the user account's original intent in the response content while avoiding the risk of unauthorized information disclosure due to standardized output, achieving precise output of response content while guaranteeing information security.

[0054] The technical solution provided in this application creates a digital clone instance corresponding to a user account as a proxy object for that user account in the digital environment. This allows the digital clone instance to assume the proxy function of business interaction when the user account cannot intervene in real time, thereby ensuring business continuity and communication accessibility for the user account when it is off-duty or in a low-response state, and avoiding business collaboration interruption or information blockage caused by the user's inability to respond in a timely manner. Based on this, this solution sets an authorization scope, allowing the digital clone instance to autonomously generate a first response only within its authorized limits. This achieves both immediate automatic response to routine business inquiries and constrains the autonomous behavior of the digital clone instance within controllable boundaries, preventing information leakage or unauthorized commitment risks due to excessive automation. Furthermore, when the interaction message exceeds the authorization scope of the digital clone instance, this solution does not adopt the direct rejection strategy commonly used in existing AI systems. Instead, it pushes the interaction message to the user account, which then intervenes to process it and generates a second response based on the instructions returned by the user account. The aforementioned mechanism, on the one hand, ensures timely responses and efficient business processing within the authorized scope of autonomy, preventing user accounts from being excessively disturbed by high-frequency repetitive queries; on the other hand, when exceeding the authorized scope, it pushes information to the user account and generates a response based on the user's instructions, ensuring the accuracy and security of processing results in complex or sensitive scenarios, and ensuring that every interactive message receives an effective response. Therefore, this solution improves business communication efficiency while clarifying the responsibility boundary between proxy behavior and user behavior through permission constraints on digital avatar instances and the final decision-making intervention of user accounts, providing a reliable technical foundation for the controllable deployment of the avatar proxy mechanism in the organizational environment.

[0055] In one embodiment, user accounts are set with different authorization scopes and response strategies for different interactive message senders or different business scenarios.

[0056] Specifically, when senders of interactive messages engage in business interactions with user accounts, different senders may have different identity attributes, organizational roles, or degrees of business affiliation, and their access needs and trust levels for user accounts may differ. For example, members of the same project team and colleagues from different departments may have different needs regarding the scope of information and level of business support that a user account can access. To accommodate these differentiated scenarios, user accounts can use the configuration interface provided by the entity executing this application to set corresponding authorization scopes and response strategies for different senders or different business scenarios.

[0057] The scope of authorization includes, but is not limited to: the range of data that the digital clone instance can access for a specific sender or specific business scenario, the types of business that can be answered on behalf of the user, and the limits of commitments that can be made. The response strategy includes, but is not limited to: the response priority of the digital clone instance for a specific sender or specific business scenario, the response time requirements, whether to enable autonomous answering or only record information, and the handling method when the scope is exceeded (such as pushing to the user account or directly rejecting the request).

[0058] For example, for members within the same project team as the user account, the user account can set a broader authorization scope, enabling the digital avatar instance to provide more detailed business data and progress synchronization information. For senders from different departments or external parties, the user account can set a narrower authorization scope, allowing the digital avatar instance to provide only summary or guiding responses, and automatically triggering push notifications to the user account when sensitive information is involved. Furthermore, for business scenarios involving project progress synchronization, the user account can configure the digital avatar instance to automatically provide standardized progress reports; for business scenarios involving resource applications or approvals, the user account can configure the digital avatar instance to only record and forward information, without making direct commitments, and push relevant interaction messages to the user account for processing.

[0059] Through this embodiment, the digital clone instance can flexibly adjust its proxy behavior according to the differences in the sender's identity and the type of business scenario under a unified authorization framework. This not only meets the differentiated needs in business collaboration, but also avoids the problems of excessive information disclosure or rigid response strategies caused by a unified authorization strategy, further improving the precision and applicability of the clone proxy.

[0060] exist Figure 1 Based on the process shown, this application also provides the following embodiments: In one embodiment, the technical solution of this application further includes: configuring at least one of the following for the digital clone instance: an activation period, a disabling period, or an automatic activation rule, based on the activation conditions set by the user account.

[0061] In this embodiment, user accounts can flexibly set the effective time of the digital clone instance through the configuration interface provided by the executing entity of this application, according to their own work rhythm and business needs.

[0062] The "Activation Period" refers to the time interval during which the digital avatar instance is active and can perform proxy operations, such as 9:00 AM to 5:00 PM on weekdays. The "Disabled Period" refers to the time interval during which the digital avatar instance is inactive. During the disabled period, the digital avatar instance does not respond to any interactive messages; these messages can be temporarily stored by the system or processed according to other preset methods. The "Automatic Activation Rule" refers to the conditions that trigger the digital avatar instance to automatically enter an active state. For example, when it detects that a user account is on leave, in a meeting, or working across time zones, the digital avatar instance automatically activates the proxy function without requiring manual operation from the user account.

[0063] Through this embodiment, user accounts can flexibly control the working rhythm of digital clone instances according to changes in their own responsiveness. This avoids unnecessary interference with the user's normal work during the user's on-duty hours, while ensuring the automation and continuity of business communication when the user is off-duty or during low-responsiveness periods, thus improving the applicability of the digital clone proxy mechanism and the user experience.

[0064] In one embodiment, the technical solution of this application further includes: when generating a first response or a second response, adding source identification information, wherein the source identification information is used to indicate that the response content is generated by the digital clone instance or generated with the confirmation of the user account.

[0065] In this embodiment, when the digital avatar instance outputs a first or second response to the sender of the interactive message, it adds source identification information to the output response content so that the sender can clearly know how the response content was generated. The source identification information can be presented in a preset location within the response content as a label, footnote, prompt, or other visual form. When the response content is independently generated by the digital avatar instance within its authorized scope, the source identification information indicates that the content was generated by the digital avatar instance, for example, marked "Reply by digital avatar"; when the response content is generated based on an instruction returned by the user account, the source identification information indicates that the content was generated after confirmation by the user account, for example, marked "Sent by digital avatar after confirmation by user account".

[0066] Through this embodiment, the sender of the interactive message can clearly identify the source attribute of the reply content, avoiding information misjudgment or lack of trust caused by the inability to identify the identity of the replyer. At the same time, it provides verifiable evidence for tracing the reply subject and attribution of responsibility afterward, enhancing the transparency and credibility of the digital avatar proxy mechanism in organizational communication.

[0067] In one embodiment, the technical solution of this application further includes: in response to a takeover command issued by a user account, the digital clone instance suspends or terminates the currently ongoing response generation process and switches subsequent interactions to be handled by the user account. Further, after the user account has finished processing, the proxy state of the digital clone instance is restored, and it continues to respond to subsequent interaction messages.

[0068] In this embodiment, at any time during the operation of the digital clone instance, the user account can issue a takeover command to the digital clone instance through a preset takeover operation entry. The takeover operation entry includes, but is not limited to, the takeover button on the client interface, voice control commands, or preset shortcut keys. Upon receiving the takeover command, the digital clone instance immediately pauses or terminates the currently ongoing response generation process, ceases its autonomous processing of the current interactive message, and completely transfers the interactive message and its corresponding dialogue context to the user account. The dialogue context includes, but is not limited to, the content of the current interactive message, the sender's identity information, and the historical dialogue records of the interactive message. After the user account takes over, the current and subsequent interactive messages are directly processed by the user account, and the digital clone instance no longer intervenes in the response process of the current session.

[0069] Furthermore, after the user account takeover process is complete, a recovery command can be issued to the digital clone instance. The digital clone instance responds to the recovery command, unsuspends or terminates its state, restores its agent state, and continues to respond to subsequent interactive messages. The user account can restore the clone's agent state immediately after the takeover process, or it can restore it after batch processing is completed; the specific timing is determined by the user account.

[0070] This embodiment allows user accounts to proactively reclaim control of the interaction process whenever manual intervention is required, avoiding the inability to promptly correct deviations or errors that may occur during automated processing of digital avatar instances. Simultaneously, the introduction of a recovery mechanism enables the digital avatar instance to quickly return to normal operation without requiring reconfiguration or re-login after temporary intervention. This balances the necessity of human intervention with the continuity of automatic proxying, further enhancing the system's adaptability to complex business scenarios and user control over digital avatar behavior.

[0071] In one embodiment, the technical solution of this application further includes: the digital clone instance prioritizes the received interactive messages and processes the interactive messages in order of priority.

[0072] In practical applications, a digital avatar instance may receive multiple interactive messages from multiple senders simultaneously or within a short period of time during operation. To allocate processing resources reasonably and ensure that important or urgent interactive messages are responded to first, the digital avatar instance prioritizes each received interactive message and sorts them according to the evaluation results, processing each interactive message in descending order of priority.

[0073] For example, priority ranking can be determined based on multiple factors. For instance, a digital avatar instance can determine priority based on the sender's identity attributes, such as messages sent by project team members taking precedence over messages sent by non-team members; priority can also be determined based on keywords or semantic features carried in the messages, such as messages containing keywords like "urgent," "important," or "deadline" being given higher priority; furthermore, priority can be determined based on the business scenario or transaction type involved in the messages, such as messages involving approval processes or customer interactions taking precedence over general inquiry messages. These priority assessment factors can be used individually or in combination.

[0074] Through this embodiment, digital clone instances can achieve differentiated response time allocation when processing resources are limited or message load is high, avoiding the backlog or delay of important messages due to first-in-first-out linear processing, and improving the response efficiency and service quality of the clone proxy mechanism in complex business environments.

[0075] In one embodiment, the technical solution of this application further includes: the digital clone instance pushing a proxy work report to the user account, so that the user account can provide feedback to the digital clone instance based on the proxy work report; and optimizing the digital clone instance based on the feedback from the user account to the digital clone instance.

[0076] In this embodiment, the digital avatar instance, according to a preset time period or upon receiving instructions from the user account, summarizes various operational data generated during its proxy operation, generates a proxy work report, and pushes it to the user account. The proxy work report includes, but is not limited to: the number of interaction messages processed by the digital avatar instance within the reporting period, the distribution of various interaction messages, statistics on the types of high-frequency issues, a list of items pushed to the user account that exceeded the authorized scope, the response time of each processing, and a summary of the processing results.

[0077] After receiving the proxy work report, the user account can evaluate the performance of the digital clone instance based on the content recorded in the report, and provide feedback to the digital clone instance or the system. Feedback includes, but is not limited to: scoring the accuracy of the digital clone instance's response, correcting or supplementing specific response content, indicating adjustments to the handling strategies for high-frequency problem types, and proposing modifications to the scope of authorization.

[0078] The implementing entity of this application optimizes and adjusts the knowledge system, processing logic, or authorization configuration of the digital avatar instance based on feedback from user accounts. For example, when a user account points out a deviation in a response, the digital avatar instance uses this feedback as a basis for correction, updating the corresponding knowledge unit in its knowledge base or adjusting the corresponding processing logic; when a user account indicates that a certain type of question does not need to be pushed to the user account, the digital avatar instance adjusts its autonomous handling strategy for that type of question accordingly. Through the above mechanism, the response accuracy and processing strategy of the digital avatar instance can be continuously iterated and optimized as user feedback accumulates, gradually approaching the user account's true expectations and business needs.

[0079] In one embodiment, the technical solution of this application further includes: in response to detecting that a preset security constraint condition is met, the digital clone instance switches from the current working mode to a restricted response mode; wherein, the preset security constraint condition includes at least one of the following: the risk level of the interaction scenario exceeds a threshold, the completeness of the information that the digital clone instance can obtain is lower than a preset standard, and the boundaries of the operation permissions involved in the interaction message are unclear; in the restricted response mode, the output capability of the digital clone instance is restricted to only performing at least one of the following operations: outputting information guiding responses, outputting status feedback responses, and outputting rejection responses.

[0080] In this embodiment, the digital clone instance continuously monitors and evaluates the security status of the current interactive environment during operation. When it detects that the current interactive environment meets preset security constraints, the digital clone instance automatically triggers a mode switch, switching from normal response mode to restricted response mode, in order to reduce the risk of misjudgment due to insufficient information or complex scenarios.

[0081] The preset security constraints include, but are not limited to, at least one of the following situations: First, the risk level of the interaction scenario exceeds the threshold, that is, the digital clone instance identifies high-risk words (such as "compensation," "breach of contract," "commitment," etc.) or high-risk business areas (such as financial approval, legal compliance, etc.) in the current interaction message through semantic analysis, and determines that the risk level of the current interaction scenario exceeds the preset security threshold based on the risk identification results; Second, the completeness of the information that the digital clone instance can obtain is lower than the preset standard, that is, the issue involved in the current interaction message exceeds the coverage of the clone instance's existing knowledge system, or the relevant information that the clone instance can retrieve is insufficient to support the generation of an accurate response; Third, the boundaries of the operation permissions involved in the interaction message are unclear, that is, the clone instance cannot clearly determine whether the operation involved in the interaction message falls within its authorized scope, or the interaction message involves a complex operation request that crosses permission boundaries.

[0082] In restricted response mode, the output capabilities of the digital avatar instance are limited, and it can only perform at least one of the following operations: output guiding responses, such as providing the sender with guidance on obtaining relevant information or suggesting subsequent operation paths; output status feedback responses, such as informing the sender of the current status of the avatar instance (e.g., "Insufficient information at present, unable to reply completely at this time"); output rejection responses, such as explicitly informing the sender that the issue is beyond the processing capabilities of the avatar instance and suggesting contacting the user account. In restricted response mode, the digital avatar instance does not generate complete responses involving substantive business content or decision-making judgments.

[0083] Through this embodiment, the digital clone instance can proactively adopt a conservative strategy when facing high-risk, incomplete information, or ambiguous permission interaction scenarios, and output responses in a restricted manner, avoiding inaccurate or inappropriate responses under insufficient conditions, and further ensuring the security and reliability of the clone proxy mechanism in complex scenarios.

[0084] In one embodiment, the technical solution of this application further includes: constructing a digital personality model of the user account based on the user account's private domain data, the digital personality model including language style features, business processing logic features and decision-making mode features; when the digital clone instance generates the first response, the response content is style-adapted based on the digital personality model.

[0085] Specifically, a digital personality model is a set of features obtained by digitally abstracting and modeling the personalized characteristics of a user account in business communication. Private domain data refers to personalized data resources accumulated by a user account in its daily business activities, including but not limited to the user account's historical communication records (such as instant messaging messages and email exchanges), work documents (such as reports, plans, and minutes), and business processing records (such as approval opinions and task notes). This private domain data embodies the unique expression habits, thinking patterns, and work styles of a user account, and forms the data foundation for constructing a digital personality model.

[0086] By performing feature extraction operations on the private domain data of user accounts, personalized characteristics of user accounts are mined to generate a digital personality model. The digital personality model includes, but is not limited to, the following feature types: Language style features, which characterize the user account's habitual wording preferences, sentence structures, tone tendencies, and punctuation usage habits in written or oral expression, such as a user's tendency to use formal and rigorous expressions or a relatively relaxed communication style; Business processing logic features, which characterize the user account's habitual analytical frameworks, processing procedures, and judgment criteria when handling various business issues, such as a user's habit of summarizing the overall progress before detailing specific aspects when responding to project progress inquiries; Decision-making pattern features, which characterize the user account's habitual decision-making tendencies and trade-off standards when facing business scenarios requiring decision-making, such as a user's tendency towards conservative response strategies or a response strategy that provides multiple options for the other party to choose from. It should be noted that the above feature types are only illustrative examples, and the actual feature dimensions of the digital personality model can be adaptively expanded according to the data type and richness of the user account's private domain data.

[0087] During the generation of the first response by the digital avatar instance, after completing content analysis and knowledge retrieval of the interactive messages, the digital avatar instance performs style adaptation processing on the initially generated response content based on the digital personality model. Style adaptation processing refers to adjusting and optimizing the wording, sentence structure, information organization, and expression style of the response content according to the language style characteristics, business processing logic characteristics, and decision-making mode characteristics of the user account recorded in the digital personality model, so that the final output of the first response is consistent with the user account's real communication style in terms of expression and logical habits.

[0088] Through this embodiment, the responses generated by the digital avatar instance are not only accurate and relevant in terms of information content, but also highly consistent with the user account in terms of expression style and logical habits. Therefore, when the sender of the interactive message receives the first response, they can feel a similar experience to communicating with the actual user account, reducing the awkwardness and disconnect of automated replies and improving the consistency and naturalness of the interactive experience.

[0089] In one embodiment, the technical solution of this application further includes: in response to a received lifecycle management instruction, performing a state change operation on the digital clone instance; wherein the state change operation includes at least one of the following: enabling the digital clone instance, freezing the digital clone instance, and destroying the digital clone instance.

[0090] Specifically, lifecycle management commands are operation instructions issued by the system or user accounts to control the running status of digital clone instances. Lifecycle management commands can be generated under various triggering conditions: for example, based on preset project cycle conditions, i.e., when a project in which the user account participates starts or ends, a corresponding enable or destroy command is automatically generated; based on time conditions, i.e., when the usage period of the digital clone instance expires, a freeze or destroy command is automatically generated; based on organizational status changes, i.e., when the user account undergoes organizational restructuring, job changes, or resignation, a corresponding status change command is generated based on the change in organizational status. Lifecycle management commands can also be initiated manually by the user account.

[0091] Upon receiving a lifecycle management instruction, the digital clone instance executes the corresponding state change operation based on the instruction type. State change operations include at least one of the following: Activating the digital clone instance, which activates a clone instance that has been created but is currently inactive, enabling it to acquire proxy capabilities; Freezing the digital clone instance, which temporarily suspends an active clone instance. During the freeze period, the clone instance does not respond to any interactive messages but retains its configuration information and historical data for later unfreezing and resumption of operation; Destroying the digital clone instance, which completely deletes the digital clone instance and its associated configuration data, learning data, and historical records, releasing the corresponding system resources.

[0092] Through this embodiment, the running status of the digital clone instance can be dynamically adjusted according to the project cycle, organizational changes, or user needs. This avoids the data accumulation and security risks caused by the long-term idleness of the clone instance when it is no longer needed. It also ensures that the digital clone instance can be cleaned up in a timely manner when organizational members change or positions are adjusted, preventing information security risks caused by historical configurations and permission residues. This provides a reliable technical guarantee for the standardized management of the digital clone proxy mechanism.

[0093] In one embodiment, the technical solution of this application further includes: constructing and maintaining a memory for the digital clone entity, wherein the memory is used to persistently store memory data that can be accessed across multiple interactive sessions.

[0094] In the technical solution provided in the embodiments of this application, the digital clone instance is different from the general artificial intelligence assistant. It has a hierarchical and manageable memory, which can not only take into account the short-term context, but also remember and apply user preferences for a long time, thereby providing users with personalized and continuous intelligent services.

[0095] Specifically, the memory of a digital avatar instance is a specially constructed data storage and management system. Its core function is to persistently store memory data, enabling memories to persist and be retrieved across multiple independent conversations with the user. This design addresses the technical shortcomings of traditional intelligent assistants, which can only maintain context within a single conversation and cannot accumulate long-term user preferences. It empowers digital employees to "remember" user habits, historical decisions, and key information, thereby providing a continuous and intelligent interactive experience. Furthermore, different digital avatar instances can have independent memory. This means that each digital avatar instance can build its own exclusive memory data system based on its service targets, functional positioning, and interaction history. On the one hand, independent memory allows digital avatar instances to provide highly personalized memory support for the specific users or teams they serve, avoiding confusion or interference between memory data across different digital avatar instances.

[0096] In one embodiment, constructing and maintaining the memory of a digital clone instance includes: acquiring information from at least one source to generate memory data, and storing the generated memory data in the memory of the digital clone instance; these sources include, but are not limited to, interaction events involving the digital clone instance, synchronization data from external systems, or data imported from files. The technical solution provided by this embodiment, through a multi-source information acquisition mechanism, can effectively ensure the richness and comprehensiveness of the memory data of the digital clone instance.

[0097] Interaction events involving digital avatar instances refer to various interaction processes between digital avatar instances and users or other digital avatar instances. When specific triggering conditions are met, digital avatar instances can extract key information from the interaction and generate memory data. For example, based on different triggering scenarios, the triggering conditions can be summarized into the following four types: First, key factual information change triggering, memory generation is triggered when changes are detected in key factual information explicitly stated by the user (such as adjustments to time nodes, modifications to budget amounts, updates to project goals, changes in responsible persons, etc.); Second, personal preference expression triggering, memory generation is triggered when users use preference keywords such as "I like," "I'm used to," "Please do not" to express personal preferences in the interaction; Third, important decision and conclusion triggering, memory generation is triggered when a resolution is clearly reached in the dialogue (such as "decided," "confirmed," "passed," etc.); Fourth, recurring problem solution triggering, memory generation is triggered when the same problem is repeatedly raised (semantic similarity exceeds 80% and occurs twice within 30 days), and the system records the habit preference accordingly.

[0098] Synchronizing data from external systems refers to the automatic synchronization of structured data between the digital avatar instance and other business systems via application programming interfaces (APIs) or data buses. For example, synchronizing user information such as job title, department, and supervisor from the human resources system updates the digital avatar instance's understanding of the user's identity; synchronizing the company's meeting room resource list and real-time occupancy status from the meeting room booking system enables the digital avatar instance to accurately process relevant requests without requiring users to repeatedly state their preferences. Importing data from files refers to the digital avatar instance extracting, analyzing, and structuring content from historical documents uploaded by users or stored in specified locations, thereby achieving batch initialization of memory data. For example, importing and parsing users' historical email archives extracts key information about project discussions, decision-making conclusions, and to-do commitments to form the user's "work history" memory; batch parsing past meeting minutes extracts meeting resolutions, action items, and responsible persons as a source of memory for project context and team habits.

[0099] Regarding the memory storage structure, the memory adopts a hierarchical storage system, including at least a first logical layer and a second logical layer. The first logical layer is used to store raw memory data directly derived from user interaction; the second logical layer is used to store memory data derived from the processing of interaction information by the digital avatar instance. Further, the second logical layer includes: a sublayer for storing memory data across different application scenarios; and / or, a sublayer for storing memory data related to a specific application scenario.

[0100] The first logical layer stores raw memory data directly derived from user interactions. This type of data is characterized by its "rawness" and "directness," meaning it includes information directly input by the user through active completion, explicit notification, or preference settings, or factual statements directly expressed by the user during interactions that have not undergone deep processing by the digital avatar instance. Examples include job information and title preferences filled in by the user in the personalization settings interface, or explicitly stating in a conversation, "I usually have meetings at 10 AM." The data in the first logical layer can be stored using a structured database to ensure accuracy and manageability.

[0101] The second logical layer stores memory data derived from the processing of interactive information by the digital avatar instances. This "interactive information" is broad, encompassing direct interactive flow content, raw memory data already stored in the first logical layer, and other relevant contextual information, as well as the accumulated content from multiple rounds of original interactive conversations. The digital avatar instances use intelligent processing methods such as induction, summarization, reasoning, and association to extract higher-level patterns and rules from this interactive information, forming derived memory data. For example, by analyzing multiple discussions about the same project in different scenarios (one-on-one chat, group chat, document comments), the digital avatar instances automatically summarize the project's key progress and core participants, forming general implicit memory; or by observing users' operational habits in office approval scenarios, they summarize the user's focus on approval matters, forming scenario-specific memory. The memory data in the second logical layer can be stored using a vector database for subsequent semantic-based intelligent retrieval.

[0102] Furthermore, the second logical layer can be subdivided into different sub-layers to meet diverse memory needs. The sub-layer storing memory data across different application scenarios primarily records general implicit memories such as users' work habits, communication preferences, and project history information. These memories are not limited by specific tools or scenarios and can be reused in various interaction contexts. The sub-layer storing memory data related to specific application scenarios focuses on preferences under specific tools or scenarios, such as meeting location preferences, document format preferences, and approval process habits. These memories are closely bound to specific application scenarios and are accurately retrieved when the corresponding scenario intent is recognized.

[0103] In one embodiment, the memory may further include a third logical layer that accesses a specific knowledge base.

[0104] A specific knowledge base refers to a centralized, structured database or system specifically built for storing and retrieving knowledge, such as a corporate wiki, product manual repository, or FAQ repository. The introduction of this logical layer enables digital avatar instances to access enterprise-level general knowledge resources, gaining more comprehensive information support when answering specialized questions or performing complex tasks.

[0105] Through the above layered design, the memory of the digital clone instance achieves refined management and differentiated storage of memory data of different types, sources, and purposes, laying a solid foundation for subsequent intention-based precision memory applications.

[0106] Based on a multi-source acquisition and memory hierarchical system, the memory of digital clone instances has the ability to be dynamically updated to ensure that the stored memory data is always accurate, relevant, and effective. This corresponds to the fact that in practical applications, user preferences may shift, business information may change, and historical accumulation may diminish in value over time. In addition, a quality scoring system is introduced to determine the quality score of the memory data in the memory (scoring criteria include, but are not limited to: direct user feedback on memory-related outputs, the frequency of successful memory data retrieval or the most recent retrieval time, and the degree of consistency between the memory data and its source information or current context, etc.). Memory data is managed based on the quality score, such as demoting low-scoring memories so that they are ranked lower or not retrieved in subsequent searches, and promoting high-scoring memories for priority application, thereby enabling the memory to have the ability to self-purify and optimize.

[0107] Through the aforementioned memory construction and maintenance mechanism, digital clone instances can reuse user preferences, historical decisions, and key information across sessions, demonstrating the ability to "remember" users during interactions. This avoids the problem of users repeatedly providing the same information in each session, allowing each interaction to benefit from historically accumulated knowledge and preferences, thereby providing a truly human-like, scalable, and intelligent service experience.

[0108] Figure 2 This is a block diagram illustrating an embodiment of a data processing apparatus provided in this application. (See diagram below.) Figure 2 As shown, the device includes: The clone creation module 21 is used to create a digital clone instance corresponding to a user account, wherein the digital clone instance serves as a proxy object of the user account in the digital environment; The autonomous response module 22 is used to respond to the interactive message, wherein the digital clone instance generates a first response within the authorized scope of the digital clone instance; The collaborative decision-making module 23 is used to push the interactive message to the user account when the interactive message exceeds the authorization scope of the digital clone instance, and generate a second response based on the instructions returned by the user account.

[0109] In one possible implementation, the device further includes: The source identification module is used to add source identification information when the first response or the second response is generated. The source identification information is used to indicate that the response content was generated by the digital clone instance or was generated with the confirmation of the user account.

[0110] In one possible implementation, the device further includes: The manual takeover module is used to respond to a takeover command issued by the user account. The digital clone instance suspends or terminates the currently ongoing response generation process and switches subsequent interactions to be handled by the user account.

[0111] In one possible implementation, the device further includes: The proxy recovery module is used to restore the proxy state of the digital clone instance after the user account has been processed, so that it can continue to respond to subsequent interaction messages.

[0112] In one possible implementation, the device further includes: The enable condition configuration module is used to configure at least one of the enable period, disable period, or automatic enable rule for the digital clone instance based on the enable conditions set by the user account.

[0113] In one possible implementation, the device further includes: The priority scheduling module is used by the digital clone instance to prioritize the received interaction messages and process the interaction messages in order of priority.

[0114] In one possible implementation, the device further includes: The reporting and feedback module is used for the digital clone instance to push the agent work report to the user account, so that the user account can provide feedback to the digital clone instance based on the agent work report; Based on the user account's feedback regarding the digital clone instance, the digital clone instance is optimized.

[0115] In one possible implementation, the device further includes: A security constraint module is used to switch the digital clone instance from the current working mode to a restricted response mode in response to the detection that a preset security constraint condition is met; wherein the preset security constraint condition includes at least one of the following: the risk level of the interaction scenario exceeds a threshold, the completeness of the information that the digital clone instance can obtain is lower than a preset standard, and the boundaries of the operation permissions involved in the interaction message are unclear; In the restricted response mode, the output capability of the digital clone instance is limited to performing only at least one of the following operations: outputting a guiding response, outputting a status feedback response, or outputting a rejection response.

[0116] In one possible implementation, the clone creation module 21 creates a digital clone instance uniquely corresponding to the user account, including: Obtain the user account's identity and permission information; The digital clone instance is created based on the identity identifier, and the authorization scope of the digital clone instance is configured according to the permission information.

[0117] In one possible implementation, the collaborative decision-making module 23 generates a second response based on the instruction returned by the user account, including: The instructions returned by the user account are subjected to semantic parsing and reconstruction operations. Based on the identity and permission information of the sender of the interaction message, the disclosable information content is extracted to generate a second response.

[0118] In one possible implementation, the digital clone instance in the autonomous response module 22 generates a first response within the authorized scope of the digital clone instance, including: The digital clone instance performs type identification on the interactive messages; Based on the identification results, at least one intelligent agent is invoked from among the multiple intelligent agents configured within the digital clone instance; wherein the invoked intelligent agent generates a sub-response within its corresponding task processing capability range; The digital clone instance integrates the sub-responses generated by each intelligent agent to generate the first response.

[0119] In one possible implementation, the device further includes: The personality modeling module is used to construct a digital personality model of the user account based on the user account's private domain data. The digital personality model includes language style features, business processing logic features, and decision-making pattern features. The style adaptation module is used to perform style adaptation processing on the response content based on the digital personality model when the digital clone instance generates the first response.

[0120] In one possible implementation, the device further includes: The management module is used to perform state change operations on the digital clone instance in response to received lifecycle management instructions; The state change operation includes at least one of the following: enabling the digital clone instance, freezing the digital clone instance, and destroying the digital clone instance.

[0121] In one possible implementation, the device further includes: The memory maintenance module is used to build and maintain the memory of the digital clone entity, which is used to persistently store memory data that can be accessed across multiple interactive sessions.

[0122] like Figure 3As shown in the figure, this application provides an electronic device, including a processor 111, a communication interface 112, a memory 113, and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114. Memory 113 is used to store computer programs; In one embodiment of this application, when the processor 111 executes the program stored in the memory 113, it implements the digital clone proxy method provided in any of the foregoing method embodiments, including: Create a digital clone instance corresponding to the user account, which serves as a proxy object for the user account in the digital environment; In response to the interaction message, the digital clone instance generates a first response within the authorized scope of the digital clone instance; In addition, if the interaction message exceeds the authorization scope of the digital clone instance, the interaction message is pushed to the user account, and a second response is generated based on the instruction returned by the user account.

[0123] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the digital clone proxy method provided in any of the foregoing method embodiments.

[0124] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0125] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0126] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0127] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A digital avatar proxy method, characterized in that, The method includes: Create a digital clone instance corresponding to the user account, which serves as a proxy object for the user account in the digital environment; In response to the interaction message, the digital clone instance generates a first response within the authorized scope of the digital clone instance; In addition, if the interaction message exceeds the authorization scope of the digital clone instance, the interaction message is pushed to the user account, and a second response is generated based on the instruction returned by the user account.

2. The method according to claim 1, characterized in that, The method further includes: In the case of generating the first response or the second response, source identification information is added, which is used to indicate that the response content was generated by the digital clone instance or was generated with the confirmation of the user account.

3. The method according to claim 1, characterized in that, The method further includes: In response to a takeover command issued by the user account, the digital clone instance suspends or terminates the currently ongoing response generation process and switches subsequent interactions to be handled by the user account.

4. The method according to claim 3, characterized in that, The method further includes: After the user account is processed, the proxy state of the digital clone instance is restored, and it continues to respond to subsequent interaction messages.

5. The method according to claim 1, characterized in that, The method further includes: Based on the activation conditions set by the user account, configure at least one of the following for the digital clone instance: an activation period, a disabling period, or an automatic activation rule.

6. The method according to claim 1, further comprising: The digital clone instance prioritizes the received interaction messages and processes them sequentially according to their priority.

7. The method according to claim 1, characterized in that, The method further includes: The digital clone instance pushes a proxy work report to the user account, so that the user account can provide feedback to the digital clone instance based on the proxy work report; Based on the user account's feedback regarding the digital clone instance, the digital clone instance is optimized.

8. The method according to claim 1, characterized in that, The method further includes: In response to the detection that preset security constraints are met, the digital clone instance switches from the current working mode to a restricted response mode; wherein, the preset security constraints include at least one of the following: the risk level of the interaction scenario exceeds a threshold, the completeness of the information that the digital clone instance can obtain is lower than a preset standard, and the boundaries of the operation permissions involved in the interaction message are unclear; In the restricted response mode, the output capability of the digital clone instance is limited to performing only at least one of the following operations: outputting a guiding response, outputting a status feedback response, or outputting a rejection response.

9. The method according to claim 1, characterized in that, The creation of a digital clone instance uniquely corresponding to a user account includes: Obtain the user account's identity and permission information; The digital clone instance is created based on the identity identifier, and the authorization scope of the digital clone instance is configured according to the permission information.

10. The method according to claim 1, characterized in that, The generation of the second response based on the instruction returned by the user account includes: The instructions returned by the user account are subjected to semantic parsing and reconstruction operations. Based on the identity and permission information of the sender of the interaction message, the disclosable information content is extracted to generate a second response.

11. The method according to claim 1, characterized in that, The digital clone instance generates a first response within the authorized scope of the digital clone instance, including: The digital clone instance performs type identification on the interactive messages; Based on the identification results, at least one intelligent agent is invoked from among the multiple intelligent agents configured within the digital clone instance; wherein the invoked intelligent agent generates a sub-response within its corresponding task processing capability range; The digital clone instance integrates the sub-responses generated by each intelligent agent to generate the first response.

12. The method according to claim 1, characterized in that, The method further includes: A digital personality model of the user account is constructed based on the user account's private domain data. The digital personality model includes language style features, business processing logic features, and decision-making pattern features. When the digital clone instance generates the first response, it performs style adaptation processing on the response content based on the digital personality model.

13. The method according to claim 1, characterized in that, The method further includes: In response to the received lifecycle management command, a state change operation is performed on the digital clone instance; The state change operation includes at least one of the following: enabling the digital clone instance, freezing the digital clone instance, and destroying the digital clone instance.

14. The method according to claim 1, characterized in that, The method further includes: The memory of the digital clone entity is constructed and maintained, and the memory is used to persistently store memory data that can be accessed across multiple interactive sessions.

15. A digital clone agent device, characterized in that, The device includes: The clone creation module is used to create digital clone instances corresponding to user accounts, and the digital clone instances serve as proxy objects for the user accounts in the digital environment; An autonomous response module is used to respond to interactive messages, wherein the digital clone instance generates a first response within the authorized scope of the digital clone instance; The collaborative decision-making module is used to push the interaction message to the user account when the interaction message exceeds the authorization scope of the digital clone instance, and generate a second response based on the instructions returned by the user account.

16. An electronic device, characterized in that, include: A processor and a memory, the processor being configured to execute a digital clone agent program stored in the memory to implement the digital clone agent method according to any one of claims 1-14.

17. A storage medium, characterized in that, The storage medium stores one or more programs, which can be executed by one or more processors to implement the digital clone agent method according to any one of claims 1-14.

Images