Information processing method and apparatus, medium, and electronic device

Abstract

An information processing method, device, medium and electronic equipment, relating to the technical field of artificial intelligence, the method comprising: receiving feedback information of a user; based on the feedback information, obtaining associated information associated with the feedback information; wherein the associated information at least includes operation information of the user and first result information, the first result information being generated based on the operation information; obtaining second result information, the second result information being generated based on at least part of the associated information, and displaying the second result information. Through the feedback information and the related management information, the processing result corresponding to the feedback content can be automatically and efficiently obtained.

Description

Technical Field

[0001] The technical solution relates to the field of artificial intelligence technology, specifically to an information processing method, device, medium, and electronic device. Background Technology

[0002] With the widespread adoption of the internet and the improvement of people's living standards, intelligent customer service has become an important bridge for communication between businesses and users. Intelligent customer service is a system that provides standardized answers to user feedback, capable of automatically classifying feedback, recognizing intent, matching business knowledge, and providing functional solutions. However, intelligent customer service cannot address the specific problems raised in user feedback. Summary of the Invention

[0003] This summary section is provided to briefly introduce the concepts, which will be described in detail in the detailed description section below. This summary section is not intended to identify key or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.

[0004] Firstly, an information processing method is provided, the method comprising: Receive user feedback information; Based on the feedback information, associated information related to the feedback information is obtained; wherein, the associated information includes at least the user's operation information and first result information, and the first result information is generated based on the operation information; Obtain second result information, which is generated based at least on a portion of the aforementioned associated information. Display the second result information.

[0005] Secondly, an information processing apparatus is provided, the apparatus comprising: The receiving module is configured to receive user feedback information; The first acquisition module is configured to acquire associated information related to the feedback information based on the feedback information; wherein the associated information includes at least the user's operation information and first result information, and the first result information is generated based on the operation information; The second acquisition module is configured to acquire second result information, which is generated based at least on a portion of the aforementioned associated information. The display module is configured to display the second result information.

[0006] Thirdly, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processing device, implements the steps of the method described in the first aspect.

[0007] Fourthly, an electronic device is provided, comprising: A storage device on which computer programs are stored; A processing device for executing the computer program in the storage device to implement the steps of the method described in the first aspect.

[0008] Through the above technical solution, upon receiving feedback information input by the user, related information can be obtained based on the feedback information. This related information includes at least the user's operation information and a first result information. The first result information can be generated based on the operation information. On this basis, a second result information can be generated based on at least some of the related information. That is, upon receiving feedback information input by the user, corresponding result information can be generated specifically for the feedback information and displayed. This not only enables automatic information processing but also allows for flexible and targeted solutions to user needs.

[0009] Other features and advantages of the technical solution will be described in detail in the following detailed implementation section. Attached Figure Description

[0010] The above and other features, advantages, and aspects of the various embodiments of the technical solutions will become more apparent when combined with the accompanying drawings and the following detailed description. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic, and the originals and elements are not necessarily drawn to scale. In the drawings: Figure 1 This is an example diagram illustrating information processing achieved manually.

[0011] Figure 2 It is a flowchart of an information processing method.

[0012] Figure 3 This is an example diagram of system prompts in an information processing method.

[0013] Figure 4 This is an example diagram of a tool registered by an intelligent system through a model context protocol in an information processing method.

[0014] Figure 5 This is an example diagram illustrating the configuration and initialization operations of an intelligent system in an information processing method.

[0015] Figure 6 This is an example diagram illustrating the asynchronous execution of an intelligent system in an information processing method.

[0016] Figure 7 This is an example diagram illustrating the deployment of a model context protocol service in an information processing method.

[0017] Figure 8This is an example diagram illustrating the isolation between model context protocol services and tool tenants in an information processing method.

[0018] Figure 9 This is a sequence diagram of the registration of the model context protocol server and tools in an information processing method.

[0019] Figure 10 This is a specific feedback investigation example diagram in an information processing method where an intelligent system receives a user feedback request.

[0020] Figure 11 This is an example diagram of the execution flow of an intelligent system in an information processing method.

[0021] Figure 12 This is an example diagram illustrating the interaction between an intelligent system, a model context protocol server, and tools in an information processing method.

[0022] Figure 13 This is an example diagram illustrating the process of an intelligent system calling a troubleshooting tool in an information processing method.

[0023] Figure 14 This is an example diagram illustrating the process of an intelligent system calling a troubleshooting tool in an information processing method.

[0024] Figure 15 This is an example diagram of the user interface displayed for the first user in an information processing method.

[0025] Figure 16 This is an example diagram of the interface display for a second user in an information processing method.

[0026] Figure 17 This is an example diagram illustrating user A's input feedback information in an information processing method.

[0027] Figure 18 This is an example diagram of the first screening strategy in an information processing method.

[0028] Figure 19 This is an example diagram illustrating user B's input feedback information in an information processing method.

[0029] Figure 20 An example diagram of the second screening strategy in an information processing method. Figure 21 This is an example diagram illustrating user C's input feedback information in an information processing method.

[0030] Figure 22 This is an example diagram of the third screening strategy in an information processing method.

[0031] Figure 23 It is a block diagram based on an information processing device.

[0032] Figure 24 This is a schematic diagram of the structure of an electronic device. Detailed Implementation

[0033] The technical solution will now be described in more detail with reference to the accompanying drawings. Although certain scenarios are shown in the drawings, it should be understood that the technical solution can be implemented in various forms and should not be construed as limited to the scenarios described herein. Rather, these scenarios are provided to provide a more thorough and complete understanding of the technical solution. It should be understood that the accompanying drawings and the scenarios described are for illustrative purposes only and are not intended to limit the scope of protection of the technical solution.

[0034] It should be understood that the steps described in the method implementation may be performed in different orders and / or in parallel. Furthermore, the method implementation may include additional steps and / or omit the steps shown. The scope of the technical solution is not limited in this respect.

[0035] The term "comprising" and its variations as used herein can be open-ended, meaning "including but not limited to". The term "based on" can mean "at least partially based on". The term "one case" means "at least one case"; the term "another case" means "at least one additional case"; the term "some cases" means "at least some cases". Definitions of other terms will be given in the following description.

[0036] It should be noted that the concepts of "first" and "second" mentioned here are only used to distinguish different devices, modules or units, and are not used to limit the order of the functions performed by these devices, modules or units or their interdependencies.

[0037] It should be noted that the terms "one" and "more" used here are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0038] The names of messages or information exchanged between the multiple devices in the implementation are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0039] It is understandable that before using the technical solutions provided here, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in accordance with relevant laws and regulations, and their permission should be obtained.

[0040] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose, based on the prompt message, whether to provide personal information to the software or hardware such as electronic devices, applications, servers, or storage media performing the operations described herein.

[0041] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0042] It is understood that the above notification and user permission acquisition process are merely illustrative and do not constitute a limitation on the implementation of the technical solution. Other methods that comply with relevant laws and regulations may also be applied to the implementation of the technical solution.

[0043] At the same time, it is understood that the data involved in the technical solution (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of relevant laws, regulations and related provisions.

[0044] In some implementations, intelligent customer service primarily relies on large language models and knowledge bases to semantically understand user input and provide basic consultation suggestions or Q&A services. Currently, a large proportion of user feedback consists of suspected defects, which are mainly identified through manual investigation by the business development team, meaning they cannot be directly resolved through dialogue with the large model. However, this manual investigation not only consumes significant manpower and time from the development team but also results in lengthy feedback processing times, leading to a poor user experience.

[0045] Please see Figure 1 When processing information, the troubleshooting work is mainly carried out by developers using various troubleshooting tools, which leads to an inability to accurately understand ambiguous or vague defect feedback. Defect feedback mainly consists of complex and personalized user feedback, which cannot be covered by general standardized processing procedures. For example, if the user's feedback does not mention keywords such as "refund," it is impossible to determine the user's intention after encountering a problem.

[0046] In addition, based on Figure 1It is evident that troubleshooting defect-related issues primarily relies on manual processes. However, manual investigation introduces additional processing time constraints. For instance, prolonged investigation times, or the possibility of overlooking issues due to limited development team resources, can negatively impact user experience. Furthermore, the intelligent customer service system uses standard template responses, resulting in inflexible content. Moreover, in cases of intent recognition errors, these templated responses can significantly increase user dissatisfaction.

[0047] To address the aforementioned issues, an information processing method, apparatus, medium, and electronic device are proposed. When receiving feedback information from a user, this information processing method can generate a second result information by utilizing relevant related information from the feedback information, thereby accurately and efficiently generating the processing result corresponding to the feedback information.

[0048] The technical solution will be further explained and illustrated below with reference to the accompanying drawings.

[0049] Figure 2 This is a flowchart of an information processing method. This information processing method can be applied to electronic devices with processing capabilities, such as terminals or servers. Furthermore, this information processing method can be executed by an information processing device, which can be implemented by software and / or hardware, and the software and / or hardware can be configured within the electronic device. (Refer to...) Figure 2 The information processing method may include the following steps.

[0050] In step S210, feedback information from the user is received.

[0051] In some implementations, user feedback can be a targeted feedback question entered by the user when encountering a practical problem while using the intelligent system. For example, the feedback can be entered by the user through a designated application, or it can be triggered by an agent through a designated application upon receiving user feedback.

[0052] In step S220, based on the feedback information, the associated information related to the feedback information is obtained.

[0053] As an alternative approach, upon receiving user input feedback, associated information can be obtained based on that feedback. Here, the associated information may include at least the user's operation information and first result information, wherein the first result information can be generated based on the operation information.

[0054] In step S230, the second result information is obtained.

[0055] In one scenario, the second result information can be generated based on at least some of the related information; that is, the second result information can be obtained by generating at least some of the related information.

[0056] For example, in response to the first result information belonging to a first category of information, this technical solution can obtain second result information, wherein the second result information can be generated based on the user's operation information and the first result information. Here, the first category of information can be a situation where a regeneration operation can be performed after the first result information is generated. For example, when the generated first result information does not meet the user's needs, or when the generation of the first result information fails, the second result information can be obtained by performing a retry operation.

[0057] Optionally, when the feedback information is related to content generation and the generated result does not meet the user's expectations, in response to the first result information belonging to the first category of information, the second result information is obtained.

[0058] As an optional approach, the first result information may include first multimedia information, and the user's operation information may include input information. Based on this, in response to the first multimedia information belonging to a first category of information, the first multimedia information is adjusted according to a first strategy and the input information to obtain second multimedia information. Here, the input information can be used to generate the first result information, and it may be a prompt message used to generate the first result information.

[0059] Optionally, the multimedia information may include images, videos, or audio. Furthermore, the first strategy can be implemented using a multimedia adjustment model, which may include an image adjustment model, a video adjustment model, or a dialogue adjustment model.

[0060] As an example, if the first image represented by the feedback information does not meet the user's needs, the first image is adjusted based on the first input information and the first strategy to obtain a second image. The first input information can be a prompt message used to generate the target image. Here, the user's needs can be requirements for the image content or the image format, etc. In other words, when the feedback information indicates dissatisfaction with the first image generated at the first moment, the first image can be updated.

[0061] Here, the update of the first image can be achieved through an image adjustment model. This involves inputting the first input information into the image adjustment model, and then using the model to update the first image based on the first input information—essentially performing an image regeneration operation to obtain the second image. For example, if the user's feedback is "I want the generated image to be in color, but the actual generated image is black and white," the image adjustment model can colorize the black and white first image to obtain the colorized second image.

[0062] As another example, if the first video represented by the feedback information does not meet the user's requirements, the first video is adjusted based on second input information and a second strategy to obtain a second video. The second input information can be a prompt used to generate the target video. Here, the user's requirements could be their demands for video content or video format, etc. In other words, when the feedback information indicates dissatisfaction with the first video generated at the first moment, the first video can be updated.

[0063] Here, the update of the first video can be achieved through a video adjustment model. This involves inputting the second input information into the video adjustment model, and then using the model in conjunction with the second input information to update the first video—essentially performing a video regeneration operation to obtain the second video. For example, if the user's feedback is "The generated video is not clear enough, please regenerate," the video adjustment model can generate a second image with higher clarity.

[0064] As an alternative approach, the first result information may include first dialogue information, and the user's operation information may include input information. Based on this, in response to the first dialogue information belonging to a first category of information, the first dialogue information is adjusted based on a second strategy and the input information to obtain second dialogue information. Furthermore, the second strategy can be implemented by a dialogue adjustment model.

[0065] For example, if the feedback indicates that the first dialogue information does not meet the user's needs, the first dialogue information is adjusted based on a third input information and a second strategy to obtain second dialogue information. The third input information can be a prompt used to generate the target dialogue information. Here, the user's needs can be the user's requirements for the dialogue information. That is, when the feedback information indicates dissatisfaction with the first dialogue information generated at the first moment, the first dialogue information can be updated.

[0066] Here, the update of the first dialogue information can be achieved through a dialogue adjustment model. This involves inputting the third input information into the dialogue adjustment model, which then combines this third input information to update the first dialogue information, resulting in the second dialogue information – essentially, regenerating the dialogue information. For example, if the user's feedback is "The generated plot was too simplistic; please regenerate it," the dialogue adjustment model can update the simplistic plot content to generate a more complex second dialogue.

[0067] The aforementioned second image, second video, and second dialogue information can be collectively referred to as second result information.

[0068] As an alternative approach, the associated information may also include virtual character information. During the process of obtaining the second result information, this technical solution can generate the second result information based at least on the feedback information and the character attributes of the virtual character. In other words, if the attributes of the virtual characters are different, the corresponding generated second result information may also be different.

[0069] For example, the attributes of a virtual character may include identity attributes, physiological attributes, personality attributes, and growth attributes. These attributes can be generated based on user needs or can be generated through dialogue between the virtual character and the user. For instance, if the virtual character's personality is humorous, then the generated second result information will include humorous information; that is, the second result information can be generated based on the virtual character's tone attribute.

[0070] Optionally, this technical solution can also generate second result information by combining the input scenario of the feedback information and the attributes of the virtual character. Here, different input scenarios may result in different generated second result information. For example, if the user inputs feedback information during a dialogue with the intelligent system, the second result information can be generated based on the attributes of the virtual character during the dialogue, using the unique attributes of the virtual character to generate the corresponding second result information. Alternatively, if the user inputs feedback information at the beginning of a dialogue with the intelligent system, the second result information can be generated using the default method.

[0071] Furthermore, since the meanings of the feedback information representations differ, the methods for generating the second result information also differ. For example, if the image generated by the feedback information representation does not meet the user's needs, the second result information can be generated according to the first method, where the description in the first method is associated with the image. Similarly, if the plot content generated by the feedback information representation does not meet the user's needs, the second result information can be generated according to the second method, where the description in the second method is associated with the plot.

[0072] After obtaining the second result information through adjustment, this technical solution can display the second result information, that is, proceed to step S240.

[0073] In step S240, the second result information is displayed.

[0074] In some implementations, the second result information can be directly output to the user, or it can be output to the agent first, so that the agent can forward the second result information to the user. That is, the second result information can be directly displayed on the user's electronic device, or it can be displayed on the agent's electronic device first, and then forwarded to the user by the agent's operation so that it can be displayed on the user's electronic device.

[0075] Furthermore, the second result information can be output based on the category of the feedback information. Different categories of feedback information will result in different ways of outputting the second result information. For example, if the feedback information pertains to a visual problem, the second result information can be output through a display device. Conversely, if the feedback information pertains to a speech problem, the second result information can be output through an audio device. There are no explicit restrictions on how the second result information is output; it can be chosen based on the actual situation.

[0076] In one scenario, this technical solution can display a first interface, which can be used to display first result information; that is, the first result information can be displayed through the first interface. In other words, when user input information is received and first result information is generated based on that input information, the first result information can be displayed on the first interface. For example, if the input information is "generate an image including a cat sitting on a sofa," after generating an image containing a cat sitting on a sofa based on that input information, the image can be displayed on the first interface.

[0077] Based on this, in response to a user's triggering of the second result information, the first result information can be replaced by the second result information and displayed on the first interface. That is, when the second result information is generated through feedback information and input information, this technical solution can display the second result information. When a user's input triggering operation on the second result information is detected, the first result information displayed on the first interface can be replaced with the second result information. For example, the generated second image can replace the first image generated at the first moment.

[0078] In another scenario, this technical solution can display a second interface. Here, the second interface can be used to receive feedback information and display second result information. That is, this technical solution can input feedback information through the second interface, and after generating second result information using the feedback information, the second result information can be displayed on the second interface.

[0079] Based on this, in response to the user's triggering of the second result information, the first result information can be replaced by the second result information and displayed on the first interface. That is, the second result information displayed on the second interface can replace the first result information displayed on the first interface.

[0080] In summary, the second result information generated by this technical solution can be displayed on the same interface as the first result information, or on different interfaces. When a user triggers an operation on the second result information, the second result information can be used to replace the first result information and displayed on the first interface.

[0081] As an alternative approach, in response to the first result information belonging to the second category of information, second result information can be obtained. In this case, the second result information can be generated based on the user's operation information and feedback information. Here, the second category of information may be a situation where result information cannot be regenerated, that is, the second result information in this case is incompatible with the first result information. In other words, when the first result information is detected to be second category information, this technical solution can generate second result information based on the user's operation information and feedback information.

[0082] In one scenario, the operation information may include associated operation information, which could be operation information related to the feedback information, such as an operation related to purchasing a membership. Here, the feedback information may be a feedback question. Upon receiving a user's feedback question, troubleshooting can be performed on that question to obtain the results.

[0083] For example, in response to the first result information belonging to the second category of information, this technical solution can generate the investigation result corresponding to the feedback information. Here, the investigation result can be generated based on the operation information and the query data sent by the investigation tool. The query data can be obtained by calling the investigation tool corresponding to the feedback information based on the model context protocol.

[0084] It should be noted that the information processing method can be applied to an information processing system, which can be a feedback and troubleshooting system based on an AI Agent (Artificial Intelligence Agent). This system may include the AI ​​Agent, an MCP Server (Model Context Protocol Server), and a Tool (troubleshooting tool). The troubleshooting tool can be used to query data related to the call request.

[0085] For example, the platform and data query capabilities needed for troubleshooting business issues can be encapsulated into tools and provided to the Agent for invocation through MCP (Model Context Protocol). This enables the intelligent customer service system to autonomously troubleshoot suspected defects, which not only improves the timeliness of feedback processing but also enhances the user experience.

[0086] The Agent can be built based on a target language development framework. This framework can provide atomic components, component orchestration, and aspect-based extension capabilities. For example, an Agent can be created on the target platform, and a System Prompt (SP) can be filled in. This SP can be found in [reference needed]. Figure 3 ,based on Figure 3It can be seen that the system prompts may include the Agent's role, task, processing logic, tool call rules, and precautions.

[0087] Before performing information processing, a configuration operation can be performed. This configuration operation can be used to configure tools for the Agent in a tool list, where the tool list can include tool names, tool descriptions, and tool parameters. For example, the tool list can include tool information for multiple tools, and each tool's tool information can include tool name, tool description, and tool parameters. For instance, the tool information for the first tool could include the tool name: "Effect Evaluation Tool," which is an asynchronous tool, and the tool description could be: "Sampling and scoring the user's recent performance based on their uid, without returning any results; this tool must be terminated immediately after invocation." As another example, the tool information for the second tool could include the tool name: "Log Error Information Query Tool," and the tool description could be: "Querying error information in the logs based on the logid (Log ID)."

[0088] Here, the tool configuration can be determined based on business needs; different business needs will require different tools. For example, when configuring a tool list for the Agent of the first application, information about the first tool, the second tool, and the third tool can be added to the tool list. The first, second, and third tools can be respectively as follows: Figure 4 The user points query tool (functional module 1), server / client log query tool (functional module 2), and user creation effect monitoring tool (functional module 3) shown can be obtained through MCP registration.

[0089] based on Figure 4 It is known that the user points query tool can be used to query points in Redis (Remote Dictionary Server) and to query transaction records in Relational Database Service (RDS); the server / client log query tool can be used to query creation records or status in RDS; and the user creation effect monitoring tool can be used to detect creation effects, such as similarity and aesthetics.

[0090] It should be noted that the tools can be downstream tools of the MCP Server, meaning that the tools configured in the Agent can support both synchronous and asynchronous modes. For example, synchronous tools can directly return results, while asynchronous tools can directly return a ToolCallID (tool call representation). After execution, this technical solution can send the execution result (query result) through a message queue (MQ). Once the Agent detects the asynchronous nature, it can terminate the current execution. Subsequent Chats, upon receiving MQ activation processing, can re-invoke the Agent to initiate the subsequent execution of this task.

[0091] Additionally, configuration operations can also be used to configure the Agent. To better illustrate the Agent configuration and initialization process, the following is provided: Figure 5 The example diagram shown is based on Figure 5 As can be seen, Agent configuration can be achieved through the AgentConfig Platform, and the configuration can then be persisted to a database (DB). Based on persistence, lazy loading can be performed. Furthermore, after Agent configuration is complete, it can be initialized. For example, initialization can be performed in real-time after Agent configuration, and / or all Agents can be initialized when the service starts.

[0092] It should be noted that during system initialization, the Agent's tool list, functional components, and model can be read to construct the intelligent system graph.

[0093] In one scenario, the Agent can support asynchronous execution of the tool; that is, when the tool is executing asynchronously, the Agent's execution can be terminated. See [link to documentation] for details. Figure 6 .based on Figure 6 It can be seen that when the Agent receives the execution result from the investigation tool, if it determines that the investigation tool is an asynchronous tool, that is, if it determines that the execution of the investigation tool is asynchronous, then the Agent can terminate its execution. In other words, the Agent does not need to wait for the execution result of the investigation tool; it can perform other operations.

[0094] Furthermore, if it's determined that the current investigation tool's operation is not asynchronous, the Agent can continue with other logic. That is, it waits for the current investigation tool to complete, receives the tool's execution result, and then continues with other logic based on that result. For example, if the first tool is synchronous, after receiving its execution result, the Agent can call the second tool and receive its execution result. Therefore, the Agent can support both asynchronous and synchronous execution of tools.

[0095] As explained above, MCP can provide encapsulated tools to the Agent for invocation. Therefore, before performing information processing, it can execute actions such as... Figure 7 The MCP service deployment shown here. The MCP service (MCP Server) can support both synchronous and asynchronous modes. Based on... Figure 7 It can be seen that the intelligent system (Agent) can obtain tool capabilities through the Model Context Protocol Service (MCPServer). That is, the tool function instance (tool_faas) can register its own capabilities with the corresponding MCP server (mcp_server), and can also report its own status to the online status management component (online_manage) every preset time interval (30s).

[0096] Here, the online status management component can be used to synchronize status to the first platform. The AI ​​intelligent system can call the tool capabilities provided by tool_faas through mcp_server, while the management component of the second platform can be used to store the management data of the MCP service in the corresponding database. This enables standardized invocation of MCP tools, real-time status monitoring, and collaborative management of multiple platforms such as the first and second platforms.

[0097] In summary, the Agent can register its required investigation tools through the MCP Server to ensure multi-tenant isolation of the tools. In other words, this technical solution can support multiple MCP Servers and multiple MCP Tools, meaning that multi-tenant tool isolation can be achieved through multiple MCP Servers.

[0098] Figure 8 This illustrates the isolation relationship between the Model Context Protocol service and the tool tenant, based on... Figure 8As can be seen, one tenant space can correspond to multiple Model Context Protocol servers (Mcp_Server), and one MCP server can be associated with multiple tools. This hierarchical relationship enables the isolated management of MCP services and tools in multi-tenant scenarios, ensuring that the resources of different tenants do not interfere with each other. Specifically, a tenant space can include attributes such as a space identifier (space_id); an MCP server can include attributes such as a service identifier (server_id) and the space identifier (space_id) of the associated space; and a tool can include attributes such as a tool identifier (tool_id) and the service identifier (server_id) of the associated Mcp_Server.

[0099] In another scenario, both the model context protocol service and the tools can be configurable to facilitate tool adjustments. Furthermore, the registry center can utilize a second platform for visual configuration, which not only increases readability but also facilitates logic such as field type validation. Simultaneously, different model context protocol services can register different tools; for example, the first model context protocol service (mcp_server_1) can register the first tool function instance (tool_faas_1), and the second model context protocol service (mcp_server_2) can register the second tool function instance (tool_faas_2).

[0100] The registration process for the above tools can be found in [link / reference]. Figure 9 ,based on Figure 9 As can be seen, the tool registration process can be mainly divided into six parts. The first part corresponds to steps 1 to 8, which enables the creation of a new tenant space in the second platform center. The second part corresponds to steps 9 to 12, which enables the registration of the required tools in the newly created tenant space. The third part corresponds to step 13, where the system can load space information through the Model Context Protocol (MCP) server during runtime. The fourth part corresponds to steps 14 to 18, where the MCP server can be used to obtain and bind the MCP services or tools registered under the space. The fifth part corresponds to steps 19 to 20, where the MCP server can be used to load tools. The sixth part corresponds to steps 21 to 22, where the MCP server can be used to monitor whether the tools have changed and dynamically load them when changes occur.

[0101] The above describes the complete process of R&D registration of MCP resources, MCP server initialization, and scheduled synchronization of tool configurations. Participants include R&D, the configuration center (config_center), the database (db), the Model Context Protocol server (mcp_serve), and the Model Context Protocol client (mcp_client). In this process, R&D can first create namespaces (corresponding tenant spaces), and the configuration center can create space records in the database and respond. Next, R&D can create the Model Context Protocol server, and the configuration center can create mcp_server records in the database and respond. Afterward, R&D can register tools with the Model Context Protocol server, and the configuration center can create / update tool records in the database and respond.

[0102] Based on this, the Model Context Protocol (MGP) server can load the configured spatial information to retrieve the corresponding MGP server from the database, and then obtain the tools associated with that server. Subsequently, each MGP server can listen to these tools. Furthermore, the MGP server can use a timer with a specified interval (1 minute) to load the tool configuration for the corresponding tenant from the database. If the tool list changes, it can notify the MGP client, which can then load the updated tool list.

[0103] After completing the processes of Agent configuration, Model Context Protocol (MTP) service deployment, MTP service registration, and tool registration, information processing operations can be performed when a call request is received.

[0104] As an optional approach, upon receiving feedback information, and if the first result information belongs to the second category, the user feedback information can be assembled to obtain candidate input information. Based on this, a screening tool is determined according to the candidate input information, wherein the screening tool can be used to query data related to the feedback information.

[0105] It should be noted that different user-input feedback information will require different investigation tools, resulting in different investigation results. Here, the investigation results can be considered as the second result information mentioned above.

[0106] Optionally, upon receiving feedback, the feedback type can be obtained, and based on this, it can be determined whether the feedback type is a problem feedback. If the feedback type is determined to be a problem feedback, information processing operations can be performed, namely, executing steps such as determining and calling the corresponding troubleshooting tool for the request, and outputting the troubleshooting results. Optionally, if the feedback type is determined to be a consultation / suggestion type, the specific intent can be identified through the Agent, and relevant content can be retrieved by calling the target business knowledge base to generate a response script for the agent. The agent can then reply to the user, and the process ends.

[0107] As an example, please see Figure 10 Developers (business personnel) can register MCP services and troubleshooting tools on the MCP service platform for subsequent Agent troubleshooting of online issues. Additionally, this technical solution can organize business documents and technical solution documents, performing document slicing and vectorization on a designated platform, and uploading the processed data to the platform's knowledge base for subsequent Agent troubleshooting to retrieve business knowledge. Afterwards, when an agent receives online user feedback, they can trigger an Agent through application A.

[0108] Based on this, if the intelligent system determines that the feedback is a consultation or suggestion, it can identify the user's specific intent, access the target business knowledge base to retrieve relevant content, generate a response script, and output the script to the agent. The agent can then refer to the script and send a reply to the user, ending the process.

[0109] Optionally, if the intelligent system determines that the feedback information is a defect, it can identify the intent, query business knowledge, and determine the troubleshooting steps and required tools. Then, it can query information such as user server / client behavior logs through the tools registered by the developers on the platform. Based on the queried information, the intelligent system can complete the troubleshooting accurately and efficiently. Finally, the intelligent system can send its conclusions to the agent so that the agent can reply to the user, and the process ends.

[0110] Here, the call request triggered by the feedback information can be an upstream call, which can be user feedback. That is, the upstream call can be triggered when the user inputs feedback information. Based on this, the intelligent system can perform pre-processing operations such as parameter conversion and context loading.

[0111] As an alternative approach, upon receiving an asynchronous callback request, this technical solution can obtain context information. Specifically, when the received call request is an asynchronous callback request, context information can be obtained. This context information may include the first query data from the first tool. Based on this, the second tool can be determined according to the context information, and the second query data sent by the second tool can be received. Then, a troubleshooting result is generated based on the second query data and output.

[0112] In summary, when it is determined that the call request is a non-asynchronous callback (feedback information), the call request can be encapsulated first, then the troubleshooting tool can be determined based on the large language model, and the troubleshooting operation can be performed based on the query data returned by the troubleshooting tool. Optionally, when it is determined that the call request is an asynchronous callback, i.e., when an asynchronous callback request is received, context information can be generated based on the temporarily stored query data returned by the asynchronous tool, and then input into the large language model to perform subsequent operations.

[0113] It's important to note that before determining the investigation tool corresponding to a call request, the intelligent system can decide whether to execute the tool invocation operation based on the request itself. If it decides to execute the tool invocation operation, it can then identify the corresponding investigation tool. In other words, upon receiving a call request, it can first determine whether to invoke a tool. If it does, it can further determine the tool's invocation mode—whether it's asynchronous or synchronous. Based on this, it can then perform targeted information processing operations according to the different invocation modes.

[0114] As an alternative approach, after identifying the investigation tool corresponding to the call request, the investigation tool can be invoked based on the Model Context Protocol (MCP). This instructs the investigation tool to query data related to the call request, and the intelligent system can then receive the query data sent by the investigation tool.

[0115] As described above, after identifying the investigation tool, it can be determined whether the tool's calling mode is asynchronous. If it is determined to be asynchronous, the query data sent by the investigation tool can be temporarily stored.

[0116] For example, when the invocation mode of the first tool is determined to be asynchronous, the intelligent system can temporarily store the first query data. Subsequently, when it receives an asynchronous callback request, it can generate context information based on the temporarily stored first query data.

[0117] Optionally, if the invocation mode of the first tool is determined to be synchronous, the first query data from the first tool can be received, and investigation results can be generated based on the first query data. In other words, if the investigation tool is determined to be asynchronous, the query data transmitted by the asynchronous tool can be temporarily stored. Conversely, if the investigation tool is determined to be synchronous, subsequent operations can continue after obtaining the query data from the synchronous tool.

[0118] In one scenario, a call request may or may not have a corresponding troubleshooting tool. That is, if it's determined that a tool needs to be called, then it can be assumed that a corresponding troubleshooting tool exists for that call request. Conversely, if it's determined that a tool doesn't need to be called, then it's assumed that a corresponding troubleshooting tool doesn't exist for that call request. Furthermore, when it's determined that a tool needs to be called, the same call request can correspond to one troubleshooting tool, or it can correspond to multiple troubleshooting tools. The exact number of troubleshooting tools is not explicitly limited here.

[0119] It should be noted that the query data sent by the investigation tool can be standard query data obtained after preprocessing. That is, when the investigation tool obtains the initial query data through the query, it can preprocess the initial query data to obtain standard query data. Based on this, the intelligent system can receive the standard query data sent by the investigation tool.

[0120] For example, two forms of interface returns can be supported here to assist the intelligent system in understanding. The first form of interface can be defined by JSON data + IDL (Interface Definition Language), and the second form of interface can be described in natural language.

[0121] As an example, the creation module and other non-integration modules return fewer and easier-to-understand fields. Agents can better understand task status, error types, and multi-dimensional creation effects through IDL. Therefore, after obtaining the initial query data, such tools can convert the initial query data into JSON data + IDL definition.

[0122] As another example, the information screening of the points module requires the agent to understand and process timestamps, and to perform screening by combining the user feedback time with the current time (temporal issues). In practice, the agent performs poorly in processing temporal information. Therefore, the data returned by the points module interface can be translated into a fixed form of natural language description and sent to the intelligent system, that is, the initial query data is converted into a natural language description.

[0123] In summary, after obtaining the initial query data, we can first determine the category of the data output by the screening tool. Based on this, we can preprocess the initial query data according to the category, that is, preprocess the initial query data into standard query data.

[0124] Optionally, upon receiving query data from the investigation tool, the intelligent system can also determine whether the query data meets preset conditions. If it does, the system can generate investigation results based on the query data. Conversely, if the query data does not meet the preset conditions, the agent can preprocess the query data to obtain standard query data, and then generate investigation results based on this standard query data. Here, the preprocessing of the query data can be determined according to the function and category of the investigation tool.

[0125] For example, when the investigation tool is determined to be a first-category tool, the query data format is converted to JSON data + IDL format; conversely, if the investigation tool is determined to be a second-category tool, the query data is converted to natural language description. Here, first-category tools are characterized by returning fewer and easier-to-understand fields; second-category tools have timestamp processing requirements, such as needing to combine user feedback time and current time for comprehensive determination.

[0126] It's important to note that after completing the data query operation, the investigation tool can instruct the MCP service via MQ. The MCP service can then encapsulate the execution results into a message list and append them to the Message List. This appending of query data can be implemented through the MCP service. Subsequent calls to the Agent can generate investigation results based on the updated MessageList. In one scenario, the Agent can be capable of sensing environmental changes and replanning and executing accordingly.

[0127] To chain multiple asynchronous executions, a QueryTask entity can be used to represent a single user request; that is, the QueryTask entity can represent the entire processing flow of the intelligent system. Without asynchronous execution, one QueryTask corresponds to one Agent execution; with asynchronous execution, one QueryTask can correspond to two or more rounds of Agent execution.

[0128] For example, when performing information processing, the intelligent system can write streaming results to Redis. The Redis instance can be configured with the following structure: `zset (Sorted Set); key: query_task_id; score: timestamp; value: streaming message content; timeout: 1 day`. Simultaneously, the Agent can return the upstream query_task_id. The upstream can then establish a streaming connection using this query_task_id, allowing the intelligent system to continuously read content from Redis and return it.

[0129] As an optional approach, after receiving the query data sent by the investigation tool, an investigation result (second result information) can be generated based on the query data, and then displayed. As described above, when a single request corresponds to multiple investigation tools, these tools can execute in parallel or asynchronously. The intelligent system can analyze the query data returned by these tools to generate the final investigation result.

[0130] To better illustrate the process by which the Agent generates investigation results, the following is given: Figure 11 The example diagram shown is based on Figure 11 It is known that when the intelligent system receives an upstream request (user feedback request), it can perform pre-processing such as parameter conversion and context loading. Additionally, upon receiving a call request, it can determine whether the call is an asynchronous callback. If not, it can enter the SP (System Hint) processing (ChatModel Template), and then input it into the ChatModel (dialogue model) to process the input parameters, such as performing parameter conversion. In other words, if the call is a non-asynchronous callback, the intelligent system can use the ChatModel Template to fill the user-provided variable values ​​into a predefined message template to generate a standard format message for interaction with the language model. This assembles the user feedback request (feedback information) to obtain the target input information.

[0131] Optionally, if it is determined that the request is an asynchronous callback request, the Agent can execute subsequent processes based on the Message List in the callback request, i.e., perform a context recovery operation. That is, it obtains context information from the MessageList and inputs this context information into the large language model. During this process, the large language model can determine whether to call a tool based on the context information (ChatModel information). If it is determined that no tool call is needed, it can directly encapsulate the large model's input content and reply, i.e., execute the termination node.

[0132] Optionally, if it is determined that a tool needs to be invoked, the Agent can invoke the relevant tool through the MCP Server to query the user's associated data, i.e., obtain the query data. During this process, the Agent can first determine whether the tool is asynchronous. If the investigation tool is asynchronous, the context can be saved, and subsequent processes can be triggered when the tool callback occurs. If the investigation tool is synchronous, the tool's return result and context can be passed through the ChatModel to execute subsequent processes, ultimately generating and returning the information processing result. This information processing result can be the investigation result, i.e., the query data and context are input into the ChatModel to generate the investigation result.

[0133] As an example, the above information processing can be achieved through the interaction between the intelligent system, the model context protocol server, and the tool. Please refer to the interaction process. Figure 12 .based on Figure 12 It can be seen that when the intelligent system receives feedback information from the user, it can first determine the type of feedback information. If the feedback information is determined to be a suggestion, the intelligent system can generate a response and output it to the user. If the feedback information is determined to be a defect, the intelligent system can determine whether investigation is needed. If investigation is needed, it can determine the investigation steps and required tools. Based on this, the intelligent system can send a query instruction to the Model Context Protocol (MCP) server. This query instruction can include information about the investigation tools. Sending a query instruction to the MCP server enables the invocation of the investigation tools, i.e., instructing the investigation tools to perform data queries. Afterwards, the intelligent system can receive the query results sent by the MCP server and append them to the message list. Based on the context information of this message list, the intelligent system can generate the investigation results.

[0134] Here, the intelligent system can include an Agent Server (intelligent system service) and an intelligent system agent. The Agent Server can perform append operations upon receiving query results and invoke the intelligent system agent, thus triggering a call request. Furthermore, upon receiving a call request, if it is determined that the call request is an asynchronous callback request, the agent can re-enter the investigation steps and the determination process of the investigation tools until it is determined that the call is synchronous, at which point the above loop ends.

[0135] For example, when receiving feedback from a user, the agent can trigger an investigation through a robot. The robot can simultaneously provide the user with information about the investigation in progress, which means that asynchronous execution can be performed. In other words, the intelligent system can perform operations such as thinking process, tool invocation, information organization and analysis, and generating a response. Finally, the response information can be output to the A application card. This not only reduces labor costs but also improves the user feedback experience.

[0136] In other words, the intelligent system can not only achieve intent recognition, knowledge retrieval, and response generation, but also automatically resolve issues. Furthermore, during the troubleshooting process, users can directly interact with the intelligent system, which automatically completes the identification, retrieval, investigation, and response generation without the need for multi-party intervention or R&D. The information processing system can complete these tasks and generate responses within minutes, thus improving the user experience to a certain extent.

[0137] By utilizing information processing systems, more personalized user questions can be addressed, significantly reducing the cost of manual feedback processing. Furthermore, using artificial intelligence systems to process feedback not only greatly shortens processing time but also delivers results to users faster, improving the user experience. Moreover, intelligent systems possess high intent recognition accuracy, enabling the generation of personalized dialogue messages for different feedback types, further enhancing the user experience.

[0138] As described above, for different user feedback requests (feedback information), during the information processing, one or more investigation tools are used to obtain query data. Then, based on this query data, a comprehensive investigation result can be generated. To better illustrate the process of the Agent calling the investigation tools, the following is provided: Figure 13 and Figure 14 The example diagram is shown.

[0139] As an example, based on Figure 13As we can see, when a user reports "Member points not credited," the Agent can determine the troubleshooting tool, meaning the Agent can enter the first "Thought" stage. For example, it can consider, "The user reported that member points haven't been credited, which is a points / commercialization issue. We need to determine whether we need to use a tool for analysis." It should be noted that due to the characteristics of points / commercialization feedback, the Agent can skip this process to conserve tokens. After determining the first tool, the Agent can enter the first "Action" stage, which involves calling the first tool, such as calling the initial problem screening tool (inputting the original feedback). Based on this, the Agent can enter the first "Observation" stage, which involves summarizing the query data from the previous troubleshooting tool, such as, "Identified as a points / commercialization issue; the user's feedback information needs to be further investigated using points-related tools."

[0140] Next, the agent can enter the second thinking stage, such as considering "the need to use tools to analyze the reasons for the points anomaly: calling the [Query User Subscription Records] tool to confirm the membership activation and activation time." After identifying the second tool, the "Query User Subscription Records tool," the agent can enter the second execution stage, which involves calling the second tool, such as calling the Query User Subscription Records tool to retrieve points details. Based on this, the agent can enter the second observation stage, which involves summarizing the data from the previous investigation tool, such as "A user renewed their membership on November 3rd, and 1080 points will be effective on November 5th at 00:00; a user renewed their membership on October 2nd, and 1080 points have already been effective, with 827 points remaining."

[0141] Based on this, the agent can first enter the third thinking stage, such as considering the discrepancy between "827 points and the user's reported 835 points," and then calling the "[Get Points Details]" tool to check recent points credit details. After identifying the third tool, "Get Points Details," the agent can enter the third execution stage, which involves calling the third tool, such as the "Get Points Details" tool. Following this, the agent can enter the third observation stage, summarizing the data from the previous investigation tool, such as "Credits: User logged in on November 3rd, earned 60 points, 60 points remaining (effective)," etc.

[0142] The first, second, and third tools mentioned above can be point query tools.

[0143] Finally, the Agent can proceed to the fourth "Thought" stage. For example, it could consider: "Currently effective are the remaining 827 points due on October 5th and the 60 daily login points from November 4th, totaling 887. This is largely consistent with the user feedback of 885, indicating a misunderstanding due to the points not yet expiring, not a bug." Based on this, the Agent can then proceed to the fourth "Action" stage, generating a response message, such as: "Your membership points for this month will be effective on November 5th at 00:00:00. Currently available points are the remaining points from last month and today's login points. You can view your points details in the homepage - top right corner - My Points section. Thank you for your feedback."

[0144] As another example, based on Figure 14 As can be seen, when a user feedback request is "the user reports a recent decline in performance," the agent can determine the investigation tool, i.e., the agent can enter the first Thought stage. For example, it can consider "firstly using a preliminary screening tool to clarify the problem type and confirm whether investigation is needed." After determining the first tool, the agent can enter the first Action stage, i.e., calling the first tool, such as calling the problem preliminary screening tool (inputting the original feedback audio). Based on this, the agent can enter the first Observation stage, i.e., summarizing the query data from the previous investigation tool, such as "the tool identified the result as [performance issue], requiring a dedicated investigation process."

[0145] Next, the Agent can enter the second Thought stage, such as considering that "the effectiveness issue needs to be verified with objective data, and the sampling scoring tool can be used to obtain indicator trends." After identifying the fourth tool, the "effectiveness sampling scoring tool," the Agent can enter the second Action stage, which involves calling the fourth tool, such as calling the effectiveness sampling scoring tool (entering the uid). Based on this, the Agent can enter the second Observation stage, which involves summarizing the data queried by the previous investigation tool, such as "the tool returns scores and trends for various core indicators, and the analysis shows that the indicators are normal, meaning the investigation result is no deterioration and meets expectations."

[0146] Building on this foundation, the Agent can first enter the third Thought stage, such as considering whether "normal metrics indicate stable performance, and further examining user input suggestions and reference images." After identifying the fifth tool, the "Creative Suggestion Understanding and Evaluation Tool," the Agent can enter the third Action stage, which involves invoking the fifth tool, such as calling the Creative Suggestion Understanding and Evaluation Tool, inputting user suggestions, and receiving creative recommendations. Based on this, the Agent can then enter the third Observation stage, summarizing the query data from the previous investigation tool, such as "the large language model returns user input evaluations."

[0147] The first, fourth, and fifth tools mentioned above can be tools for detecting creative effects.

[0148] Finally, the agent can enter the fourth thinking stage, such as considering that "the user prompts are vague and simple, lacking descriptions of scenarios, actions, and expressions." Based on this, the agent can enter the fourth execution stage, which involves generating response information, such as "The tool has no obvious effect; you can further refine the prompts by considering aspects such as environment, actions, and expressions."

[0149] As an optional approach, after generating the investigation results, a first response message can be generated based on those results and output to the first user. Here, the first user can be the user who inputs feedback information, i.e., the user who submitted the feedback issue. The first response message serves to inform the first user of the current progress of the investigation and suggested solutions.

[0150] Please see Figure 15 If the user inputs feedback information as "Program A automatically deducted the fee but did not give me points", the first reply information output by the electronic device to the first user could be "Your membership points were effective on December 8, 2025 at 00:00:00. You currently have 1057 points remaining. You can go to the homepage - points in the upper right corner - my points to view the details of points earned and consumed. Thank you for your feedback."

[0151] It should be noted that the output of the first response information is similar to the output of the second result information, and it can also be based on the attribute information of the virtual character.

[0152] Optionally, after generating the investigation results, a second response message can be generated based on the investigation results and output to a second user. Here, the second user can be the user who is fixing the problem, such as a developer. The second response message serves to provide suggestions for fixing the problem.

[0153] In addition, the output for the second user can include not only the second response, but also feedback on the problem, the first response to the first user, and query data obtained by the Agent when calling the troubleshooting tools. This can help the second user locate the problem more quickly. For details on the dialogue information for the second user, please refer to [link / reference needed]. Figure 16 .

[0154] In summary, when generating troubleshooting results, on the one hand, the troubleshooting progress and solution suggestions can be provided to users to ensure a good user experience; on the other hand, fix suggestions can be provided to developers to ensure the timeliness of problem fixing.

[0155] In another scenario, feedback can be triggered through a target feedback group. Users can submit issue feedback to an external feedback group, and the agent, upon receiving this feedback, can generate investigation results based on the feedback and output these results to the user. Alternatively, users can enter feedback information within a designated window of a specified application, which can also trigger the sending of a user feedback request.

[0156] As an alternative approach, upon receiving a user feedback request (feedback information), the intelligent system can determine the category to which the user feedback request belongs. Based on this, it can determine the corresponding investigation strategy for that category and finally implement the investigation based on the corresponding strategy. In other words, different user-input feedback information will result in different investigation strategies executed by the Agent, thus ensuring the accuracy of information processing to a certain extent. Furthermore, for the same type of problem, the investigation strategy may differ depending on the user, the scenario, and the context.

[0157] As an example, when the user's feedback is a type 1 problem, and analysis of the user's historical data confirms that it meets the first condition, the first screening strategy can be executed. For example... Figure 17 As shown, user A's feedback message is "Frequent generation failures," and analysis determines that the first user meets the first condition. Therefore, the following can be used: Figure 18 The strategy shown is the first screening strategy. For example, the first condition could be that user A's approval rate exceeds a preset threshold during the creation process.

[0158] based on Figure 18Therefore, the first investigation strategy can be as follows: If it's determined that the issue submitted by User A is a creation-related issue, the agent can query User A's most recently incomplete creation records and analyze their status, i.e., extract error information and log information. Based on this, an API is called to query the error information in the logs using the login ID, and then it's determined whether the problem can be located through the logs. If it can be located, the agent can output the investigation results based on the error information. Conversely, if it cannot be located, i.e., the status is abnormal, the agent can check for any risk control restrictions. If risk control is confirmed, the agent can output the investigation result "The problem was caused by risk control"; if no risk control is confirmed, the agent can mark this feedback information as an abnormal status requiring further investigation. Based on this, the investigation ends.

[0159] As another example, when the user's feedback is a type 1 problem, and analysis of the user's historical data determines that it does not meet the first condition, a second screening strategy can be implemented. For example... Figure 19 As shown, user B's feedback message is "Images failed to be generated, and clicking 'retry' says 'operation is too frequent to retry'". Analysis confirms that the first user does not meet the first condition. Therefore, the following can be used: Figure 20 The strategy shown is the second investigation strategy.

[0160] based on Figure 20 Therefore, the second investigation strategy can be as follows: When it is determined that the issue submitted by User B is a creation-type issue, the agent can query User B's most recent incomplete creation records and determine if there is any "approved by review" information in the creation records. If it is determined that there is an approved by review information in the creation records, the investigation result "approved by review leading to incompleteness" is directly output. Conversely, if it is determined that there is no approved by review information in the creation records, the record status can be analyzed, that is, error information and log information can be extracted. Based on this, the interface is called, that is, the error information in the log is queried according to the logid, and then it is determined whether the log shows information related to approved by review. If the log shows approved by review, the investigation result "approved by review leading to incompleteness" is output. Conversely, if the log shows approved by review, the investigation result "other reasons leading to incompleteness" can be output, and the log information is output. Afterwards, the Agent can summarize all the investigation information to form a final conclusion. Based on this, the investigation ends.

[0161] The first type of problem mentioned above can be a creation failure problem. For example, the first example could be troubleshooting a creation failure problem, and the second example could be troubleshooting a creation that fails the review process.

[0162] As another example, when the user's feedback is a type 2 problem, a third investigation strategy can be implemented. For example... Figure 21As shown, user C's feedback message is "The clone has been created for two days and it's still not finished." At this point, you can... Figure 22 The strategy shown is the third investigation strategy.

[0163] based on Figure 22 Therefore, the third investigation strategy can be as follows: When it is determined that the issue submitted by user C is a clone feedback investigation type issue, the most recent clone status records of user C can be queried to preliminarily determine whether the clone status is clear. If the clone status is clear, the status is directly recorded and used as the basic investigation result. Conversely, if the clone status is unclear, i.e., the status is determined to be abnormal and cannot be determined, the logid corresponding to the abnormal clone can be extracted. Based on this, the log query interface is called to obtain information based on the logid, then the log error information is analyzed, and the cause of the abnormality is added to the investigation result. Afterwards, the Agent can summarize and analyze the status judgment result and log analysis, and output the final investigation report. Based on this, the investigation ends.

[0164] The second type of problem mentioned above can be used to troubleshoot issues related to clone training. It should be noted that users A, B, and C can be users who input feedback information, or they can be developers; there is no specific restriction on which. Furthermore, the first result information corresponding to the first and second types of problems can belong to the second category of information mentioned above.

[0165] The above-mentioned investigation strategy can also be called investigation steps. When the agent determines that it needs to perform investigation operations on the feedback information, it can determine the investigation steps and tools corresponding to the feedback information, and perform subsequent investigation operations based on the investigation steps and tools to obtain the investigation results.

[0166] In one scenario, a customized intent recognition agent based on a project knowledge base can be developed. This effectively identifies user feedback intent, meaning it can recognize the original audio of feedback regarding defects being investigated. Furthermore, through the MCP (Multi-Channel Programming) framework, an information processing API (Application Programming Interface) can be implemented for large-scale models. This allows the model to input collected user feedback along with information such as the user's device and account details, enabling it to identify the current status of the user's account and whether an error has occurred. Based on this, the API documentation serves as a knowledge base, allowing the large-scale model to automatically parse the reasons for user errors and provide solutions. Moreover, the troubleshooting results parsed by the large-scale model can be processed through Prompt Engineering (PE) to refine the response scripts, providing personalized and targeted dialogue information to a certain extent, thereby improving user feedback satisfaction.

[0167] Based on the same inventive concept, an information processing device is also provided. Figure 23 This is a block diagram illustrating an information processing apparatus 2300 according to an exemplary embodiment, such as... Figure 23 As shown, the information processing device 2300 may include a receiving module 2310, a first acquisition module 2320, a second acquisition module 2330, and a display module 2340.

[0168] The receiving module 2310 is configured to receive user feedback information; The first acquisition module 2320 is configured to acquire associated information based on the feedback information; wherein the associated information includes at least the user's operation information and first result information, the first result information being generated based on the operation information; The second acquisition module 2330 is configured to acquire second result information, which is generated based at least on a portion of the associated information. The display module 2340 is configured to display the second result information.

[0169] In some implementations, the second acquisition module 2330 is further configured to acquire the second result information in response to the first result information belonging to a first category of information, wherein the second result information is generated based on the user's operation information and the first result information.

[0170] In some implementations, the first result information includes first multimedia information, the operation information includes input information, and the second acquisition module 2330 is further configured to adjust the first multimedia information based on a first strategy and the input information in response to the first multimedia information belonging to a first category of information, to obtain second multimedia information.

[0171] In some implementations, the first result information includes first dialogue information, the operation information includes input information, and the second acquisition module 2330 is further configured to adjust the first dialogue information based on a second strategy and the input information in response to the first dialogue information belonging to a first category of information, thereby obtaining second dialogue information.

[0172] In some implementations, the second acquisition module 2330 is further configured to acquire the second result information in response to the first result information belonging to the second category of information, the second result information being generated based on the user's operation information and feedback information.

[0173] In some implementations, the operation information includes associated operation information, and the second acquisition module 2330 is further configured to generate a screening result corresponding to the feedback information in response to the first result information belonging to the second category of information. The screening result is generated based on the operation information and the query data sent by the screening tool. The query data is obtained by calling the screening tool corresponding to the feedback information based on the model context protocol.

[0174] In some embodiments, the information processing apparatus 2300 may further include: The generation module is configured to, upon receiving an asynchronous callback request, obtain context information, the context information including first query data of a first tool; obtain a second tool based on the context information; invoke the second tool based on the model context protocol and receive second query data sent by the second tool; and generate the investigation result based on the second query data.

[0175] In some implementations, the generation module is further configured to temporarily store the first query data when the first tool is invoked in an asynchronous mode; and to generate the context information based on the first query data when the asynchronous callback request is received.

[0176] In some implementations, the generation module is further configured to receive first query data from the first tool when the first tool is in synchronous invocation mode; and generate the investigation results based on the first query data.

[0177] In some implementations, the associated information also includes virtual character information, and the second acquisition module 2330 is further configured to acquire second result information, which is generated based at least on the feedback information and the character attributes of the virtual character.

[0178] In some embodiments, the display module 2340 is further configured to display a first interface for displaying the first result information; in response to a user triggering the second result information, the first result information is replaced with the second result information, and the first interface is displayed.

[0179] In some embodiments, the display module 2340 is further configured to display a second interface for receiving the feedback information and displaying the second result information; in response to a user triggering the second result information, the first result information is replaced with the second result information, and the first interface is displayed.

[0180] The following is for reference. Figure 24The diagram illustrates the structure of an electronic device 2400 suitable for implementing the above-described technical solution. The terminal device may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 24 The electronic device shown is merely an example and should not be construed as limiting its functionality or scope of use.

[0181] like Figure 24 As shown, electronic device 2400 may include a processing device (e.g., a central processing unit, a graphics processor, etc.) 2401, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 2402 or a program loaded from storage device 2408 into random access memory (RAM) 2403. The RAM 2403 also stores various programs and data required for the operation of electronic device 2400. The processing device 2401, ROM 2402, and RAM 2403 are interconnected via bus 2404. Input / output (I / O) interface 2405 is also connected to bus 2404.

[0182] Typically, the following devices can be connected to I / O interface 2405: input devices 2406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 2407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 2408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 2409. Communication device 2409 allows electronic device 2400 to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 24 An electronic device 2400 with various devices is shown; however, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0183] In particular, depending on certain circumstances, the processes described in the flowchart above can be implemented as computer software programs. For example, a computer program product is provided, comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. In such embodiments, the computer program can be downloaded and installed from a network via communication device 2409, or installed from storage device 2408, or installed from ROM 2402. When the computer program is executed by processing device 2401, it performs the functions defined in the above-described methods.

[0184] It should be noted that the aforementioned computer-readable medium can be a computer-readable signal medium, a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In one case, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In another case, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0185] In some implementations, clients and servers can communicate using any currently known or future-developed network protocol such as HTTP (Hypertext Transfer Protocol) and can interconnect with digital data communication (e.g., communication networks) of any form or medium. Examples of communication networks include local area networks (“LANs”), wide area networks (“WANs”), the Internet (e.g., the Internet of Things), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future-developed networks.

[0186] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0187] The aforementioned computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: receive user feedback information; obtain associated information related to the feedback information based on the feedback information; wherein the associated information includes at least the user's operation information and first result information, the first result information being generated based on the operation information; obtain second result information, the second result information being generated based at least on a portion of the associated information; and display the second result information.

[0188] Computer program code for performing the above operations can be written in one or more programming languages ​​or a combination thereof. These programming languages ​​include, but are not limited to, object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0189] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0190] The modules mentioned above can be implemented in software or hardware. The names of the modules do not necessarily limit the module itself; for example, the second acquisition module can also be described as "acquiring second result information, which is generated based at least on a portion of the associated information."

[0191] The functions described above in this document can be performed at least in part by one or more hardware logic components. For example, exemplary types of hardware logic components that can be used, without limitation, include: field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), and so on.

[0192] In this context, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0193] The above description is merely illustrative and explains the technical principles employed. Those skilled in the art should understand that the scope of disclosure in the technical solution is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features provided herein that have similar functions.

[0194] Furthermore, although the operations are described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or in a sequential order. Multitasking and parallel processing may be advantageous in certain environments. Similarly, while several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the technical solution. Certain features described in the context of a single embodiment may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented individually or in any suitable sub-combination in multiple embodiments.

Claims

1. An information processing method, the method comprising: Receive user feedback information; Based on the feedback information, associated information related to the feedback information is obtained; wherein, the associated information includes at least the user's operation information and first result information, and the first result information is generated based on the operation information; Obtain second result information, which is generated based at least on a portion of the aforementioned associated information. Display the second result information.

2. The information processing method according to claim 1, wherein obtaining the second result information includes: In response to the first result information belonging to the first category of information, the second result information is obtained, which is generated based on the user's operation information and the first result information.

3. The information processing method according to claim 2, wherein the first result information includes first multimedia information, the operation information includes input information, and the step of obtaining the second result information in response to the first result information belonging to a first category of information includes: In response to the first multimedia information belonging to the first category of information, the first multimedia information is adjusted based on the first strategy and the input information to obtain the second multimedia information.

4. The information processing method according to claim 2, wherein the first result information includes first dialogue information, the operation information includes input information, and the step of obtaining the second result information in response to the first result information belonging to a first category of information includes: In response to the first dialogue information belonging to the first category of information, the first dialogue information is adjusted based on the second strategy and the input information to obtain the second dialogue information.

5. The information processing method according to claim 1, wherein obtaining the second result information includes: In response to the first result information belonging to the second category of information, the second result information is obtained, which is generated based on the user's operation information and feedback information.

6. The information processing method according to claim 5, wherein the operation information includes associated operation information, and the step of obtaining the second result information in response to the first result information belonging to the second category of information includes: In response to the first result information belonging to the second category of information, an investigation result corresponding to the feedback information is generated. The investigation result is generated based on the operation information and the query data sent by the investigation tool. The query data is obtained by calling the investigation tool corresponding to the feedback information based on the model context protocol.

7. The information processing method according to claim 6, further comprising: Upon receiving an asynchronous callback request, obtain context information, which includes the first query data of the first tool; The second tool is obtained based on the context information; The second tool is invoked based on the model context protocol, and the second query data sent by the second tool is received. The investigation results are generated based on the second query data.

8. The information processing method according to claim 7, further comprising: When the first tool is invoked in asynchronous mode, the first query data is temporarily stored. Upon receiving an asynchronous callback request, obtaining context information includes: Upon receiving the asynchronous callback request, the context information is generated based on the first query data.

9. The information processing method according to claim 7, wherein the method comprises: When the first tool is in synchronous invocation mode, the first query data of the first tool is received; The investigation results are generated based on the first query data.

10. The information processing method according to any one of claims 6 to 9, wherein displaying the second result information includes: Based on the investigation results, a first response message is generated and displayed to the first user. The first response message is used to indicate the progress of the investigation and provide solutions.

11. The information processing method according to any one of claims 1 to 9, further comprising: Display a first interface, which is used to display the first result information; In response to the user's triggering of the second result information, the first result information is replaced with the second result information, and the first interface is displayed.

12. The information processing method according to claim 11, further comprising: Display a second interface, which is used to receive the feedback information and display the second result information; In response to the user's triggering of the second result information, the first result information is replaced with the second result information, and the first interface is displayed.

13. An information processing apparatus, the apparatus comprising: The receiving module is configured to receive user feedback information; The first acquisition module is configured to acquire associated information related to the feedback information based on the feedback information; wherein the associated information includes at least the user's operation information and first result information, and the first result information is generated based on the operation information; The second acquisition module is configured to acquire second result information, which is generated based at least on a portion of the aforementioned associated information. The display module is configured to display the second result information.

14. A computer-readable medium having a computer program stored thereon, which, when executed by a processing device, implements the steps of the method according to any one of claims 1-12.

15. An electronic device comprising: A storage device on which computer programs are stored; A processing device for executing the computer program in the storage device to implement the steps of the method according to any one of claims 1-12.

Images