Enterprise digital management method and system based on artificial intelligence

Abstract

The embodiment of the invention discloses an enterprise digital management method and system based on artificial intelligence. The method comprises the steps of obtaining multi-modal data corresponding to a to-be-processed task, wherein the multi-modal data comprises image data and text data; determining a target resource required by the to-be-processed task based on the image data and the text data; obtaining a real-time resource state of the enterprise, and determining a resource scheduling strategy of the enterprise based on the real-time resource information and the target resource; and generating a plurality of to-be-processed sub-tasks based on the resource scheduling strategy, and distributing the plurality of to-be-processed sub-tasks to different departments corresponding to the enterprise. According to the embodiment of the invention, the overall task processing period can be greatly shortened, and the enterprise operation efficiency and response speed are improved.

Description

Technical Field

[0001] The present application relates to the field of enterprise management technology, and specifically to an enterprise digital management method and system based on artificial intelligence. Background Art

[0002] With the acceleration of enterprise digital transformation, task management has become a core component of improving organizational efficiency. Traditional task management systems rely on manual input or single-modal data (such as plain text input), manual monitoring, and static rules. This leads to bottlenecks such as delayed responses, missed tasks, and high labor costs. Existing tools are unable to intelligently identify the emotional tendencies and priorities in user messages. When text, images, and rich media are mixed, they struggle to deeply understand the complex semantic structure. Furthermore, they lack flexibility in task allocation, resulting in inefficient management. Summary of the Invention

[0003] To solve the above technical problems, the embodiments of the present application provide an enterprise digital management method and system based on artificial intelligence, an electronic device, a computer-readable storage medium, and a computer program product.

[0004] According to one aspect of an embodiment of the present application, an artificial intelligence-based enterprise digital management method is provided, including: obtaining multimodal data corresponding to a task to be processed, the multimodal data including image data and text data; determining the target resources required for the task to be processed based on the image data and the text data; obtaining the real-time resource status of the enterprise, and determining the resource scheduling strategy of the enterprise based on the real-time resource information and the target resources; generating multiple subtasks to be processed based on the resource scheduling strategy, and distributing the multiple subtasks to be processed to different departments corresponding to the enterprise.

[0005] According to one aspect of an embodiment of the present application, the method also includes: determining the object information and content information corresponding to the image data, and determining the semantic features corresponding to the text data; cross-modally fusing the object information, the content information and the semantic features to obtain structured semantic labels; and determining the target resources corresponding to the task to be processed based on the structured semantic labels.

[0006] According to one aspect of an embodiment of the present application, determining the resource scheduling strategy of the enterprise based on the real-time resource information and the target resource includes: determining the task attributes corresponding to the task to be processed; determining the processing priority corresponding to the task to be processed based on the task attributes, the target resource and the real-time resource information; and determining the resource scheduling strategy of the enterprise based on the processing priority.

[0007] According to one aspect of an embodiment of the present application, the method also includes: real-time monitoring of the multimodal data and newly added task processing rules sent by the preset data interface of the enterprise; updating the corresponding set of pending tasks of the enterprise based on the multimodal data and the newly added task processing rules; updating the processing priority of the pending tasks based on the set of pending tasks, so as to adjust the resource scheduling strategy of the enterprise based on the updated processing priority.

[0008] According to one aspect of an embodiment of the present application, the historical resource scheduling strategy and preset task processing rules of the enterprise are obtained; the executability of the resource scheduling is determined based on the historical resource scheduling strategy and the preset task processing rules; if the executability is greater than a preset executability threshold, the resource scheduling strategy is executed; model training is performed based on the resource scheduling strategy and the historical resource scheduling strategy, and a knowledge graph is obtained based on the trained model to determine the resource scheduling strategy of the enterprise through the knowledge graph.

[0009] According to one aspect of an embodiment of the present application, the method also includes: parsing the multimodal data to determine the data to be encrypted in the multimodal data based on the parsing results; encrypting the data to be encrypted based on differential privacy and federated learning to obtain encrypted multimodal data; and determining the target resources corresponding to the task to be processed based on the encrypted multimodal data.

[0010] According to one aspect of an embodiment of the present application, generating multiple subtasks to be processed based on the resource scheduling strategy and distributing the multiple subtasks to be processed to different departments corresponding to the enterprise include: generating multiple subtasks to be processed based on the resource type in the resource scheduling strategy; determining target processing departments based on the resource types corresponding to each of the multiple subtasks to be processed, the target processing departments including different departments corresponding to the enterprise; and distributing the multiple subtasks to be processed to different departments corresponding to the enterprise.

[0011] According to one aspect of an embodiment of the present application, an enterprise digital management system based on artificial intelligence is provided, the system comprising: an acquisition module for acquiring multimodal data corresponding to a task to be processed, the multimodal data comprising image data and text data; a determination module for determining the target resources required for the task to be processed based on the image data and the text data; a policy module for acquiring the real-time resource status of the enterprise, and determining the resource scheduling strategy of the enterprise based on the real-time resource information and the target resources; a distribution module for generating a plurality of subtasks to be processed based on the resource scheduling strategy, and distributing the plurality of subtasks to be processed to different departments corresponding to the enterprise.

[0012] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device implements the aforementioned artificial intelligence-based enterprise digital management method.

[0013] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which computer-readable instructions are stored. When the computer-readable instructions are executed by a processor of a computer, the computer executes the enterprise digital management method based on artificial intelligence as described above.

[0014] According to one aspect of an embodiment of the present application, a computer program product is also provided, including a computer program, which, when executed by a processor, implements the steps in the enterprise digital management method based on artificial intelligence as described above.

[0015] In the technical solution provided in the embodiments of the present application, the semantic information of the tasks to be processed can be fully captured, so as to accurately determine the target resources required for the tasks and avoid resource misjudgment due to missing information; combined with the dynamic perception of the real-time resource status of the enterprise, the system can generate the most optimized resource scheduling strategy based on the intelligent matching of current resource availability and target resource requirements, significantly improve resource utilization and reduce task waiting time; finally, through sub-task decomposition and cross-departmental collaborative distribution, complex tasks are efficiently decomposed into sub-tasks that can be executed in parallel, realizing seamless collaboration among multiple departments, thereby greatly shortening the overall task processing cycle and improving the enterprise's operational efficiency and response speed.

[0016] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, serving to explain the principles of the present application. It is obvious that the drawings described below are merely some embodiments of the present application, and a person of ordinary skill in the art can derive other drawings based on these drawings without inventive effort. In the drawings: Figure 1 This is a schematic diagram of an implementation environment for artificial intelligence-based enterprise digital management in enterprise management, shown in an exemplary embodiment of the present application; Figure 2 This is a flowchart of an enterprise digital management method based on artificial intelligence, as shown in an exemplary embodiment of the present application; Figure 3is a flowchart of an enterprise digital management method based on artificial intelligence, shown in another exemplary embodiment of the present application; Figure 4 is a flowchart of an enterprise digital management method based on artificial intelligence, shown in another exemplary embodiment of the present application; Figure 5 is a flowchart of an enterprise digital management method based on artificial intelligence, shown in another exemplary embodiment of the present application; Figure 6 is a flowchart of an enterprise digital management method based on artificial intelligence, shown in another exemplary embodiment of the present application; Figure 7 is a flowchart of an enterprise digital management method based on artificial intelligence, shown in another exemplary embodiment of the present application; Figure 8 is a flowchart of an enterprise digital management method based on artificial intelligence, shown in another exemplary embodiment of the present application; Figure 9 This is a block diagram of an enterprise digital management system based on artificial intelligence, shown in an exemplary embodiment of the present application; Figure 10 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0018] Exemplary embodiments will be described in detail herein, examples of which are illustrated in the accompanying drawings. When the following description refers to the drawings, identical numerals in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of systems and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0019] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0020] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0021] In this application, "plurality" refers to two or more. "And / or" describes the relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0022] First and foremost, it's important to note that with the acceleration of enterprise digital transformation, task management has become a core engine driving organizational efficiency and resource collaboration. Globally, over 76% of enterprises have incorporated digital task management into their strategic planning. This is particularly true in complex scenarios like equipment maintenance, contract approval, and safety inspections. Technological convergence is driving a paradigm shift in task management. However, existing enterprise management systems only support structured text input and are unable to automatically parse unstructured data such as images (e.g., equipment photos, contract scans) or videos (e.g., operational process monitoring). Multimodal data (e.g., dashboard values ​​and handwritten annotations in images) lacks semantic alignment, leading to low information utilization, inefficient task processing, and difficulties in collaborative task completion across departments.

[0023] Figure 1 This is a schematic diagram of the implementation environment of enterprise digital management based on artificial intelligence in enterprise management shown in an exemplary embodiment of this application. Figure 1 As shown, when a pending task is triggered, the enterprise management system 110 receives data of different modes sent by multiple preset data interfaces 120, and the data of different modes include image data and text data. Then, the enterprise management system 110 can determine the target resources required for the pending task based on the image data and text data, and obtain the real-time resource status within the enterprise, wherein the real-time resource status includes the real-time resource usage status and the remaining idle resource status of the enterprise. Then, the enterprise management system 110 can determine the resource scheduling strategy of the enterprise based on the real-time resource status and the target resources required for the pending task. Furthermore, the enterprise management system 110 can generate multiple pending subtasks based on the resource scheduling status, and then distribute the multiple pending subtasks to the corresponding different departments of the enterprise. In this way, efficient processing of enterprise tasks is achieved.

[0024] in, Figure 1 The enterprise server 110 shown can be any terminal device that supports enterprise data processing, such as a smart phone, a car computer, a tablet computer, a laptop computer, or a wearable device, but is not limited thereto.

[0025] With the acceleration of enterprise digital transformation, task management has become a core component of improving organizational efficiency. Traditional task management systems rely on manual input or single-modal data (such as plain text input), manual monitoring, and static rules. This leads to bottlenecks such as delayed responses, missed tasks, and high labor costs. Existing tools are unable to intelligently identify the emotional tendencies and priorities in user messages. When text, images, and rich media are mixed, they struggle to deeply understand the complex semantic structure. Furthermore, they lack flexibility in task allocation, resulting in inefficient management.

[0026] For example, in equipment maintenance scenarios, due to the complexity of equipment and the diversity of faults, traditional empirical judgment methods are difficult to quickly and accurately locate the cause of the fault, especially when the fault information is vague or cross-impacts, problem location is even more difficult. In addition, due to the lack of intelligent management, operation and maintenance resources are difficult to allocate reasonably, resulting in some equipment being over-maintained and some equipment being under-maintained, affecting overall operation and maintenance efficiency.

[0027] The problems pointed out above are generally applicable in general enterprise management scenarios. It can be seen that existing enterprise management relies on manual input or single-modal data (such as plain text input), and still relies on manual monitoring and static rules, with bottlenecks such as delayed response, missed tasks, and high labor costs. In order to solve these problems, the embodiments of the present application respectively propose an enterprise digital management method based on artificial intelligence, an enterprise digital management system based on artificial intelligence, an electronic device, a computer-readable storage medium, and a computer program product. These embodiments will be described in detail below.

[0028] See also Figure 2 , Figure 2 This is a flowchart of an enterprise digital management method based on artificial intelligence, as shown in an exemplary embodiment of the present application. This method can be applied to Figure 1 The implementation environment shown is specifically implemented by the enterprise management system 110 in the implementation environment. It should be understood that the method can also be applied to other exemplary implementation environments and be specifically implemented by devices in other implementation environments. This embodiment does not limit the implementation environment to which the method is applicable.

[0029] like Figure 2 As shown, in an exemplary embodiment, the enterprise digital management method based on artificial intelligence includes at least steps S210 to S240, which are described in detail as follows: Step S210: Acquire multimodal data corresponding to the task to be processed, where the multimodal data includes image data and text data.

[0030] For example, users can submit multimodal data such as text, images, tables, voice, or short videos corresponding to the pending task through a preset data interface (such as a group message). In this embodiment, the multimodal data includes image data and text data as an example. Specifically, users can use a preset data interface (such as enterprise WeChat / DingTalk group messages, API gateway) to support the parsing of heterogeneous data formats, including: structured / unstructured text such as equipment failure descriptions and contract terms; images such as equipment photos, dashboard screenshots, handwritten documents, etc.; tables (such as Excel), voice (such as maintenance recordings), short videos (such as operation process records), etc.

[0031] Step S220: determining target resources required for the task to be processed based on the image data and the text data.

[0032] For example, deep learning models (such as YOLOv8 and Faster R-CNN) can be used to locate and classify devices, components, and fault features in images, and identify device models (such as "ABC-1000 water pump").

[0033] Detect fault points (e.g., "motor overheating," "pipeline leak"). Alternatively, use optical character recognition (OCR) to extract textual information from images (e.g., instrument panel readings, equipment numbers). Combine image classification results to generate structured descriptions (e.g., "The ABC-1000 water pump pressure gauge reading is 0.5 MPa, below the normal range"). Use pre-trained language models to perform word segmentation, entity recognition, and intent classification on text. For example, extract key information from task descriptions (e.g., "motor bearings need to be replaced," "two electricians are required") and identify task priorities (e.g., "urgent," "routine"). Match the parsed textual information with resources in the enterprise resource directory (ERP system) to generate a list of candidate resources.

[0034] Optionally, in some feasible embodiments, speech data can be converted to text using corresponding technologies, and resource requirements can be further extracted. A cross-modal model can be used to map image and text features into the same semantic space and calculate similarity. For example, if "motor overheating" is detected in an image and "device temperature abnormality" is mentioned in the text, semantic alignment can confirm that they describe the same fault. By combining historical task data with the knowledge graph, implicit resource requirements can be supplemented. For example, if historical data shows that "motor overheating" typically requires a "cooling fan," "cooling fan" can be automatically added to the resource list. Based on the parsing results, target resources can be categorized into the following categories: human resources: role (e.g., electrician, engineer), skill requirements, and quantity; material resources: spare parts (e.g., motor bearings, coolant), tools (e.g., multimeter, wrench); and time resources: task start time, expected completion time, etc.

[0035] Step S230: obtaining the real-time resource status of the enterprise, and determining the resource scheduling strategy of the enterprise based on the real-time resource information and the target resources.

[0036] For example, an enterprise resource planning system (ERP / EAM) uses APIs to access real-time data on human resources, material inventory, and equipment status. This includes: Internet of Things (IoT) devices: sensors and smart tools report real-time equipment operating status (e.g., vibration and temperature); personnel location systems: GPS or RFID captures employees' real-time location and task status (idle / busy); human resources: roles, skills, current tasks, and available time; material resources: inventory quantity, storage location, and expiration date (e.g., spare parts); and equipment resources: operating status (normal / faulty) and load factor (e.g., generator power). Scheduled polling can be set, for example, synchronizing ERP system data every five minutes, with IoT devices proactively reporting abnormal status via the MQTT protocol. Comparing target resource requirements (e.g., "two electricians needed, one motor bearing") with real-time resource status can identify gap types, such as quantity gaps (e.g., insufficient available resources (e.g., only one electrician available); skills gaps (e.g., lack of personnel with specific skills (e.g., "high-voltage electrician")); and location gaps (e.g., resources distributed elsewhere (e.g., spare parts in a remote warehouse). Then, the resource allocation order is dynamically adjusted based on the urgency of the task (e.g., "Emergency Repair" > "Routine Inspection"). Basic scheduling strategies include: proximity allocation, which prioritizes resources in the same region (e.g., electricians in the same workshop); and skill matching, which selects personnel with a perfect match for the task requirements (e.g., "motor repair" skills). For more complex tasks, advanced scheduling strategies can be employed, such as task splitting, which breaks down large tasks into multiple subtasks and allocates resources in stages. For example, if a motor bearing needs to be replaced and tested, Electrician A can be assigned first to complete the replacement, followed by Electrician B to perform the test. If resource borrowing is involved, such as the need to allocate resources across departments or regions, an approval process is triggered. For example, a spare part can be allocated from another factory warehouse, generating a logistics order number. If internal resources are insufficient, a purchase requisition is automatically generated. Resources are pre-locked for high-priority tasks, and task progress is monitored in real time. Resources are reallocated if timeouts occur. This generates a resource scheduling strategy tailored to the enterprise, including task allocation: a detailed list of personnel, equipment, and spare parts. For example, time window: task start / end time, resource occupancy period; exception handling: backup resource plan (such as "if electrician A cannot come to work, activate electrician C"), and push the same specifications to the corresponding employees on multiple terminals. For example, task cards containing task description, resource list, and completion deadline are pushed to relevant personnel through enterprise IM tools (such as enterprise WeChat) and mobile APP.

[0037] Step S240: Generate multiple subtasks to be processed based on the resource scheduling strategy, and distribute the multiple subtasks to be processed to different departments corresponding to the enterprise.

[0038] For example, the overall resource scheduling strategy is broken down into executable subtasks, each corresponding to independent work steps, resource requirements, and responsible departments. The timing dependencies between subtasks must be clearly defined (e.g., "spare parts collection" must be completed before "repair"). This is modeled using a workflow engine. Subtask cards containing task details and associated documents (e.g., repair manuals) are pushed via WeChat for Work and DingTalk. Subtasks are then synchronized to the department's task dashboard and automatically assigned to individual schedules. Each subtask card can include the following information: Task ID: a globally unique identifier (e.g., "SUBTASK-20251001-001"); Task Description: a brief description of the task (e.g., "Collect SKF-6204 motor bearings and store them in the tool cabinet in the repair shop"); Resource Requirements: a detailed list of required personnel, tools, and spare parts; Time Requirements: a planned start time, deadline, and timeout threshold; Related Tasks: the IDs of predecessor tasks (e.g., "Waiting for SUBTASK-20251001-002 to complete"). Personnel from each department update task status (e.g., "In Progress," "Completed") via a mobile app or PC system, and the system automatically records operation logs to support traceability (e.g., "Who picked up the spare part and when?"). If a department is unable to complete a task on time, the following process is automatically triggered: a resource conflict alert is generated and the dispatch center is notified. The dispatch center then re-matches resources (e.g., relocating spare parts from other warehouses) and updates the scheduled time for the affected subtasks. If a predecessor task is delayed, the time for subsequent tasks is automatically adjusted, and the relevant departments are notified. The progress of multi-department tasks is displayed via Gantt charts or Kanban boards, and synchronized with management in real time.

[0039] In this embodiment, by integrating the multimodal data of the tasks to be processed (including image data and text data) to accurately determine the target resource requirements, and dynamically generating resource scheduling strategies based on the real-time resource status of the enterprise, the intelligence and collaborative efficiency of task processing can be significantly improved. By cross-validation and comprehensive analysis of image data and text data, the type and quantity of resources required for the task can be more comprehensively and accurately identified, thereby avoiding resource allocation deviations caused by insufficient information; and dynamically generating resource scheduling strategies based on the real-time resource status of the enterprise can maximize the utilization of existing resources and respond to resource shortage risks in a timely manner; further, the resource scheduling strategies are decomposed into multiple sub-tasks to be processed and distributed to corresponding departments, realizing refined division of labor and cross-departmental collaboration in task execution, ensuring clear responsibilities of each department and seamless process connection, and ultimately shortening the task processing cycle, reducing resource waste, and improving the enterprise's ability to respond to complex tasks and overall operational efficiency.

[0040] Further, based on the above embodiment, please refer to Figure 3 In one of the exemplary embodiments provided in this application, the specific implementation process of the above-mentioned enterprise digital management method based on artificial intelligence may further include steps S310 to S330, which are described in detail as follows: Step S310, determining object information and content information corresponding to the image data, and determining semantic features corresponding to the text data; Step S320: cross-modally fuse the object information, content information, and semantic features to obtain a structured semantic label; Step S330: determining the target resource corresponding to the task to be processed based on the structured semantic tag.

[0041] For example, taking the equipment maintenance scenario as an example, deep learning models (such as YOLOv8, Mask R-CNN) can be used to detect targets in images, identify objects such as equipment, components, fault features, etc., and then use the classification model to perform fine-grained classification of the detected objects (such as "ABC-1000 water pump" and "motor bearing"). Object information can include: equipment type (such as "motor"), component name (such as "bearing"), fault type (such as "overheating" and "wear"); content information can include: spatial position (such as "bearing is located on the right side of the motor", visual features (such as "bearing surface has cracks" and "dashboard reading is 0.5MPa"), and text content can be (such as "Model: SKF-6204" and "Maximum pressure: 1.0MPa"). On the other hand, pre-trained language models can also be used to perform word segmentation, named entity recognition, intent classification, etc. on text. For example, through keyword extraction, the core requirements in the task description (such as "replace bearing" and "electrician required") can be identified, and then the text data can be obtained. According to the corresponding semantic features, semantic features include: task type (such as "maintenance" and "inspection"); resource requirements (such as "electrician" and "motor bearing"); and constraints (such as "urgent task" and "must be completed within 2 hours"). Feature alignment and fusion techniques can then be used: Cross-modal modeling: Uses corresponding techniques to map image features (such as the detected "motor bearing") and text features (such as "replace bearing") into the same semantic space and calculate similarity. Knowledge graph enhancement: Associates the fused features with the enterprise knowledge graph (such as the equipment fault database and maintenance manual) to supplement implicit information (such as "worn motor bearings typically require replacement, model SKF-6204").

[0042] In addition, in some feasible embodiments, structured semantic tags can be matched with resources in the enterprise resource directory (ERP system) to filter qualified resources (such as the number of "SKF-6204 bearings" in inventory ≥ 1), and combined with historical task data (such as "motor bearing replacement usually requires the collaboration of 2 people"), implicit resource requirements can be supplemented (such as "an additional person is required to assist in lifting").

[0043] In some embodiments of the present application, by comprehensively analyzing the object information and content information of image data, extracting the semantic features of text data, and using cross-modal fusion technology to deeply integrate visual and text information to generate structured semantic labels, the accuracy and efficiency of task resource matching can be significantly improved. Cross-modal fusion breaks the information limitations of a single modality, enabling the system to fully understand the task scenario and thus more accurately infer the required resources; it avoids the complex multi-modal data one-by-one comparison process in traditional methods, significantly shortens the resource scheduling decision-making time, and reduces the waste of resources caused by information understanding deviations, ultimately achieving intelligent upgrades in task processing and optimization of operating costs.

[0044] Further, based on the above embodiment, please refer to Figure 4 In one of the exemplary embodiments provided in this application, the specific implementation process of determining the enterprise's resource scheduling strategy based on the real-time resource information and the target resources may further include steps S410 to S430, which are described in detail as follows: Step S410, determining the task attributes corresponding to the task to be processed; Step S420, determining the processing priority corresponding to the task to be processed based on the task attributes, target resources and real-time resource information; Step S430: Determine the enterprise's resource scheduling strategy based on the processing priority.

[0045] For example, based on the task description or historical classification, it is clear that the task belongs to the type of "maintenance", "inspection", "installation", "debugging", etc. For example, if the task description contains "replacement of bearings", it is classified as "maintenance". Determine the scope of the impact of the task, such as "single equipment shutdown", "production line interruption", "entire workshop shutdown", etc. For example, if the main water pump fails and the production line stops, the scope of impact is "production line", and then combine the task description with business rules to classify the urgency into "urgent", "high", "medium" and "low". For example, if the equipment shutdown causes production interruption, it is marked as "urgent". It is also necessary to analyze the degree of dependence of the task on resources such as manpower, equipment, and spare parts. For example, a task that requires a specific model of bearing is marked as "high dependence". In addition, it is also possible to determine whether the task is a repetitive task (such as a monthly inspection) or a new task that appears for the first time. For example, if a task description matches a historical inspection task, it will be marked as a "regular task." The task's priority is then determined based on the following dimensions, combined with task attributes, target resources, and real-time resource information: Urgency: This directly impacts the order in which tasks are processed, with urgent tasks prioritized; Scope of Impact: Tasks with a wide impact have a higher priority (e.g., a workshop shutdown takes precedence over a single piece of equipment failure); Resource Availability: If target resources (e.g., spare parts) are in sufficient inventory, the priority can be lowered appropriately. If critical resources (e.g., a specific type of bearing) are out of stock, the priority needs to be increased to coordinate resources. For example, maintenance tasks typically have a higher priority than inspection tasks, while tasks involving safety or compliance (e.g., annual equipment inspections) have a higher priority.

[0046] In addition, priorities can be dynamically adjusted based on real-time resource status (such as spare parts arrival and staff availability). For example, if bearings arrive, a task originally marked as "medium" priority can be upgraded to "high." Coordinate resources from multiple departments based on task attributes (such as the need for electrician + mechanic collaboration). For example, motor repair tasks require collaboration between the electrical maintenance team and the maintenance team, prioritizing the availability of resources for both parties. Monitor task progress and resource status in real time. If high-priority tasks are delayed, reallocate resources or adjust subsequent task plans. For example, if an "urgent repair" task is delayed due to non-arrival of spare parts, prioritize coordination of allocations from other warehouses or enable alternative solutions.

[0047] In some embodiments of this application, by clarifying the attributes of pending tasks and dynamically determining task processing priorities based on these attributes, target resource requirements, and the enterprise's real-time resource status, the scientific nature and responsiveness of enterprise resource scheduling can be significantly improved. Furthermore, resource scheduling strategies generated based on priority ranking can ensure that high-priority tasks receive priority resource guarantees while also avoiding local resource overloads through global resource optimization (such as cross-departmental collaboration and dynamic resource allocation), achieving a dual improvement in task processing efficiency and resource utilization. This approach is particularly suitable for enterprise-level resource management in complex multi-tasking parallel scenarios.

[0048] Further, based on the above embodiment, please refer to Figure 5 In one of the exemplary embodiments provided in this application, the specific implementation process of the above-mentioned enterprise digital management method based on artificial intelligence may further include steps S510 to S530, which are described in detail as follows: Step S510: Real-time monitoring of multimodal data sent by the enterprise's preset data interface and newly added task processing rules; Step S520: updating the pending task set corresponding to the enterprise based on the multimodal data and the newly added task processing rules; Step S530 : updating the processing priority of the pending task based on the pending task set, so as to adjust the resource scheduling strategy of the enterprise based on the updated processing priority.

[0049] For example, the enterprise continuously monitors multimodal data sent through various pre-set data interfaces (such as IoT device interfaces, manual reporting interfaces, and third-party system docking interfaces). Multimodal data includes: image data, such as on-site equipment photos and dashboard reading screenshots; text data (voice data), such as fault descriptions, operation logs, and manually reported task requirements; and sensor data, such as equipment operating parameters (such as temperature, pressure, and vibration frequency). Different collection frequencies can be set based on data type and business needs (such as real-time sensor data collection and on-demand triggering of image data). Furthermore, the enterprise management system can also obtain new or modified task processing rules in real time. These rules may come from: management decisions, such as temporary adjustments to the processing procedures for emergency tasks; system configuration, such as the priority calculation weight for adding a certain type of task; and external input, such as rule changes caused by urgent customer requests or supplier notifications. The real-time collected multimodal data is then subjected to correlation analysis (such as matching fault characteristics in image data with descriptions in text data).

[0050] Optionally, the parsed data is converted into specific tasks to be processed based on pre-set rules. For example, an image showing a "motor bearing crack" and a text description describing "abnormal equipment noise" generates a "motor bearing repair" task; sensor data showing "pressure exceeding the limit" and a manual report of a "valve failure" generates a "valve emergency repair" task. Newly identified tasks are added to the set of pending tasks, and the status of existing tasks is updated based on real-time data (for example, a task is marked as "executable" due to the arrival of a spare part). If multiple tasks are associated with the same equipment, they are merged into a comprehensive task (for example, "motor bearing replacement + lubrication"). If a task is too complex, it is split into subtasks (for example, "production line maintenance" can be split into "motor inspection" and "drive belt replacement"). Rule types include task classification rules, priority calculation rules, and resource allocation constraints. Furthermore, priority calculation factors can be adjusted based on newly acquired task processing rules (for example, adding a "customer importance" dimension or modifying the "urgency" weight). Then, the system re-evaluates task priorities across multiple dimensions. For example, task urgency is re-evaluated based on real-time data (e.g., if sensor data anomalies continue to worsen, the urgency is increased); impact is updated based on multimodal data correlation analysis (e.g., if a fault is found to affect multiple related devices, the impact is expanded); and resource dependencies are adjusted based on real-time resource information (e.g., if a spare part inventory decreases, the resource dependency of related tasks increases). Finally, based on the adjusted priority calculation results, the set of pending tasks is reordered. The priority change is notified to relevant personnel or systems (e.g., a notification to the dispatcher stating that a task's priority has been increased from 'medium' to 'high'). Critical resources are allocated first (e.g., assigning the only spare part in inventory to a high-priority task). This is especially important if cross-departmental resource coordination is involved (e.g., transferring personnel or tools from other workshops).

[0051] In some embodiments of the present application, by clarifying the task attributes of the tasks to be processed and dynamically determining the task processing priority based on the task attributes, target resource requirements and the real-time resource status of the enterprise (such as resource availability, load balancing, and bottleneck links), the scientific nature and response speed of the enterprise resource scheduling can be significantly improved. Ultimately, the resource scheduling strategy generated based on priority sorting can not only ensure that high-priority tasks obtain resource guarantees first, but also avoid local resource overload through global resource optimization (such as cross-departmental collaboration and dynamic resource allocation), thereby achieving a dual improvement in task processing efficiency and resource utilization, and is particularly suitable for enterprise-level resource management in complex multi-task parallel scenarios.

[0052] Further, based on the above embodiment, please refer to Figure 6 In one of the exemplary embodiments provided in this application, the specific implementation process of the above-mentioned enterprise digital management method based on artificial intelligence may further include steps S610 to S640, which are described in detail as follows: Step S610: Obtain the enterprise's historical resource scheduling strategy and preset task processing rules; Step S620, determining the feasibility of resource scheduling based on historical resource scheduling strategies and preset task processing rules; Step S630: If the degree of execution is greater than the preset degree of execution threshold, the resource scheduling strategy is executed; Step S640: Perform model training based on resource scheduling strategies and historical resource scheduling strategies, and obtain a knowledge graph based on the trained model to determine the enterprise's resource scheduling strategy through the knowledge graph.

[0053] For example, historical resource scheduling records are extracted from the enterprise's operations and maintenance system, ERP system, or historical archives, including data such as task allocation logic, resource allocation plans, task completion times, and resource utilization. For example, a record may indicate that "Over the past three months, emergency maintenance tasks prioritized spare part A in inventory, and the average response time was reduced by 20%." The company's existing task classification standards, priority calculation rules (e.g., "Downtime task priority = urgency × 0.6 + impact area × 0.4"), and resource allocation constraints (e.g., "A maximum of three maintenance personnel may be assigned to a single task") are then compiled to form a rule base. Historical scheduling policies are then compared with pre-set rules to assess their coverage and conflicts. For example, if the historical policy's requirement that "cross-departmental collaborative tasks require resources to be requested 24 hours in advance" conflicts with the pre-set rule that "urgent tasks can be requested immediately," the policy is deemed inexecutable. The historical policy is then verified to ensure that it still meets resource availability requirements, taking into account current resource status (e.g., spare part inventory and staff schedules). For example, if the historical policy relies on a certain spare part model, but the current inventory is zero, the policy is inexecutable. A scoring system for executable performance is established (e.g., "perfect match = 100 points, partial conflict = 60 points, severe conflict = 0 points"). If the score exceeds a preset threshold (e.g., 80 points), the policy is deemed executable. Policies that meet the executable performance criteria are directly imported into the scheduling system for execution, with real-time monitoring of task progress and resource consumption. For example, a policy that prioritizes allocating inventory spare parts to high-priority tasks can be applied to the current task set.

[0054] Furthermore, in some feasible embodiments, historical scheduling strategies, preset rules, task attributes (e.g., urgency, resource dependency), and execution performance data are integrated into a training set to extract key features (e.g., the correlation between "task type, resource allocation, and completion time"). A supervised learning algorithm (e.g., decision tree, random forest) or reinforcement learning model is used to train a resource scheduling optimization model. For example, the decision-making pattern of "prioritizing high-resource dependency tasks when spare parts are sufficient" can be learned from historical data. The trained model is converted into a structured knowledge graph, representing scheduling knowledge in the form of entities, relationships, and attributes. For example, entities include: task, resource, rule; relationships: "task, dependency, resource," "rule, constraint, task, priority," task urgency (high / medium / low), and resource availability (sufficient / shortage). Then, based on the knowledge graph, real-time reasoning is performed on new tasks and resource status, dynamically generating the optimal scheduling strategy. For example, if a new task is "high urgency + high resource dependency" and spare parts are currently sufficient, the graph recommends "immediately allocate inventory resources and initiate emergency processes." The newly implemented scheduling strategy and its results are then fed back to the knowledge graph. Through incremental learning, the model is continuously updated, forming a closed loop of "execution-feedback-optimization." For example, if a strategy causes equipment maintenance delays during the rainy season, the graph automatically adjusts the priority weight of humidity-sensitive tasks.

[0055] In one embodiment provided in the present application, by real-time monitoring of multimodal data transmitted by the enterprise's preset data interface and dynamically added task processing rules, it is possible to timely capture sudden demands and rule changes in enterprise operations, and update the set of pending tasks in real time based on these dynamic information, thereby dynamically adjusting the task processing priority; this task set and priority update mechanism driven by real-time data and rules enables the enterprise resource scheduling strategy to quickly respond to external environmental changes (such as sudden failures, urgent customer needs) and internal rule adjustments (such as temporary resource restrictions, new policy requirements), ensuring that resources are always tilted towards high-value tasks, avoiding task backlogs or resource mismatches caused by static scheduling, and ultimately improving the enterprise's adaptability to complex and changing business environments and overall operational resilience.

[0056] Further, based on the above embodiment, please refer to Figure 7 In one of the exemplary embodiments provided in this application, the specific implementation process of the above-mentioned enterprise digital management method based on artificial intelligence may further include steps S710 to S730, which are described in detail as follows: Step S710: parsing the multimodal data to determine the data to be encrypted in the multimodal data based on the parsing result; Step S720: encrypting the data to be encrypted based on differential privacy and federated learning to obtain encrypted multimodal data; Step S730: Determine the target resource corresponding to the task to be processed based on the encrypted multimodal data.

[0057] For example, the received multimodal data is first classified to identify its type (e.g., text, image, sensor signal, etc.). For example, equipment operation logs are text data, on-site photos are image data, and temperature and pressure readings are sensor data. Natural language processing techniques are used to extract key information from text data, such as fault descriptions ("abnormal motor noise," "bearing wear") and task requirements ("emergency repair," "regular inspections"). For image data, computer vision techniques can be used to analyze image content and identify equipment fault characteristics (e.g., cracks, leaks) or task scenarios (e.g., maintenance personnel operations, equipment installation locations). Sensor data analyzes abnormal patterns in time series data (e.g., sudden temperature increases, abnormal vibration frequencies), and combines threshold rules to determine equipment status (e.g., "motor temperature exceeds safety threshold"). Then, based on data sensitivity and business rules, data requiring encryption is marked. For example, detailed descriptions of equipment failures in text (e.g., "Bearing model XYZ has a design defect") may involve trade secrets and require encryption. Facial or license plate information in images (e.g., facial features of maintenance personnel) may involve privacy and require desensitization. Controllable noise is added to marked sensitive text or numerical data (e.g., equipment failure rate, repair duration) to ensure that individual pieces of data cannot be reverse-engineered. For example, adding noise to "Number of bearing failures = 5" becomes "Number of bearing failures = 5 ± 1." The privacy budget (ε) is then adjusted to balance data availability and privacy protection. A smaller ε value results in stronger privacy protection, but may also increase data distortion.

[0058] Optionally, if multimodal data is dispersed across departments or devices (e.g., sensor data in each workshop is stored independently), a federated learning framework can be used for encrypted aggregation. For example, each data source can locally train a model (e.g., a neural network for predicting equipment failure) based on encrypted data, sharing only the model parameters, not the original data. Secure aggregation: Using secure multi-party computation or homomorphic encryption, the model parameters from each party are aggregated to generate a global model, preventing data leakage. The differentially private text / numeric data is combined with the encrypted model parameters generated by federated learning to form an encrypted multimodal dataset. Within an authorized environment (such as an enterprise intranet), the encrypted data is decrypted to restore key features. For example, differential privacy inversion or federated learning global model inference can yield a "high bearing failure risk level." If partial data confidentiality is still required, a trusted execution environment or hardware security module can be used for restricted decryption. The decrypted features are used to infer the task type. For example, if "bearing failure risk level = high" and "temperature surge" occur, an "urgent repair task" is inferred. Resource requirements are calculated by combining historical task data with pre-set rules. For example, an "urgent repair task" requires two senior maintenance technicians, one set of bearing spare parts, and one specialized tool. If the data involves cross-departmental collaboration (e.g., images showing a fault is located in a hazardous area), additional safety protection resources (e.g., protective clothing, safety monitors) should be added.

[0059] In some embodiments of the present application, by parsing multimodal data and accurately identifying the sensitive information to be encrypted therein, combined with differential privacy technology and a federated learning framework, the value of multimodal data can be fully utilized to determine the target resources for the tasks to be processed while protecting data privacy and security. This not only solves the privacy leakage risk in traditional data sharing, but also achieves a dual improvement in data security and business efficiency through the collaboration of cross-modal data fusion and privacy protection technology. It is particularly suitable for enterprise scenarios with high data sensitivity requirements and cross-departmental collaboration, and optimizes the accuracy and real-time nature of resource scheduling decisions while ensuring compliance.

[0060] Further, based on the above embodiment, please refer to Figure 8 In one of the exemplary embodiments provided in this application, the specific implementation process of generating multiple pending subtasks based on the resource scheduling strategy and distributing the multiple pending subtasks to different departments of the enterprise may also include steps S810 to S830, which are described in detail as follows: Step S810: Generate multiple to-be-processed subtasks based on the resource type in the resource scheduling policy; Step S820: determining a target processing department based on the resource types corresponding to the plurality of to-be-processed subtasks, wherein the target processing department includes different departments corresponding to the enterprise; Step S830: distribute the multiple to-be-processed subtasks to the corresponding different departments of the enterprise.

[0061] For example, the required resource types and their associated tasks are extracted from the resource scheduling policy. For example, the policy requires "assigning two electrical engineers, one vibration analyzer, and 5 kg of lubricant" to complete the "motor bearing repair task." For example, based on resource type, this can be categorized as follows: human resources: electrical engineers, maintenance technicians; equipment resources: vibration analyzers, cranes; and material resources: lubricant, bearing spare parts. The resource scheduling policy then breaks down the task into multiple subtasks based on resource type, each corresponding to a specific resource type. For example: Subtask 1: "Assign two electrical engineers to conduct on-site inspections" (human resources); Subtask 2: "Transfer vibration analyzers to the repair site" (equipment resources); Subtask 3: "Prepare and inspect 5 kg of lubricant" (material resources). The execution order or dependencies between subtasks must be clearly defined. For example, Subtask 2 (allocating equipment) must be executed after Subtask 1 (assigning personnel) to avoid idle personnel; Subtask 3 (preparing materials) must be completed first, otherwise the repair cannot begin. Establish a mapping between enterprise departments and resource types. For example, the Electrical Department is responsible for electrical-related resources such as electrical engineers and vibration analyzers; the Equipment Management Department manages general equipment such as cranes and vehicles; and the Warehousing Department stores materials such as lubricants and spare parts. Then, automatically match the target department based on the resource type required for each subtask. For example, subtask 1 (electrical engineer) belongs to the Electrical Department; subtask 2 (vibration analyzer) belongs to the Equipment Management Department (if that department manages the equipment); and subtask 3 (lubricant) belongs to the Warehousing Department.

[0062] Optionally, if a subtask requires collaboration across multiple departments, define the primary and collaborating departments. For example, "Crane allocation" requires the Equipment Management Department (primary responsibility) and the Safety Department (collaborating, responsible for operating permits). Subtasks are automatically pushed to departmental workstations via the enterprise task management system (such as ERP or OA). For example, the Electrical Department Manager receives a to-do item in the system: "Assign two engineers to motor repair." For urgent tasks, instant reminders are triggered via corporate WeChat, SMS, and other channels. Each subtask must include the following information: Task Description: clearly defining the task (e.g., "Test motor bearing vibration levels"); Resource Requirements: specific resource types and quantities (e.g., "One vibration analyzer"); Time Requirements: planned start / finish times (e.g., "Arrive at site before 2:00 PM today"); Dependency Notes: identifying predecessor tasks (e.g., "Lubricant required from Warehousing Department before 1:00 PM").

[0063] Optionally, some feasible implementations include post-distribution tracking and feedback, as well as execution status monitoring. Department heads update subtask statuses (e.g., "In Progress," "Completed"), which the system automatically aggregates to the overall task dashboard. If a department is unable to allocate resources as needed (e.g., "All electrical engineers are on duty"), the system reports the anomaly and triggers an escalation process (e.g., calling in an external contractor). Furthermore, if multiple subtasks compete for the same resources (e.g., a vibration analyzer is occupied by other tasks), the system will reallocate resources or adjust schedules based on task priority. The system will also monitor task backlogs across departments and dynamically adjust distribution strategies. For example, if the electrical department is overloaded, some tasks can be transferred to collaborative suppliers.

[0064] In some embodiments of the present application, by refining the resource types in the resource scheduling strategy into multiple sub-tasks to be processed, and accurately matching the target processing departments based on the resource requirements of the sub-tasks, modularization and professional division of labor in task execution can be achieved, avoiding buck-passing or duplication of work due to unclear responsibilities; further, through a systematic distribution mechanism, sub-tasks are pushed to corresponding departments, which not only improves the transparency and efficiency of cross-departmental collaboration, but also dynamically adjusts task priorities based on the real-time load of the department, thereby ensuring the reasonable allocation of resources while maximizing the use of the professional capabilities and resource reserves of each department, and ultimately forming an efficient closed loop of "strategy decomposition, department collaboration, and dynamic optimization", which significantly shortens the task processing cycle and reduces communication costs, and is particularly suitable for resource-intensive or cross-functional collaboration-intensive enterprise scenarios.

[0065] Figure 9 This is a block diagram of an enterprise digital management system based on artificial intelligence, as shown in an exemplary embodiment of the present application. The system can be applied to Figure 1 The implementation environment shown is specifically configured in the enterprise management system 110. The system can also be applied to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the implementation environment to which the system is applicable.

[0066] like Figure 9 As shown, the exemplary artificial intelligence-based enterprise digital management system includes: an acquisition module 910, used to acquire multimodal data corresponding to the task to be processed, and the multimodal data includes image data and text data; a determination module 920, used to determine the target resources required for the task to be processed based on the image data and text data; a strategy module 930, used to obtain the real-time resource status of the enterprise, and determine the resource scheduling strategy of the enterprise based on the real-time resource information and the target resources; a distribution module 940, used to generate multiple subtasks to be processed based on the resource scheduling strategy, and distribute the multiple subtasks to be processed to different departments corresponding to the enterprise.

[0067] According to one aspect of an embodiment of the present application, the above-mentioned determination module 920 is also used to determine the object information and content information corresponding to the image data, and to determine the semantic features corresponding to the text data; to cross-modally fuse the object information, content information and semantic features to obtain structured semantic labels; and to determine the target resources corresponding to the task to be processed based on the structured semantic labels.

[0068] According to one aspect of an embodiment of the present application, the above-mentioned policy module 930 is also used to determine the task attributes corresponding to the task to be processed; determine the processing priority corresponding to the task to be processed based on the task attributes, target resources and real-time resource information; and determine the enterprise's resource scheduling strategy based on the processing priority.

[0069] According to one aspect of an embodiment of the present application, the above-mentioned policy module 930 is also used to monitor in real time the multimodal data and new task processing rules sent by the enterprise's preset data interface; update the enterprise's corresponding set of pending tasks based on the multimodal data and the new task processing rules; update the processing priority of the pending tasks based on the set of pending tasks, so as to adjust the enterprise's resource scheduling strategy based on the updated processing priority.

[0070] According to one aspect of an embodiment of the present application, the above-mentioned policy module 930 is also used to obtain the enterprise's historical resource scheduling strategy and preset task processing rules; determine the executability of resource scheduling based on the historical resource scheduling strategy and preset task processing rules; if the executability is greater than the preset executability threshold, execute the resource scheduling strategy; perform model training based on the resource scheduling strategy and the historical resource scheduling strategy, and obtain a knowledge graph based on the trained model to determine the enterprise's resource scheduling strategy through the knowledge graph.

[0071] According to one aspect of an embodiment of the present application, the above-mentioned determination module 920 is also used to parse multimodal data to determine the data to be encrypted in the multimodal data based on the parsing results; encrypt the data to be encrypted based on differential privacy and federated learning to obtain encrypted multimodal data; and determine the target resources corresponding to the task to be processed based on the encrypted multimodal data.

[0072] According to one aspect of an embodiment of the present application, the above-mentioned distribution module 940 is also used to generate multiple sub-tasks to be processed based on the resource type in the resource scheduling strategy; determine the target processing department based on the resource type corresponding to each of the multiple sub-tasks to be processed, and the target processing department includes different departments corresponding to the enterprise; and distribute the multiple sub-tasks to be processed to different departments corresponding to the enterprise.

[0073] It should be noted that the enterprise digital management system based on artificial intelligence provided in the above embodiment and the enterprise digital management method based on artificial intelligence provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here. In actual applications, the enterprise digital management system based on artificial intelligence provided in the above embodiment can distribute the above functions to different functional modules as needed, that is, divide the internal structure of the system into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0074] An embodiment of the present application also provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by one or more processors, the electronic device implements the artificial intelligence-based enterprise digital management method provided in the above-mentioned embodiments.

[0075] Figure 10 The following is a schematic diagram showing the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application. Figure 10 The computer system 1000 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0076] like Figure 10 As shown, computer system 1000 includes a central processing unit (CPU) 1001, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, based on programs stored in read-only memory (ROM) 1002 or programs loaded from storage 1008 into random access memory (RAM) 1003. RAM 1003 also stores various programs and data required for system operation. CPU 1001, ROM 1002, and RAM 1003 are interconnected via bus 1004. An input / output (I / O) interface 1005 is also connected to bus 1004.

[0077] The following components are connected to the I / O interface 1005: an input section 1006 including a keyboard, mouse, and the like; an output section 1007 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 1008 including a hard disk; and a communication section 1009 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 1009 performs communication processing via a network such as the Internet. A drive 1010 is also connected to the I / O interface 1005 as needed. Removable media 1011, such as a magnetic disk, optical disk, magneto-optical disk, or semiconductor memory, is installed in the drive 1010 as needed, so that computer programs read from the media can be installed in the storage section 1008 as needed.

[0078] In particular, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the methods illustrated in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication section 1009 and / or installed from removable media 1011. When executed by the central processing unit (CPU) 1001, the computer program performs the various functions defined in the system of the present application.

[0079] It should be noted that the computer-readable medium described in the embodiments of this application may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may, for example, be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media 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), flash memory, optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. This propagated data signal may take a variety of forms, including, but not limited to, an electromagnetic signal, an optical signal, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0080] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0081] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0082] Another aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned artificial intelligence-based enterprise digital management method. The computer-readable storage medium may be included in the electronic device described in the above embodiments, or may exist independently and not be incorporated into the electronic device.

[0083] Another aspect of the present application provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to implement the artificial intelligence-based enterprise digital management method provided in each of the above embodiments.

[0084] The above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. Ordinary technicians in this field can easily make corresponding changes or modifications based on the main ideas and spirit of the present application. Therefore, the scope of protection of the present application shall be based on the scope of protection required by the claims.

Claims

1. An enterprise digital management method based on artificial intelligence, characterized in that: include: Acquire multimodal data corresponding to the task to be processed, wherein the multimodal data includes image data and text data; determining target resources required for the task to be processed based on the image data and the text data; Acquiring the real-time resource status of the enterprise, and determining the resource scheduling strategy of the enterprise based on the real-time resource information and the target resource; A plurality of to-be-processed subtasks are generated based on the resource scheduling strategy, and the plurality of to-be-processed subtasks are distributed to different departments corresponding to the enterprise.

2. The method according to claim 1, wherein The method further comprises: Determining object information and content information corresponding to the image data, and determining semantic features corresponding to the text data; Cross-modally fusing the object information, the content information, and the semantic features to obtain a structured semantic label; A target resource corresponding to the task to be processed is determined based on the structured semantic tag.

3. The method according to claim 1, wherein The determining of the resource scheduling strategy of the enterprise based on the real-time resource information and the target resource includes: Determining the task attributes corresponding to the task to be processed; Determining a processing priority corresponding to the task to be processed based on the task attributes, the target resource, and the real-time resource information; A resource scheduling strategy for the enterprise is determined based on the processing priority.

4. The method according to claim 3, wherein The method further comprises: Real-time monitoring of multimodal data sent by the enterprise's preset data interface and newly added task processing rules; Updating a set of pending tasks corresponding to the enterprise based on the multimodal data and the newly added task processing rules; The processing priorities of the tasks to be processed are updated based on the set of tasks to be processed, so as to adjust the resource scheduling policy of the enterprise based on the updated processing priorities.

5. The method according to any one of claims 1 to 4, characterized in that The method further comprises: Obtain the historical resource scheduling strategies and preset task processing rules of the enterprise; Determining the feasibility of the resource scheduling based on the historical resource scheduling strategy and the preset task processing rules; If the degree of execution is greater than a preset degree of execution threshold, executing the resource scheduling strategy; Model training is performed based on the resource scheduling strategy and the historical resource scheduling strategy, and a knowledge graph is obtained based on the trained model to determine the resource scheduling strategy of the enterprise through the knowledge graph.

6. The method according to any one of claims 1 to 4, characterized in that The method further comprises: parsing the multimodal data to determine the data to be encrypted in the multimodal data based on the parsing result; Encrypting the data to be encrypted based on differential privacy and federated learning to obtain encrypted multimodal data; The target resource corresponding to the task to be processed is determined based on the encrypted multimodal data.

7. The method according to claim 1, wherein The generating of a plurality of to-be-processed subtasks based on the resource scheduling strategy and distributing the plurality of to-be-processed subtasks to different departments corresponding to the enterprise includes: Generate multiple subtasks to be processed based on the resource type in the resource scheduling strategy; Determining a target processing department based on the resource types corresponding to the plurality of to-be-processed subtasks, wherein the target processing department includes different departments corresponding to the enterprise; Distribute the multiple to-be-processed subtasks to different departments corresponding to the enterprise.

8. An enterprise digital management system based on artificial intelligence, characterized by: The system comprises: An acquisition module, configured to acquire multimodal data corresponding to a task to be processed, wherein the multimodal data includes image data and text data; a determination module, configured to determine target resources required for the task to be processed based on the image data and the text data; A policy module, configured to obtain the real-time resource status of the enterprise and determine the resource scheduling policy of the enterprise based on the real-time resource information and the target resource; A distribution module is used to generate a plurality of to-be-processed subtasks based on the resource scheduling strategy, and distribute the plurality of to-be-processed subtasks to different departments corresponding to the enterprise.

Images