An agent-based service management method, device, equipment and medium

Abstract

This invention discloses a business management method, apparatus, device, and medium based on intelligent agents. The business management method includes the following steps: acquiring business operation data of multiple business objects in a target enterprise to be managed; identifying differences in each business operation data based on multiple preset difference dimensions to obtain the difference identification results corresponding to each difference dimension; generating correction schemes for each business object through a large model based on the difference identification results; and coordinating the execution of the correction schemes for each business object through multiple preset business intelligent agents. This method, through the collaborative operation of multiple business intelligent agents, corrects the difference identification results of multiple difference dimensions, achieving accurate location and dynamic response to business anomalies, significantly improving the efficiency of enterprise business management and decision-making quality.

Description

Technical Field

[0001] This invention relates to the fields of enterprise management and artificial intelligence application technology, and in particular to a business management method, apparatus, equipment and medium based on intelligent agents. Background Technology

[0002] In modern enterprise management, there is often a disconnect between strategy formulation and execution. While existing technologies or management models (such as KPIs, OKRs, and BSC) have addressed goal management to some extent, they still face discrepancies and disconnects across the entire business chain, including distorted goal transmission, a black box execution process, delayed feedback cycles, and severe data silos. This leads to the dilution of strategic intent through layers of decomposition, causing the execution results to drift further and further away from the initial goals. Furthermore, existing systems are mostly static data recording tools, lacking dynamic perception and autonomous decision-making capabilities, making it difficult to achieve real-time perception and adaptive alignment of discrepancies during dynamic business evolution. Summary of the Invention

[0003] This invention provides a business management method, apparatus, device, and medium based on intelligent agents to solve the problems of various differences and gaps in existing enterprise operation and management, which make it impossible to achieve real-time perception and adaptive alignment of differences in the dynamic evolution of business.

[0004] In a first aspect, this application provides a business management method based on intelligent agents, comprising the steps of: acquiring business operation data of multiple business objects in a target enterprise to be managed; identifying differences in each of the business operation data based on multiple preset difference dimensions to obtain difference identification results corresponding to each difference dimension; generating correction schemes for each of the business objects through a large model based on the difference identification results; and coordinating the execution of the correction schemes for each of the business objects through multiple preset business intelligent agents.

[0005] Secondly, this application provides a business management device based on intelligent agents, comprising: a data acquisition module for acquiring business operation data of multiple business objects in a target enterprise to be managed; a difference identification module for identifying differences in each of the business operation data based on multiple preset difference dimensions, and obtaining difference identification results corresponding to each of the difference dimensions; a scheme generation module for generating correction schemes for each of the business objects through a large model based on the difference identification results; and a scheme execution module for collaboratively executing the correction schemes for each of the business objects through multiple preset business intelligent agents.

[0006] Thirdly, this application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described agent-based business management method.

[0007] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described agent-based business management method.

[0008] The aforementioned technical solution for business management based on intelligent agents, including a method, apparatus, computer equipment, and storage medium, comprises the following steps: acquiring business operation data of multiple business objects within the target enterprise to be managed; identifying differences in each business operation data based on multiple preset difference dimensions to obtain the difference identification results corresponding to each difference dimension; generating correction schemes for each business object through a large model based on the difference identification results; and collaboratively executing the correction schemes for each business object through multiple preset business intelligent agents. This method, through the collaborative operation of multiple business intelligent agents, corrects the difference identification results across multiple difference dimensions, achieving accurate location and dynamic response to business anomalies, significantly improving the efficiency of enterprise business management and decision-making quality. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart of a business management method based on intelligent agents according to an embodiment of the present invention; Figure 2 This is a specific flowchart of step S2 in a business management method based on intelligent agents according to an embodiment of the present invention; Figure 3 This is a specific flowchart of step S3 in a business management method based on intelligent agents according to an embodiment of the present invention; Figure 4 This is a specific flowchart of step S4 in a business management method based on intelligent agents according to an embodiment of the present invention; Figure 5 This is a schematic diagram of a business management device based on an intelligent agent according to an embodiment of the present invention; Figure 6 This is a schematic diagram of a computer device according to an embodiment of the present invention. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] In one embodiment, such as Figure 1 As shown, a business management method based on intelligent agents is provided, including the following steps: Step S1: Obtain business operation data of multiple business objects in the target enterprise to be managed.

[0013] It should be noted that the target enterprise refers to the enterprise entity that needs to implement intelligent business management, and its business objects cover independently operating business units such as departments, production lines, and project teams; business operation data includes multi-source heterogeneous indicators such as real-time capacity, energy consumption, order fulfillment rate, and inventory turnover rate.

[0014] Before step S1, multiple business objects need to be defined, along with the metric dimensions and data collection rules for each object, to ensure the integrity and real-time nature of the operational data. These metric dimensions include, but are not limited to, capacity, yield, delivery cycle, cost percentage, and customer satisfaction. The data collection rules specify the collection frequency, source system, verification logic, and exception handling mechanism to ensure that the metric data is comparable, reliable, and usable.

[0015] In this embodiment, by defining the indicator dimensions of each business object, a mapping relationship between business objects and indicators can be established, thereby obtaining the business operation data of the corresponding business object in the corresponding indicator dimension; this mapping relationship supports dynamic updates to adapt to organizational structure adjustments and business scenario evolution.

[0016] In some embodiments, this mapping relationship also supports cross-business object indicator linkage analysis. For example, when the yield of a certain production line declines, it automatically triggers backtracking verification of the material qualification rate and process parameters of related suppliers. All collected data is preprocessed by edge computing nodes and then uniformly connected to the enterprise knowledge graph to form a business semantic network with spatiotemporal attributes, providing a structured cognitive foundation for subsequent intelligent agent collaborative decision-making. This semantic network uses "business object-indicator-event-root cause" as the four-tuple skeleton, embedding causal reasoning models and time series pattern recognition algorithms, enabling intelligent agents to autonomously identify abnormal propagation paths. For example, when customer satisfaction drops sharply, the system traces back along the graph to a delayed delivery of a certain batch of products, then links it to excessive vibration of upstream production line equipment and spare parts inventory alarms, ultimately locating the supplier's logistics scheduling disorder.

[0017] Step S2: Based on multiple preset difference dimensions, identify the differences in the operational data of each business and obtain the difference identification results corresponding to each difference dimension.

[0018] It should be noted that the dimensions of difference include information, execution, cognition, and competition. Business operation data includes the node update time of business objects, execution process data, work report data, and business indicator data. Among them, the information dimension focuses on the deviation of data timeliness and completeness, mainly reflected in the information gap caused by poor information flow, resulting in node update delays and missing key fields; the execution dimension focuses on the gap between task implementation and resource matching. Although information is synchronized, there are lags or deviations at the execution level, which are manifested as overdue tasks, resource misallocation, or deviations in standard execution; the cognition dimension reveals the deviation of decision-makers' understanding of business goals and indicator logic. Different depths of understanding of the same goal lead to inaccurate goal decomposition, misjudgment of priorities, or misallocation of resource investment; the competition dimension focuses on the relative gap between business object indicators and industry benchmarks or historical best levels, which is reflected in the competitive gap in horizontal comparison of core indicators such as capacity utilization rate, yield volatility, and on-time delivery rate.

[0019] In this embodiment, by identifying differences in the information, execution, cognition, and competition dimensions, structural breakpoints and root cause clusters in business operations can be accurately located, and corresponding difference identification results can be obtained.

[0020] like Figure 2 As shown, specifically, step S2 includes the following sub-steps: Step S21: Compare the node update time with the preset update execution time to identify delayed nodes and obtain the information difference identification result of the information dimension.

[0021] It should be noted that the node update time is the last data refresh time of the corresponding node of the business object in the knowledge graph, and the preset update execution time is the time benchmark for when the node should complete data synchronization. Delayed nodes refer to business nodes whose actual data update time is later than the preset update execution time. Their delay duration, frequency, and related upstream and downstream nodes are all marked in real time to obtain information difference identification results. The information difference identification results are direct evidence of the existence of information differences in the information dimension, reflecting the degree of obstruction of information flow at the nerve endings of the organization, and can include the distribution heatmap of delayed nodes, average delay duration, high-frequency delay links, and related abnormal event labels, etc.

[0022] In this embodiment, the node update time is compared with the preset update execution time. When the node update time of the corresponding node is later than the update execution time, it is determined that there is an information difference in the corresponding node and it is marked as a delayed node, thereby obtaining the information difference identification result in the information dimension.

[0023] Furthermore, the integrity of the fields of the synchronized data of the delayed node can be verified simultaneously, and the field missing rate of the synchronized data of the node relative to the standard information data can be calculated. If the field missing rate exceeds the missing threshold, it is marked as a data incomplete node, which strengthens the dual judgment dimension of information difference.

[0024] In this application, the accuracy of information difference identification is improved to the millisecond-level response threshold by comparing the node update time with the update execution time at the millisecond level. At the same time, the integrity verification of the fields of the corresponding node's synchronized data is combined to achieve dual verification of time and synchronization status, ensuring that information difference identification is both fast and accurate.

[0025] Step S22: Compare the execution process data with the preset plan node data, identify the progress deviation rate, and obtain the execution deviation identification result of the execution dimension.

[0026] It should be noted that execution process data refers to real-time process data collected throughout the entire TTC (Task to Cash) lifecycle of the business flow, covering key action sequences such as task initiation, resource scheduling, stage duration, and abnormal interruptions. Planned node data is a structured dataset containing the time nodes, resource allocation, and delivery standards to be completed at each stage of the pre-defined process. The schedule deviation rate is a quantified measure of the deviation between the execution process data and the planned node data across three dimensions: time axis, resource quantity, and delivery quality, reflecting the strength of the deviation from the pre-defined path at the business execution level. The execution difference identification result is a quantitative representation of the existence of execution differences in the execution dimensions, reflecting the degree of inaccuracy during business execution.

[0027] In this embodiment, the actual completion time, actual resource consumption, and actual delivery quality score of each task node in the execution process data are normalized and compared with the corresponding time window, resource quota, and quality threshold in the planning node data. The three-dimensional deviation is calculated by weighting to obtain the progress deviation rate. When the progress deviation rate exceeds the preset progress threshold of 5%, it is determined to be an execution deviation node and the root cause analysis engine is triggered. The root cause analysis engine automatically retrieves the system logs, operation traces, and personnel behavior trajectories associated with the node to generate a probability distribution map containing three types of root causes: operation delay, resource mismatch, and quality fluctuation, and obtain the execution deviation identification result of the execution dimension.

[0028] Furthermore, the deviation dimension weights are dynamically allocated based on time axis (40%), resource quantity (35%), and delivery quality (25%) to ensure that deviations on the critical path are identified first. This weight allocation mechanism is derived from regression analysis of historical business data. The time axis has the highest weight, reflecting the sensitivity of the critical path to overall timeliness; resource quantity is the next highest, reflecting the rigid constraints of cost control; although the quality dimension has a low weight, it is subject to a veto clause, meaning that when the delivery quality score is below the threshold, regardless of whether deviations in other dimensions exceed the limits, it is directly determined as an execution deviation node.

[0029] In other embodiments, a dynamic weight adjustment mechanism can be introduced to automatically calibrate the three-dimensional weight coefficients based on real-time business scenarios. For example, in an emergency delivery scenario, the timeline weight can be temporarily increased to 60%, the resource quantity can be reduced to 25%, and the quality can be maintained at 25% but trigger a more stringent threshold judgment. In cost-sensitive projects, the resource quantity weight is conversely strengthened to 50%. All weight changes must be approved by the risk control module and recorded to ensure that the strategy adjustment is compliant and traceable.

[0030] In this application, by comparing and dynamically weighting the execution process data with the planned node data in real time, a closed loop for performance health assessment covering the entire TTC cycle is constructed to obtain the results of poor performance identification, providing project managers with quantifiable, traceable and intervention-friendly decision-making basis.

[0031] Step S23: Match the work report data with the preset strategic target data, identify comprehension biases, and obtain the cognitive difference identification results of the cognitive dimension.

[0032] It should be noted that the work report data is structured collection of employees' daily reports, including descriptions of goal progress, statements of key achievements, and problem feedback text. This data can be semantically analyzed using a large-scale model to extract cognitive characteristics such as employees' understanding of goals (keywords), the mapping relationship between achievements and strategic indicators, and problem attribution tendencies. Strategic goal data is a clearly defined quantitative indicator system within the company's strategic goals (SP / BP), covering dimensions such as revenue growth, market share, customer satisfaction, and innovation investment. This data can also be semantically analyzed using a large-scale model to extract strategic keywords, goal hierarchy relationships, and logical constraints between indicators. Understanding bias refers to significant deviations that occur after semantically matching the cognitive characteristics extracted from the work report with the strategic goal data analysis results, such as keyword misalignment, inverted hierarchy relationships, or broken logical constraints. The cognitive difference identification result is a quantitative representation of the cognitive dimension bias, obtained using semantic matching degree as the core indicator.

[0033] In this embodiment, a large model is used to analyze the semantic matching degree between work report data and strategic goal data. A semantic matching degree is generated by weighting three indicators: keyword co-occurrence frequency, hierarchical path similarity, and logical constraint satisfaction rate. When the semantic matching degree is lower than a preset matching degree threshold, it is determined that there is a comprehension bias, thus obtaining a cognitive difference identification result.

[0034] Furthermore, the matching score threshold is dynamically set based on historical deviation distribution and industry benchmarks. The initial threshold is set at 0.68. When the semantic matching score is below this value for three consecutive periods, a comprehension deviation is determined, and an adaptive threshold adjustment mechanism is automatically triggered, with an adjustment range of 0.02 and an upper limit of 0.75. If the matching score rises above the new threshold for five consecutive periods, the initial setting is restored. This mechanism avoids management disturbances caused by misjudgments while ensuring a sensitive response to systemic cognitive biases.

[0035] In this application, through the deep integration of semantic matching and dynamic threshold mechanism between work report data and strategic goal data, the recognition of cognitive gaps is upgraded from static comparison to a continuously evolving cognitive calibration process, thereby realizing dynamic calibration of cognitive dimensions and spiral improvement of organizational learning capabilities.

[0036] Step S24: Compare the business indicator data with the preset benchmark indicator data, identify the competitiveness index, and obtain the competition difference identification result of the competition dimension.

[0037] It's important to note that business metrics data refers to real-time operational data collected from various business entities within the enterprise, encompassing quantifiable behavioral indicators such as order volume, conversion rate, average order value, and repurchase frequency. Benchmark metrics data refers to pre-set reference data used as a comparison standard, sourced from authoritative industry databases, historical averages, or highly challenging metrics, providing a benchmark for competitive comparisons. The competitiveness index is a standardized score generated by comparing business metrics data with benchmark metrics data, reflecting the relative gap between the enterprise's current operational performance and industry benchmarks or its own ultimate goals. The competitive difference identification result is a quantitative representation of deviations in the competitiveness dimension, using standardized scores as the core indicator to characterize the absolute gap and trend deviation between business performance and benchmark targets.

[0038] In this embodiment, quantifiable behavioral indicators such as order volume, conversion rate, average order value, and repurchase frequency are collected from various business objects of the enterprise as business indicator data. The business indicator data is standardized and compared with the benchmark indicator data to calculate the indicator difference value of each business indicator data relative to the benchmark indicator data. Then, based on the indicator difference value of each business indicator, a competitiveness index is generated according to a preset standardization algorithm. Finally, based on the competitiveness index, a competition difference identification result is generated for the competition dimension. The competition difference identification result includes the competitiveness index value, the absolute gap between business performance and the benchmark target, and trend deviation information.

[0039] Furthermore, the indicator difference value is obtained by substituting the mean and standard deviation of the business indicator data and the benchmark indicator data into the indicator difference value formula, which is as follows: ; in, The difference value of the indicator. The actual values ​​of the business metrics corresponding to the business metric data. This represents the mean of the benchmark indicator corresponding to the benchmark indicator data. This represents the standard deviation of the benchmark indicator corresponding to the benchmark indicator data. The formula for this indicator difference value can eliminate dimensional differences, making different business units and multidimensional indicators comparable.

[0040] Furthermore, the calculated index difference values ​​are transformed into a competitiveness index of 0 to 100 points through a mapping function. Negative values ​​correspond to the lagging range below the benchmark, while positive values ​​correspond to the leading range above the benchmark. The median of 50 points represents complete consistency with the benchmark. This mapping ensures that the competitiveness index is intuitive to interpret and manageable.

[0041] In this application, by dynamically comparing business indicator data with benchmark indicator data, a quantitative assessment of the enterprise's business objects in the competitive dimension is achieved. The standardized competitiveness index intuitively reflects the gap between the enterprise's operational performance and industry benchmarks or its own goals, providing data support for the generation of subsequent competitive correction schemes.

[0042] In some embodiments, the difference identification results of the above four dimensions can form a "four-dimensional attribution heatmap". The information dimension presents the spatiotemporal distribution of data breakpoints, the execution dimension calibrates the path dependence of task drift, the cognitive dimension maps the hierarchical decay of target understanding, and the competition dimension quantifies the convergence speed of indicator gaps.

[0043] Step S3: Based on the results of each difference identification, generate a correction plan for each business object through the large model.

[0044] It should be noted that the difference identification results are based on a comparison of business operation data across four dimensions: information, execution, cognition, and competition. These results represent various sets of deviation data, including information difference identification results, execution difference identification results, cognition difference identification results, and competition difference identification results. The large model refers to a privately deployed generative artificial intelligence model, pre-loaded with the target enterprise's business process specifications, strategic goals, and preset inference rules, capable of generating structured solution content based on input data. Correction solutions refer to a set of structured instructions generated by the large model to correct deviations for different types of difference identification results, including information correction solutions, execution correction solutions, cognition correction solutions, and competition correction solutions.

[0045] like Figure 3 As shown, specifically, step S3 includes the following sub-steps: Step S31: Based on the information difference identification results, generate an information correction scheme that includes the synchronous adjustment of the delay nodes of the corresponding business objects through the large model.

[0046] In this embodiment, the information difference identification results, including delayed node information and node update time difference data, are first input into the large model. The large model then performs logical processing based on the distribution of delayed nodes, node update time differences, and preset business synchronization rules, outputting an information correction scheme that includes delayed node synchronization adjustments. The information correction scheme includes a list of delayed nodes, synchronization adjustment order, synchronization time limit requirements, and data synchronization triggering conditions between nodes. This information correction scheme can automatically adapt to multi-source heterogeneous system environments and is compatible with mainstream middleware protocols, ensuring cross-system data consistency and timeliness.

[0047] Step S32: Based on the execution difference identification results, generate an execution correction scheme that includes the business execution path of the corresponding business object through the large model.

[0048] In this embodiment, the execution deviation identification results, including schedule deviation rate data and the difference between planned and actual nodes, are first input into the large model. The large model then performs logical processing based on the schedule deviation rate and the planned node data of business execution, combined with the preset business execution closed-loop logic, and outputs an execution correction plan that includes the business execution path of the corresponding business object. The execution correction plan includes the milestone event adjustment plan for business execution, the time correction data of key nodes, the responsible entity information of each node, and the node connection requirements. This execution correction plan supports dynamic rolling updates. When a new schedule deviation is detected, it can automatically trigger secondary reasoning and plan iteration to ensure that the business flow always fits the established rhythm.

[0049] Step S33: Based on the cognitive difference recognition results, generate a cognitive correction scheme that includes target alignment suggestions for the corresponding business objects through a large model.

[0050] In this embodiment, the matching degree between work report data and strategic goal data, as well as the cognitive gap identification results of the misunderstanding content, are first input into the large model. The large model then performs logical processing based on the misunderstanding content and the core requirements of the strategic goal, combined with the preset goal alignment logic, and outputs a cognitive correction scheme that includes goal alignment suggestions for the corresponding business objects. The cognitive correction scheme includes the interpretation of the core points of the strategic goal, the explanation of the correction of the misunderstanding content, the suggestion for adjusting employee tasks, and the implementation steps for goal alignment.

[0051] Step S34: Based on the competition difference identification results, generate a competition correction scheme that includes the competitiveness improvement path of the corresponding business object through the large model.

[0052] In this embodiment, the competition difference identification results, including competitiveness index data and differences between business indicators and benchmark indicators, are first input into the large model. The large model then performs logical processing based on the differences between the competitiveness index, business indicators and benchmark indicators, combined with the preset competitiveness improvement logic, and outputs a competition correction scheme that includes the competitiveness improvement path of the corresponding business object. The competition correction scheme includes the improvement target of business indicators, key improvement links, implementation steps and phased verification requirements.

[0053] In this application, the above steps utilize a large-scale privately deployed model to generate structured correction schemes for different types of differences, providing clear instructions for the subsequent collaborative execution of business intelligence agents, avoiding the subjectivity and disorder of deviation correction, and ensuring the pertinence and executability of deviation correction.

[0054] It should be noted that before executing step S4, each business intelligence agent needs to be constructed based on the preset business flow and data flow.

[0055] The business flow is a closed-loop chain consisting of core business processes, upstream and downstream collaborative nodes, and decision feedback loops. The data flow, on the other hand, is a dynamic execution framework defined by the target enterprise's business execution logic sequence and data semantic mapping relationships. Together, the business flow and data flow define the target enterprise's business execution logic sequence and the data interaction paths between various business agents, thus standardizing the collaborative logic of these agents. A business agent refers to an AI entity with specific business processing capabilities, built based on a pre-defined business flow and data flow, including project agents, job-specific agents, compliance agents, performance evaluation agents, and budget agents.

[0056] In this embodiment, the business flow is a closed loop of "SP-BP-TTC" (Strategic Plan-Business Plan-Task to Cash). The data flow starts from SP, is dynamically calibrated by BP, and then terminates at the TTC execution layer, with real-time data acquisition and semantic parsing nodes embedded throughout. Each node exchanges data at millisecond levels via standardized API interfaces and automatically triggers corresponding intelligent agents to coordinate responses based on business semantics. For example, when a payment delay occurs in the TTC stage, the system immediately coordinates with the performance evaluation agent to adjust performance weights, the budget agent to reset cash flow forecasts, and the compliance agent to initiate risk control review.

[0057] It's important to note that the Strategic Plan (SP) is used to formulate the company's overall strategic goals and direction, serving as the starting point for business processes. The Business Plan (BP) is used to develop specific business strategies and implementation plans based on strategic goals, ensuring alignment with the strategy through dynamic calibration. The Time-to-Choice (TTC) is used to translate the business plan into concrete execution tasks, ultimately achieving business monetization and cash flow management.

[0058] Furthermore, each business intelligence agent can load the corresponding task parsing engine, deviation response strategy library and real-time collaboration protocol according to its role in the closed-loop chain, so as to ensure that after receiving the deviation correction scheme generated in the S3 stage, it can accurately deconstruct the instruction semantics, automatically match resource capabilities, and synchronously trigger the cross-node actions in step S4, forming a millisecond-level closed loop of "identification-decision-execution-feedback".

[0059] Step S4: The corrective action plan for each business object is executed collaboratively by multiple pre-defined business intelligence agents.

[0060] It should be noted that collaborative execution refers to the process in which multiple business intelligence agents, based on a pre-defined division of labor, execute the corresponding instructions in the correction scheme in sequence or in parallel to jointly complete the deviation correction.

[0061] like Figure 4 As shown, specifically, step S4 includes the following sub-steps: Step S41: The project intelligent agent and the compliance intelligent agent work together to execute the information correction plan for the corresponding business object.

[0062] In this embodiment, the project agent receives the synchronization adjustment instruction for delayed nodes in the information correction scheme, triggers data synchronization operations for the delayed nodes according to the synchronization adjustment sequence, and provides real-time feedback on the node synchronization status during the synchronization process. Project information is aligned from top to bottom and from outside to inside, automatically generating daily and weekly reports. Information is aligned by combining target achievement rates, eliminating information gaps. Simultaneously, the compliance agent performs compliance verification on the synchronization process of delayed nodes, checking whether the node synchronization operations comply with data security specifications and access control requirements. Non-compliant synchronization operations are intercepted and reported. Through real-time linkage between the project agent and the compliance agent, information synchronization is ensured to be both efficient and secure, and the entire information deviation correction process is traceable, auditable, and fully auditable.

[0063] Step S42: The project agent, job agent, and compliance agent work together to execute the execution correction plan for the corresponding business object.

[0064] In this embodiment, the project agent receives the business execution path adjustment instructions from the execution correction plan, advances business execution according to the adjusted milestone events and key nodes, and synchronously updates business execution progress data. The job agent adjusts employee task assignments and execution requirements according to the business execution path adjustment plan, synchronously updates employee task status data, and dynamically analyzes the matching degree between employee capabilities and task requirements based on the "person-job matching model," adjusting performance bonuses and penalties accordingly. The compliance agent performs compliance verification on the business execution process, verifying whether the operations at business execution nodes comply with business specifications and risk control requirements, issuing warnings and providing feedback on non-compliant operations.

[0065] Step S43: The cognitive correction scheme for the corresponding business object is executed collaboratively by the job-specific intelligent agent and the compliance intelligent agent.

[0066] In this embodiment, the job-specific intelligent agent receives goal alignment suggestions from the cognitive correction scheme, pushes strategic goal interpretations, deviation correction explanations, and task adjustment suggestions to employees, and simultaneously records employee feedback data. The compliance intelligent agent performs compliance verification on the content of the goal alignment suggestions, verifying whether the content complies with enterprise management regulations and information release requirements, and intercepts and reports non-compliant content.

[0067] Step S44: The project agent, compliance agent, and assessment agent work together to execute the competitive correction plan for the corresponding business object.

[0068] In this embodiment, the project agent receives the competitiveness enhancement path instructions from the competition correction plan, advances the implementation of key enhancement steps according to the enhancement path, and synchronously updates the progress data of business indicator enhancement. The compliance agent performs compliance verification on the implementation process of the competitiveness enhancement steps, verifies whether the implementation operations comply with industry standards and corporate compliance requirements, and issues warnings and provides feedback on non-compliant implementation operations. The assessment agent receives the progress data of business indicator enhancement, tracks and assesses the progress of indicator enhancement according to the phased verification requirements, and synchronously generates assessment result data, realizing automated assessment of all employees and shortening the "goal-result" feedback cycle from monthly to real-time.

[0069] In this application, the above steps enable the collaborative execution of multiple business intelligence agents. Different types of deviation correction schemes are executed by combinations of corresponding business intelligence agents, ensuring the professionalism and efficiency of deviation correction. At the same time, compliance risks in the deviation correction process are avoided through full-process verification by the compliance intelligence agent.

[0070] Furthermore, the business intelligence agent also includes a budget intelligence agent, and the business management method also includes: based on the resource requirements of each correction scheme, the budget intelligence agent performs resource scheduling during the execution of each correction scheme.

[0071] Resource requirements refer to the amount of financial, human, and technical resources required for each corrective action plan. The budgeting intelligence dynamically calculates, allocates, and monitors resource usage based on historical execution data and current business conditions, determining product / service budgets to ensure precise matching of resource investment with corrective action goals, avoiding redundancy or shortages, providing resource guarantees for the execution of corrective action plans, and achieving closed-loop management of deviation correction; simultaneously, it generates resource usage efficiency analysis reports to support continuous optimization of subsequent corrective action strategies.

[0072] Furthermore, a comprehensive intelligent agent is set up as the global coordination hub to schedule the dynamic coordination of various business intelligent agents according to task priorities and resource constraints, and to integrate multi-dimensional data streams such as positions, projects, compliance, assessments, and budgets in real time to ensure the real-time consistency of information and actions of all employees.

[0073] In summary, the aforementioned agent-based business management method includes the following steps: acquiring business operation data of multiple business objects within the target enterprise to be managed; identifying differences in each business operation data based on multiple preset difference dimensions to obtain the difference identification results corresponding to each difference dimension; generating correction schemes for each business object through a large model based on the difference identification results; and collaboratively executing the correction schemes for each business object through multiple preset business intelligent agents. This method, by identifying the difference identification results corresponding to each difference dimension throughout the entire enterprise operation chain and correcting the difference identification results of multiple difference dimensions through the collaborative operation of multiple business intelligent agents, achieves accurate location and dynamic response to business anomalies, and realizes automatic correction and multi-terminal aligned collaborative management, significantly improving the enterprise's business management efficiency and decision-making quality.

[0074] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0075] In one embodiment, an agent-based service management device is provided, which corresponds one-to-one with the agent-based service management method described in the above embodiments. For example... Figure 5 As shown, the agent-based business management device includes a data acquisition module 101, a difference recognition module 102, a solution generation module 103, and a solution execution module 104. Detailed descriptions of each functional module are as follows: The data acquisition module 101 is used to acquire business operation data of multiple business objects in the target enterprise to be managed.

[0076] The difference recognition module 102 is used to perform difference recognition on various business operation data based on multiple preset difference dimensions, and obtain the difference recognition results corresponding to each difference dimension.

[0077] The solution generation module 103 is used to generate correction solutions for each business object based on the results of each difference identification through a large model.

[0078] The scheme execution module 104 is used to collaboratively execute the correction schemes of various business objects through multiple preset business intelligent agents.

[0079] For specific limitations regarding the agent-based service management device, please refer to the limitations of the agent-based service management method above, which will not be repeated here. Each module in the aforementioned agent-based service management device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0080] Furthermore, the business management device can be presented as a business management system in software. The aforementioned agent-based business management method runs on the corresponding business management system, which includes a data input layer, a business agent cluster, an active service triggering module, a super agent hub, a service output layer, and a memory module.

[0081] The data input layer provides multi-source, multi-modal data, specifically including third-party platforms, market and industry data, company data assets, and employee personal data assets. The data types include images, audio / video, and text. Third-party platforms provide external data input to the company's data assets; market and industry data, company data assets, and employee personal data assets provide foundational data support for subsequent business processing.

[0082] The business intelligence agent cluster is connected sequentially according to a preset business flow, specifically including budget intelligence agents, project intelligence agents, and job-specific intelligence agents, used for hierarchical business processing of data from the data input layer. Specifically, the budget intelligence agent is connected to the company's data assets; the project intelligence agents are connected to both the budget intelligence agent and market / industry data; and the job-specific intelligence agents are connected to the project intelligence agent, company data assets, market / industry data, and employee personal data assets.

[0083] The proactive service triggering module is connected to the super intelligent agent's central hub and is used to provide triggering conditions for proactive services. These triggering conditions include time, latest data, latest hot events, and custom conditions.

[0084] The integrated intelligent agent is the Happy Work super-agent, which serves as the main body for global scheduling and service output of the system. Its inputs include the business processing results of the job-specific intelligent agents, the triggering conditions of the proactive service triggering module, and the data of the memory module. Its outputs include work-related services and life-related services. The work-related services include the latest meetings, messages, reminders, to-do lists, training, company news, and today's dashboard. The life-related services include today's meals and exercise plans.

[0085] The memory module is used to store user interaction data, business data, and user preference characteristics, providing data support for personalized services to the central hub of the super intelligent agent.

[0086] The service output layer includes Happy Life concierge and expert roles. The Happy Life concierge is connected to the super intelligent agent central hub to provide quality life guarantee services. The expert role is used to provide consulting services in human resources, legal affairs, finance, psychology and professional fields, and supports professional problem consultation referral and full-duplex in-depth conversation.

[0087] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 6 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements an agent-based business management method.

[0088] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the agent-based business management method described in the above embodiments, for example... Figure 1 As shown in S1-S4, or Figures 2 to 4 As shown, to avoid repetition, it will not be described again here. Alternatively, when the processor executes the computer program, it implements the functions of each module / unit in this embodiment of the agent-based business management device, for example... Figure 5 The functions of the data acquisition module 101, difference identification module 102, scheme generation module 103, and scheme execution module 104 shown are not described again here to avoid repetition.

[0089] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program implements the agent-based business management method described in the above embodiments, for example... Figure 1 As shown in S1-S4, or Figures 2 to 4 As shown, to avoid repetition, it will not be described again here. Alternatively, when the computer program is executed by the processor, it implements the functions of each module / unit in this embodiment of the agent-based business management device, for example... Figure 5 The functions of the data acquisition module 101, difference identification module 102, scheme generation module 103, and scheme execution module 104 shown are not described again here to avoid repetition. The computer-readable storage medium can be non-volatile or volatile.

[0090] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0091] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0092] It should be noted that any AI models, software tools, or components not belonging to this company appearing in the embodiments of this application are merely illustrative examples and do not represent actual use. All user personal information involved in the embodiments of this application has been authorized (with the knowledge and consent) by the relevant parties or has been fully authorized by all parties, and the executing entity may obtain it through various legal and compliant means. The collection, storage, use, processing, transmission, provision, and disclosure of the information, data, and signals involved all comply with relevant laws and regulations and do not violate public order and good morals.

[0093] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A business management method based on intelligent agents, characterized in that, Including the following steps: Acquire business operation data of multiple business objects within the target enterprise to be managed; Based on multiple preset difference dimensions, the difference identification of each business operation data is performed to obtain the difference identification result corresponding to each difference dimension; Based on the differences identified, a correction scheme for each of the business objects is generated using a large model. Multiple pre-defined business intelligence agents collaboratively execute the correction schemes for each business object.

2. The business management method according to claim 1, characterized in that, The difference dimensions include information dimension, execution dimension, cognitive dimension, and competitive dimension. The business operation data includes the node update time of the business object, execution process data, work report data, and business indicator data. Based on multiple preset difference dimensions, the difference identification of each of the business operation data is performed to obtain the difference identification result corresponding to each difference dimension, including: The node update time is compared with the preset update execution time to identify delayed nodes and obtain the information difference identification result of the information dimension. The execution process data is compared with the preset plan node data to identify the progress deviation rate and obtain the execution difference identification result of the execution dimension. The work report data is matched with the preset strategic target data to identify comprehension biases and obtain the cognitive difference identification results of the cognitive dimension. The business indicator data is compared with the preset benchmark indicator data to identify the competitiveness index and obtain the competition difference identification result of the competition dimension.

3. The business management method according to claim 2, characterized in that, The step of generating a correction scheme for each of the business objects based on the difference identification results through a large model includes: Based on the information difference identification results, an information correction scheme including the synchronous adjustment of delay nodes corresponding to the business objects is generated through the large model; Based on the execution difference identification results, an execution correction scheme including the business execution path corresponding to the business object is generated through the large model; Based on the cognitive discrepancy recognition results, a cognitive correction scheme including target alignment suggestions corresponding to the business object is generated through the large model; Based on the competition difference identification results, a competition correction scheme including a competitiveness enhancement path for the corresponding business object is generated through the large model.

4. The business management method according to claim 3, characterized in that, The business intelligence agents include project intelligence agents, job intelligence agents, compliance intelligence agents, and performance evaluation intelligence agents. The step of collaboratively executing corrective action plans for each business object through multiple pre-defined business intelligence agents includes: The project intelligent agent and the compliance intelligent agent work together to execute the information correction scheme corresponding to the business object. The project agent, the job agent, and the compliance agent work together to execute the execution correction scheme corresponding to the business object. The cognitive correction scheme for the corresponding business object is executed collaboratively by the job-specific intelligent agent and the compliance intelligent agent. The project agent, the compliance agent, and the assessment agent work together to execute the competition correction scheme corresponding to the business object.

5. The business management method according to claim 1, characterized in that, The business intelligence agent also includes a budget intelligence agent, and the business management method further includes: Based on the resource requirements of each of the aforementioned correction schemes, the budget agent performs resource scheduling during the execution of each of the aforementioned correction schemes.

6. The business management method according to claim 1, characterized in that, The business management method also includes: Based on the preset business flow and data flow, each business intelligent agent is constructed.

7. The business management method according to claim 1, characterized in that, The business management method also includes: Define each of the aforementioned business objects, and define the metric dimensions for each of the aforementioned business objects; The acquisition of business operation data of multiple business objects in the target enterprise to be managed includes: Based on the metric dimensions of each business object, obtain the corresponding business operation data for that metric dimension.

8. A business management device based on intelligent agents, characterized in that, include: The data acquisition module is used to acquire business operation data of multiple business objects in the target enterprise to be managed; The difference recognition module is used to perform difference recognition on each of the business operation data based on multiple preset difference dimensions, and obtain the difference recognition result corresponding to each difference dimension; The solution generation module is used to generate correction solutions for each of the business objects based on the difference identification results and through a large model. The scheme execution module is used to collaboratively execute the correction schemes of each business object through multiple preset business intelligent agents.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the business management method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the business management method as described in any one of claims 1 to 7.

Images