Enterprise intelligent management method and system based on artificial intelligence

Abstract

The present invention relates to the field of digital management technology, and in particular to an enterprise intelligent management method and system based on artificial intelligence. The method comprises the following steps: obtaining enterprise transaction data of an enterprise management resource library; analyzing associated enterprise transaction feature data based on enterprise transaction data; analyzing customer associated feature data, enterprise contract data and project task data through associated enterprise transaction feature data; designing differentiated customer maintenance operation management decisions through customer associated feature data; performing enterprise contract anomaly analysis based on enterprise contract data, and designing enterprise contract anomaly feedback decisions through abnormal enterprise contract data; designing project task scheduling demand optimization decisions based on project task data; and feeding back differentiated customer maintenance operation management decisions, enterprise contract anomaly feedback decisions and project task scheduling demand optimization decisions to a terminal to execute intelligent enterprise management operations. The present invention realizes automation and intelligence of enterprise management.

Description

Technical Field

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

[0002] With the rapid development of globalization and informatization, the operation and management of enterprises have become more and more complex, involving multi-domain transaction coordination, resource allocation, and customer relationship maintenance. In the enterprise management model, enterprises process a large amount of transactional data manually or with the support of simple information systems. Artificial Intelligence (AI) technology, including sub-fields such as machine learning, natural language processing, data mining, and computer vision, is being widely used at all levels of enterprise management. These technologies extract valuable information from massive amounts of enterprise data and make predictions and analyses based on historical data, thereby assisting managers in making more accurate enterprise management decisions through artificial intelligence technology. However, the existing enterprise intelligent management methods lack personalized and precise means for different users through customer tracking and maintenance, resulting in poor customer maintenance management effects. In addition, the review of enterprise contracts relies on manual or simple rules, which easily ignores potential risks in the contract. The tracking of project tasks also lacks in-depth analysis of task dependencies and scheduling requirements, resulting in poor follow-up execution efficiency of project tasks when facing complex project dependencies or resource constraints. Summary of the invention

[0003] Based on this, the present invention provides an enterprise intelligent management method and system based on artificial intelligence to solve at least one of the above technical problems.

[0004] To achieve the above purpose, an enterprise intelligent management method based on artificial intelligence includes the following steps:

[0005] Step S1: Acquire enterprise transaction data from an enterprise management resource library; perform feature analysis and processing on related enterprise transactions based on the enterprise transaction data to generate related enterprise transaction feature data;

[0006] Step S2: extract customer-related features based on the associated enterprise transaction feature data to generate customer-related feature data; perform classified customer portrait analysis and processing based on the customer-related feature data to generate classified customer portrait data; perform differentiated customer maintenance operation management strategy intelligent analysis based on clustered customer behavior data to obtain differentiated customer maintenance operation management decisions;

[0007] Step S3: Collect and classify the enterprise contract data corresponding to the customer portrait data according to the characteristic data of the associated enterprise transactions; perform enterprise contract anomaly analysis based on the enterprise contract data, and generate abnormal enterprise contract data and valid enterprise contract data respectively, wherein the abnormal enterprise contract data includes abnormal nodes of contract key clause attributes and contract event execution feasibility data; perform enterprise contract anomaly feedback strategy analysis through abnormal nodes of contract key clause attributes corresponding to the abnormal enterprise contract data and contract event execution feasibility data, so as to obtain enterprise contract anomaly feedback decision;

[0008] Step S4: collecting and monitoring project task data corresponding to effective enterprise contract data based on the associated enterprise transaction characteristic data; performing project task scheduling demand optimization strategy analysis based on the project task data to obtain project task scheduling demand optimization decisions;

[0009] Step S5: Based on the differentiated customer maintenance operation management decisions, enterprise contract exception feedback decisions and project task scheduling demand optimization decisions, the intelligent enterprise management operations are fed back to the terminal, wherein the intelligent enterprise management operations include enterprise customer maintenance operation management operations, enterprise contract exception feedback operations and project task scheduling demand optimization operations.

[0010] Furthermore, step S1 includes the following steps:

[0011] Step S11: Acquire enterprise transaction data from the enterprise management resource library;

[0012] Step S12: Perform heterogeneous data analysis on the enterprise transaction data to generate enterprise transaction heterogeneous data;

[0013] Step S13: extracting enterprise transaction traceability tags from enterprise transaction data to generate enterprise transaction traceability tag data;

[0014] Step S14: establishing an enterprise transaction data matrix according to the enterprise transaction heterogeneous type data, and transferring the enterprise transaction data to the enterprise transaction data matrix through the enterprise transaction traceability tag data to perform enterprise transaction data association mapping to obtain an associated enterprise transaction data matrix;

[0015] Step S15: performing correction processing on similar redundant data of associated nodes on the associated enterprise affairs data matrix to generate a corrected associated enterprise affairs data matrix;

[0016] Step S16: Perform feature analysis processing based on the corrected associated enterprise transaction data matrix to generate associated enterprise transaction feature data.

[0017] Further, step S2 includes the following steps:

[0018] Step S21: extracting customer-related features according to the related enterprise transaction feature data to generate customer-related feature data, wherein the customer-related feature data includes customer attribute label data and customer behavior data;

[0019] Step S22: Performing behavior clustering analysis on the customer behavior data to generate clustered customer behavior data;

[0020] Step S23: performing classified customer portrait analysis and processing based on the customer attribute label data and the clustered customer behavior data to generate classified customer portrait data;

[0021] Step S24: Performing intelligent analysis of differentiated customer maintenance operation management strategies based on the clustered customer behavior data to obtain differentiated customer maintenance operation management decisions.

[0022] Further, step S24 includes the following steps:

[0023] Step S241: performing customer group stickiness index analysis based on the clustered customer behavior data to generate customer group stickiness index data;

[0024] Step S242: setting a customer group stickiness weight parameter of the customer group stickiness index data; designing a customer group stickiness evaluation model through the customer group stickiness weight parameter and the customer group stickiness index data;

[0025] Step S243: performing personalized customer behavior prediction processing based on the clustered customer behavior data to generate personalized customer behavior prediction data;

[0026] Step S244: using the customer group stickiness evaluation model to perform differentiated enterprise customer stickiness feature analysis processing on the personalized customer behavior prediction data to generate differentiated enterprise customer stickiness feature data;

[0027] Step S245: Performing intelligent analysis of differentiated customer maintenance operation management strategies based on differentiated enterprise customer stickiness feature data to obtain differentiated customer maintenance operation management decisions.

[0028] Further, step S243 includes the following steps:

[0029] Define the probability coefficient of customer group behavior type based on clustered customer behavior data and generate the probability coefficient of customer group behavior type;

[0030] Perform customer behavior frequency feature analysis based on clustered customer behavior data to generate customer behavior frequency feature data;

[0031] Calculate the personalized customer behavior conditional probability based on the customer group behavior type probability coefficient and customer behavior frequency characteristic data to generate personalized customer behavior conditional probability data;

[0032] Personalized customer behavior prediction processing is performed based on personalized customer behavior conditional probability data to generate personalized customer behavior prediction data.

[0033] Further, step S3 includes the following steps:

[0034] Step S31: collecting and classifying the corresponding enterprise contract data in the customer portrait data according to the associated enterprise transaction feature data;

[0035] Step S32: performing enterprise contract parsing processing on the enterprise contract data based on a preset semantic parsing model to generate enterprise contract parsing data; performing contract key clause attribute identification based on the enterprise contract parsing data to generate contract key clause attribute data;

[0036] Step S33: performing contract clause rule matching exception marking processing based on the contract key clause attribute data, and generating a contract key clause attribute exception node;

[0037] Step S34: Performing a feasibility analysis on the execution of a contract event based on the attribute data of the key contract terms, and generating feasibility data on the execution of the contract event;

[0038] Step S35: when an abnormal node of a key contract clause attribute is identified or the contract event execution feasibility data is less than a preset contract event execution feasibility threshold, the enterprise contract data is marked as abnormal enterprise contract data; otherwise, the enterprise contract data is marked as valid enterprise contract data;

[0039] Step S36: Analyze the enterprise contract abnormality feedback strategy through the abnormal nodes of the contract key clause attributes corresponding to the abnormal enterprise contract data and the contract event execution feasibility data to obtain the enterprise contract abnormality feedback decision.

[0040] Further, step S33 includes the following steps:

[0041] Acquire a contract terms and rules knowledge base; perform contract terms and rules matching anomaly analysis on the contract key terms and rules attribute data according to the contract terms and rules knowledge base; when it is identified that the contract key terms and rules attribute data do not match the contract terms and rules knowledge base, mark the corresponding contract key terms and rules attribute data as a contract key terms and rules attribute anomaly node.

[0042] Further, step S34 includes the following steps:

[0043] Obtain the historical contract execution event library; perform contract execution logic relationship analysis on the key contract clause attribute data to obtain contract execution logic relationship data;

[0044] Perform contract event execution reasoning analysis based on contract execution logic relationship data to generate contract event execution reasoning data;

[0045] The contract event execution feasibility analysis is performed on the contract event execution reasoning data according to the historical contract execution event library to generate the contract event execution feasibility data.

[0046] Further, step S4 includes the following steps:

[0047] Step S41: collecting and monitoring project task data corresponding to valid enterprise contract data based on the associated enterprise transaction feature data;

[0048] Step S42: performing a project task scheduling requirement dependency analysis according to the project task data to generate project task scheduling requirement dependency analysis data, wherein the project task scheduling requirement dependency analysis data includes the project task scheduling requirement pre-dependency analysis data, scheduling requirement resource dependency analysis data, and scheduling requirement timing dependency analysis data;

[0049] Step S43: performing a weight evaluation analysis of the project task scheduling requirements based on the project task scheduling requirements dependency analysis data to design a weight parameter of the project task scheduling requirements;

[0050] Step S44: performing a project task scheduling response characteristic analysis on the project task data according to the project task scheduling requirement dependency analysis data to generate project task scheduling response characteristic data;

[0051] Step S45: performing a quantitative analysis of the project task scheduling deviation on the project task scheduling response characteristic data to generate quantitative data of the project task scheduling deviation;

[0052] Step S46: Performing a project task scheduling requirement optimization strategy analysis based on the project task scheduling requirement weight parameter and the project task scheduling deviation quantification data to obtain a project task scheduling requirement optimization decision.

[0053] This specification provides an enterprise intelligent management system based on artificial intelligence, which is used to execute the enterprise intelligent management method based on artificial intelligence as described above. The enterprise intelligent management system based on artificial intelligence includes:

[0054] The associated enterprise affairs feature analysis module is used to obtain enterprise affairs data from the enterprise management resource library; perform feature analysis and processing on associated enterprise affairs based on the enterprise affairs data to generate associated enterprise affairs feature data;

[0055] The customer management module is used to extract customer-related features based on the characteristic data of related enterprise transactions to generate customer-related feature data; to analyze and process classified customer portraits based on the customer-related feature data to generate classified customer portrait data; to perform intelligent analysis of differentiated customer maintenance operation management strategies based on clustered customer behavior data to obtain differentiated customer maintenance operation management decisions;

[0056] The contract management module is used to collect and classify the corresponding enterprise contract data in the customer portrait data according to the characteristic data of the associated enterprise affairs; perform enterprise contract anomaly analysis based on the enterprise contract data, and generate abnormal enterprise contract data and valid enterprise contract data respectively, wherein the abnormal enterprise contract data includes abnormal nodes of key contract terms and contract event execution feasibility data; perform enterprise contract anomaly feedback strategy analysis through the abnormal nodes of key contract terms and contract event execution feasibility data corresponding to the abnormal enterprise contract data, so as to obtain enterprise contract anomaly feedback decision;

[0057] The project task management module is used to collect and monitor the project task data corresponding to the effective enterprise contract data based on the characteristic data of the associated enterprise affairs; analyze the project task scheduling demand optimization strategy based on the project task data to obtain the project task scheduling demand optimization decision;

[0058] The enterprise intelligent management module is used to execute intelligent enterprise management operations based on differentiated customer maintenance operation management decisions, enterprise contract exception feedback decisions, and project task scheduling demand optimization decision feedback to the terminal, wherein the intelligent enterprise management operations include enterprise customer maintenance operation management operations, enterprise contract exception feedback operations, and project task scheduling demand optimization operations.

[0059] The beneficial effect of this application is that the enterprise transaction data is obtained through the enterprise management resource library, and the multi-dimensional feature analysis and processing of these data are performed to automatically generate the characteristic data of related enterprise transactions. The whole process utilizes the data-driven artificial intelligence model, which can quickly and efficiently complete the processing and analysis of data, replacing the traditional manual method, thereby greatly improving the overall management efficiency of the enterprise, and the enterprise management becomes more refined and accurate, reducing redundant work and improving the response speed of decision-making. Through the analysis of the characteristic data of related enterprise transactions, the behavioral characteristics of customers are extracted, and then customer portraits are classified. Based on the cluster analysis of these portraits and customer behaviors, enterprises can better understand the needs and behavior patterns of different types of customers, and predict and classify them through customer stickiness evaluation models, which provides support for enterprises to formulate differentiated customer maintenance strategies. The intelligent customer management model enables enterprises to accurately identify customer groups, optimize resource allocation, and improve customer satisfaction and loyalty, thereby occupying a more advantageous position in market competition. In the process of enterprise contract data parsing, a contract clause rule matching mechanism based on semantic analysis is adopted, which can timely discover abnormal nodes of key clauses in the contract and perform intelligent reasoning analysis on the feasibility of contract execution. Once potential problems in contract execution are discovered, the system will provide feedback strategy suggestions so that the enterprise can take remedial measures, improve the accuracy of contract management, prevent contract disputes, and reduce the legal and financial risks of the enterprise. By using the characteristic data of related enterprise affairs and contract data, the dependencies, scheduling requirements and response characteristics of project tasks are analyzed. Through intelligent analysis of multi-dimensional weight evaluation and deviation quantification, an optimized project task scheduling strategy is designed, which improves the flexibility and adaptability of project management and reduces resource waste and time delays during project implementation. Through automatic scheduling optimization of artificial intelligence, enterprises can ensure that projects are completed efficiently under limited resources and time conditions and maximize business goals. Through artificial intelligence technology, a closed-loop management system with data as the core is constructed, involving all aspects of corporate affairs, customer management, contract management and project scheduling. The management content of multiple dimensions is highly integrated in the same system. The management decision of each link is fed back to the next stage through the analysis results of the previous stage, and connected with the terminal execution system to ensure the closed-loop implementation and continuous optimization of the decision, which can ensure the real-time and flexibility of enterprise management. It can also continuously improve and improve the operation of each link through the feedback mechanism, and promote the gradual transformation of enterprises to intelligent and data-driven. Therefore, the enterprise method based on artificial intelligence of the present invention ensures the high efficiency of customer maintenance management and the accuracy of customer needs by designing intelligent decisions for personalized customer maintenance operations for different users. By identifying abnormal clauses and executable risks in enterprise contracts through intelligent analysis means, the accuracy and security of contract management are improved, and the execution ability and management efficiency of enterprise contracts are guaranteed.Through multi-dimensional analysis of the dependencies of project task data, the scheduling requirements of project tasks are comprehensively evaluated, and according to the importance of scheduling requirements, the resource scheduling requirements of project tasks are intelligently adjusted to improve the follow-up execution efficiency of project tasks. Cross-domain data association analysis and decision optimization have achieved efficient and intelligent enterprise management, and integrated analysis has greatly reduced the island effect between various management systems. It has a high degree of adaptability and can flexibly respond to the management needs of enterprises of different sizes and types. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 A schematic diagram of the steps of an enterprise intelligent management method based on artificial intelligence according to the present invention;

[0061] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0062] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG.

[0063] Figure 4 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0064] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0065] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0066] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0067] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0068] To achieve this, please refer to Figures 1 to 4 The present invention provides an enterprise intelligent management method based on artificial intelligence. In the embodiments of the present invention, please refer to Figure 1 FIG. 1 is a schematic diagram of the process flow of the enterprise intelligent management method based on artificial intelligence of the present invention, and the enterprise intelligent management method based on artificial intelligence includes the following steps:

[0069] Step S1: Acquire enterprise transaction data from an enterprise management resource library; perform feature analysis and processing on related enterprise transactions based on the enterprise transaction data to generate related enterprise transaction feature data;

[0070] In an embodiment of the present invention, the enterprise transaction data in the enterprise management resource library is obtained, and the enterprise management resource library is accessed through the enterprise data management system. The resource library contains detailed records of various enterprise transactions, such as customer order information, supplier contracts, financial transactions, project progress and other data. The enterprise transaction data is stored in both structured and unstructured forms. The system extracts customer order records and contract information from the structured database by combining SQL query and NoSQL query tools, and extracts relevant data from unstructured documents (such as PDF, Word files) through text mining technology (such as Apache Tika). During the data acquisition process, all transaction data are converted into a standardized data format for unified feature analysis. After completing the data acquisition, the system performs association analysis on the enterprise transaction data and labels each type of transaction data, such as marking customer orders as "order data" and contract information as "contract data". Then, the system constructs a graph structure based on a graph database, takes each transaction data as a node, and takes the relationship between transactions as an edge to generate a relationship graph of enterprise transactions. Based on the relationship graph, the system applies the PageRank algorithm in social network analysis to calculate the importance of each transaction node and identify key transaction nodes that have a strong correlation with other transaction data. These key transactions are analyzed through feature extraction models (such as PCA dimensionality reduction models) to generate feature data of related corporate transactions. These feature data are the basis for subsequent customer portrait analysis and management decisions.

[0071] Step S2: extract customer-related features based on the associated enterprise transaction feature data to generate customer-related feature data; perform classified customer portrait analysis and processing based on the customer-related feature data to generate classified customer portrait data; perform differentiated customer maintenance operation management strategy intelligent analysis based on clustered customer behavior data to obtain differentiated customer maintenance operation management decisions;

[0072] In an embodiment of the present invention, customer-related features are extracted based on the feature data of the associated enterprise transactions, the feature data of the associated enterprise transactions are clustered, and the enterprise transactions with similar features are grouped using the K-means algorithm. For example, order features, payment terms, etc. are clustered together to identify customers with similar behaviors. Each clustering result contains data features of several transactions, which reflect the customer's behavior patterns, such as high-frequency purchases, delayed payments, etc. Then, the system further extracts features from the transaction data in each cluster group, and summarizes the behavior data of each customer, such as order frequency, order amount, contract fulfillment status, etc., into customer attribute features. Through feature extraction tools, such as the Scikit-learn library in Python, the system extracts key associated feature data of customers, such as customer loyalty, customer contribution, etc. After the customer-related feature extraction is completed, customer portrait analysis is performed based on the customer-related feature data, and the customer segmentation model is used to divide the customer data into different customer groups, for example, based on the RFM model (Recency, Frequency, Monetary), customers are divided into high-value customers, loyal customers, and customers who are about to lose. For each type of customer, the system creates a profile based on its characteristic data, including purchase behavior analysis, payment habits, demand preferences, etc. For example, for high-frequency purchase customers, the profile may include characteristic information such as the number of purchases, purchase time intervals, product preferences, etc. The generated classified customer profile data provides support for corporate decision-making. Based on clustered customer behavior data, differentiated customer maintenance and operation management strategy intelligent analysis is performed, and the behavior pattern clustering analysis is performed on the behavior data of each customer group. Customer behavior is clustered using DBSCAN (density-based clustering algorithm) to identify similar patterns of customers in the purchase cycle. For example, for high-value customers, their behavior patterns may be concentrated on regular large-amount purchases, while for low-value customers, their purchase behavior may be more random and less frequent. Based on the behavior patterns of different customer groups, the system intelligently generates differentiated maintenance and operation management strategies. For example, for high-value customers, the system may formulate customized customer service and reward plans; for customers who are about to churn, the system generates incentive plans, such as sending exclusive coupons or proactively contacting for follow-up. The system generates optimal strategy decisions through a rule engine (such as the Drools rule engine) combined with a machine learning model (such as a support vector machine SVM). These decisions include customer contact frequency, maintenance channel selection (such as telephone, email, social media, etc.), specific marketing content, etc. All these strategies are packaged into a complete set of customer maintenance operation management decisions and fed back to the relevant operation team for implementation through the enterprise management platform.

[0073] Step S3: Collect and classify the enterprise contract data corresponding to the customer portrait data according to the characteristic data of the associated enterprise transactions; perform enterprise contract anomaly analysis based on the enterprise contract data, and generate abnormal enterprise contract data and valid enterprise contract data respectively, wherein the abnormal enterprise contract data includes abnormal nodes of contract key clause attributes and contract event execution feasibility data; perform enterprise contract anomaly feedback strategy analysis through abnormal nodes of contract key clause attributes corresponding to the abnormal enterprise contract data and contract event execution feasibility data, so as to obtain enterprise contract anomaly feedback decision;

[0074] In the embodiment of the present invention, the contract data is accurately collected based on the enterprise transaction feature data and classified customer portrait data generated in the early stage, and the contract number and basic information of the contract involved by each customer are determined through the transaction feature data in the enterprise management resource library, which includes the customer ID, contract type, amount, signing date, contract status, etc. of the contract. The system uses SQL query tools to query the enterprise transaction database and extract the contract information corresponding to each customer portrait data. For example, for the customer portrait marked as "high-value customer", the system extracts all the contract data of the customer, including historical contracts and contracts being performed. By matching the customer portrait data with the contract data, the system can form a complete contract data set for each customer. The data set is stored in the form of a structured table, including detailed data such as contract number, customer information, contract terms, etc., for subsequent abnormal analysis. After completing the enterprise contract data collection, the contract data is analyzed for enterprise contract abnormalities, and the key terms in each contract are analyzed in detail, especially the important contents such as payment terms, delivery terms, and breach of contract terms. The system uses a contract parsing model based on rule matching (for example, using NLP tools and Drools rule engines) to perform compliance checks on key contract terms, such as whether the payment period meets the company's standards and whether the delivery conditions are reasonable. By parsing the contract text, the system identifies abnormal information in each contract clause and marks these clauses that do not meet the company's standards or regulatory requirements as "abnormal nodes". At the same time, the system evaluates the feasibility of contract event execution based on the historical execution data of contract performance. For example, the system compares the delivery time marked in the contract with the historical project execution status. If it is identified that the delivery time is unrealistic, the system will determine that the contract has low execution feasibility. Finally, the system divides all contract data into two categories: one is contracts with abnormal nodes or low execution feasibility, marked as "abnormal enterprise contract data"; the other is contracts that meet the standards and have high execution feasibility, marked as "valid enterprise contract data". This process is performed through an automated discrimination model. All contracts are marked and stored in the contract management database, and classified and archived according to "abnormal contracts" and "valid contracts". After marking the abnormal contract data, the system conducts an abnormal feedback strategy analysis on the abnormal enterprise contract data, and conducts a detailed analysis of the data of each abnormal contract, focusing on the abnormal nodes of each contract. For example, if the payment terms show that the payment cycle is too long, or the delivery terms of the contract require completion in an unreasonably short time, the system will record these abnormal terms. At the same time, the system also analyzes the feasibility data of the contract. If the feasibility analysis results show that some contracts have a high performance risk, such as too high delivery difficulty, the system will mark these contracts as high-risk contracts.Next, the system automatically generates enterprise contract abnormality feedback decisions by comprehensively analyzing these abnormal nodes and contract execution feasibility data. It uses decision support models (such as decision tree models or rule-based expert systems) and combines historical contract processing experience to generate corresponding feedback strategies for each abnormal type. For example, if the payment terms are unreasonable, the system recommends renegotiating the payment terms with the customer to shorten the payment cycle to a reasonable range; if there are problems with the delivery time, the system recommends re-evaluating the project resource allocation or extending the delivery time to ensure the feasibility of the project. These feedback strategies automatically generate detailed feedback reports through the contract management platform, which list the abnormal terms of the contract, the specific description of the problem, the cause of the problem, and the recommended treatment measures. The final enterprise contract abnormality feedback decision is passed to the contract management department, and the contract management team modifies or renegotiates the abnormal contract based on the feedback strategy. Through this process, the system realizes the automatic identification, in-depth analysis, and intelligent feedback of contract abnormalities, improving the efficiency of contract management and the security of contract performance.

[0075] Step S4: collecting and monitoring project task data corresponding to effective enterprise contract data based on the associated enterprise transaction characteristic data; performing project task scheduling demand optimization strategy analysis based on the project task data to obtain project task scheduling demand optimization decisions;

[0076] In an embodiment of the present invention, enterprise contract data marked as valid are obtained from the enterprise management resource library. These data contain basic information of each contract, including key attributes such as contract number, contract signing date, customer information, and delivery date. Then, the system searches for project task data associated with each valid contract based on the associated enterprise transaction feature data. These task data usually include detailed contents such as project name, task list, executor, resource allocation, and time node related to the contract. The data collection process is implemented by executing SQL query in the project management database. The system extracts the project tasks corresponding to each contract and stores these project task data in a structured table form to ensure the accurate correspondence between each contract and its project tasks for subsequent scheduling demand optimization strategy analysis. After obtaining the project task data, the scheduling demand optimization strategy analysis is performed based on these project task data to obtain the project task scheduling demand optimization decision, and a comprehensive evaluation is performed on various constraints in the task data, including the pre-dependency of the task, resource occupancy, execution time window, etc. In order to optimize the scheduling of tasks, the system uses a heuristic algorithm (such as a genetic algorithm or a particle swarm optimization algorithm) to simulate and iterate the execution order of the tasks to find the optimal scheduling solution. The goal of optimization is to minimize the execution time of project tasks, improve resource utilization, and reduce waiting time between tasks. For example, in the optimization analysis, the system may identify that there are resource conflicts in certain tasks, and solve the problem of resource contention by rearranging the task order, thereby optimizing the overall scheduling. Ultimately, the system generates optimization decisions for project task scheduling requirements, which include adjusted task execution order, reallocated resource allocation plan, adjusted time plan, etc., to ensure that each project task can be completed efficiently and orderly.

[0077] Step S5: Based on the differentiated customer maintenance operation management decisions, enterprise contract exception feedback decisions and project task scheduling demand optimization decisions, the intelligent enterprise management operations are fed back to the terminal, wherein the intelligent enterprise management operations include enterprise customer maintenance operation management operations, enterprise contract exception feedback operations and project task scheduling demand optimization operations.

[0078] In an embodiment of the present invention, a differentiated customer maintenance operation management decision is pushed to a customer management terminal. The customer management terminal is responsible for executing customer maintenance operations, including differential contact with customers through a CRM system, corresponding maintenance management for different customers, and dynamically adjusting maintenance strategies according to customer responses to ensure that each customer group receives accurate services. The enterprise contract abnormality feedback decision is transmitted to the contract management terminal so as to perform the enterprise contract abnormality feedback operation. The contract management terminal will take specific processing measures for abnormal contracts according to the feedback decision. For example, in the case of abnormalities in the contract payment terms, the system will automatically generate a correction plan, such as suggesting adjustment of payment terms or re-negotiation of payment cycles. These suggestions are pushed to the contract person in charge through the contract management terminal, and the contract person in charge can negotiate with the customer according to the system's suggestions, or directly revise the contract terms. The system will also monitor the execution progress of the contract in real time to ensure that the revised terms of the contract can be effectively executed and reduce the risks in the performance of the contract. The project task scheduling demand optimization decision is pushed to the project management terminal to perform the project task scheduling demand optimization operation. After receiving the optimization decision, the project management terminal rearranges the project tasks according to the adjusted task execution order and resource allocation plan. The system re-plans each task through project management software (such as Microsoft Project or JIRA) and automatically notifies the relevant person in charge of the changed task plan. For example, if the execution time of a task is readjusted, the system will automatically update the execution time node of the task in the project management software and notify the relevant project members. At the same time, the project management terminal will also monitor the tasks in execution in real time to ensure that the tasks are implemented according to the optimized plan, and feedback any abnormal situations to the project management personnel for timely processing. Through intelligent decision support, differentiated customer maintenance, contract abnormality feedback and project task scheduling optimization decisions are executed on different management operation terminals, forming a complete set of enterprise management execution chain. Each management operation interacts through an integrated management platform, and through automated task scheduling, real-time feedback and dynamic adjustment, the efficient implementation of management decisions is ensured, which improves the intelligence level and management efficiency of enterprise operations.

[0079] Further, step S1 includes the following steps:

[0080] Step S11: Acquire enterprise transaction data from the enterprise management resource library;

[0081] Through the preset database connection program, you can directly access various data tables in the enterprise transaction database, including customer information, contract data, project task progress data, etc. Use database query language (such as SQL) to extract various transaction data related to enterprise operations from different database tables, and organize these data in a structured manner for subsequent processing and use. Enterprise transaction data includes text, tables, files and other forms, and efficient data collection is achieved through a unified API interface.

[0082] Step S12: Perform heterogeneous data analysis on the enterprise transaction data to generate enterprise transaction heterogeneous data;

[0083] In an embodiment of the present invention, a heterogeneous data analysis tool is used to classify and process the various types of collected data. Since enterprise transaction data may include structured data (such as numerical data in a table), semi-structured data (such as document data in XML or JSON format), and unstructured data (such as PDF contract text, etc.), it is necessary to use a data parser, a document scanner, and natural language processing (NLP) technology to identify and parse different types of data. By performing field analysis and pattern recognition on the data, a set of metadata describing the data type is generated to obtain heterogeneous type data of enterprise transactions. These heterogeneous types of data serve as labels to help the system identify and process different types of data in subsequent processing.

[0084] Step S13: extracting enterprise transaction traceability tags from enterprise transaction data to generate enterprise transaction traceability tag data;

[0085] In the embodiment of the present invention, the traceability tag is used to record the key information such as the source, generation time, and data modification history of each transaction data. Log analysis technology and data tracking tools are used to automatically extract the operation traceability information related to each piece of data through the log records and operation history in the enterprise transaction system, and mark the corresponding index tag. The generated traceability tag data is used as the basis for subsequent data management and exception tracking, which ensures the traceability of the data and facilitates audit and retrieval by management personnel.

[0086] Step S14: establishing an enterprise transaction data matrix according to the enterprise transaction heterogeneous type data, and transferring the enterprise transaction data to the enterprise transaction data matrix through the enterprise transaction traceability tag data to perform enterprise transaction data association mapping to obtain an associated enterprise transaction data matrix;

[0087] In an embodiment of the present invention, heterogeneous data of enterprise affairs are used as the dimensions of the matrix. For example, assuming that enterprise affairs involve different types of data such as customer information, contract terms, and project tasks, these types are first structured, and on this basis, different matrix dimensions are assigned to each data type. For example, customer information is the first dimension of the matrix, contract data is the second dimension, and project tasks are the third dimension. Each dimension of the matrix contains specific transaction data, and the association relationship between these data is determined by traceability tags. After the basic enterprise transaction data matrix is ​​established, the transaction data needs to be mapped to the corresponding matrix dimension through enterprise transaction traceability tag data. The traceability tag data provides a miniature tag corresponding to each transaction data, associates the relevant enterprise affairs of the enterprise management event, identifies the source and the transaction to which each data belongs through the traceability tag, and fills these data in the corresponding position of the matrix according to key fields such as contract number or customer number. For data with cross-association (such as the association between customers and contracts, and the association between contracts and project tasks), the relationship between them is determined by the traceability tag. For example, the signatory of the contract is a certain customer, and the executor of the contract is responsible for a certain project task. These associations can be established in the matrix through the traceability tag. After the mapping operation of the traceability tag data, each node in the matrix is ​​filled with specific enterprise transaction data, and the association between different types of data is clarified. In this enterprise transaction data matrix, data of different transaction dimensions can be effectively linked through the associated fields to obtain the associated enterprise transaction data matrix.

[0088] Step S15: performing correction processing on similar redundant data of associated nodes on the associated enterprise affairs data matrix to generate a corrected associated enterprise affairs data matrix;

[0089] In the embodiment of the present invention, a similarity detection algorithm (such as Jaccard similarity coefficient, cosine similarity, etc.) is used to detect the associated nodes in the matrix to identify possible duplicate data items. Subsequently, a data cleaning tool is used to merge the redundant data, remove duplicate and invalid data, ensure the uniqueness and accuracy of the enterprise transaction data, and generate a corrected associated enterprise transaction data matrix.

[0090] Step S16: Perform feature analysis processing based on the corrected associated enterprise transaction data matrix to generate associated enterprise transaction feature data.

[0091] In the embodiment of the present invention, a feature analysis algorithm is used to conduct an in-depth analysis of the corrected matrix. A data mining tool is used to perform statistical analysis on the data in the matrix to extract the main feature values ​​of each dimension of data. Based on these features, a machine learning method such as regression analysis is further used to perform pattern recognition on the data, thereby generating feature data of corporate affairs. These feature data are used to describe the characteristics and behavior patterns of corporate affairs in different dimensions, such as the behavioral characteristics of customers, the execution status of contracts, and feature data associated with corporate affairs such as project tasks.

[0092] Further, as an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S2 in FIG. 1 , in this embodiment, step S2 includes the following steps:

[0093] Step S21: extracting customer-related features according to the related enterprise transaction feature data to generate customer-related feature data, wherein the customer-related feature data includes customer attribute label data and customer behavior data;

[0094] In the embodiment of the present invention, the associated enterprise transaction feature data obtained in the previous step is analyzed to extract the feature data related to the customer. This includes identifying the customer's attribute labels (such as age, gender, occupation, consumption habits, etc.) and behavior data (such as purchase behavior, feedback behavior, service usage frequency, etc.) from a large amount of transaction data through data mining technology, using attribute recognition algorithms to convert unstructured customer-related data into a structured label format, and using behavior pattern analysis technology, such as sequence analysis, event association technology, etc., to build a customer behavior data model. These generated data are organized into customer-related feature data, including the customer's static attributes and dynamic behavior patterns, to provide a basic data set for the next step of analysis.

[0095] Step S22: Performing behavior clustering analysis on the customer behavior data to generate clustered customer behavior data;

[0096] In an embodiment of the present invention, when performing cluster analysis processing of customer behavior data, an unsupervised machine learning algorithm such as K-means or a hierarchical clustering algorithm is used to group customer behavior data. The parameters of the clustering algorithm, such as the number of clusters (K value), are determined based on prior knowledge of business needs and data distribution. For example, the K value is determined by setting various types of behavior type standards to analyze different purchasing behavior patterns. The algorithm groups customer behavior data such as purchase frequency, type of promotional activities participated in, and reaction time to generate clustered customer behavior data. This data reflects the behavioral trends of different groups of customers and provides a basis for formulating personalized customer service strategies.

[0097] Step S23: performing classified customer portrait analysis and processing based on the customer attribute label data and the clustered customer behavior data to generate classified customer portrait data;

[0098] In an embodiment of the present invention, the customer's attribute labels (such as age, gender, geographic location, industry) are combined with their behavior data, and a corresponding customer portrait is generated for each cluster group through feature matching. Portrait analysis is processed through machine learning classification algorithms (such as decision trees, logistic regression, etc.) or feature weighted models (such as weighted TF-IDF) to ensure the accuracy and interpretability of customer portraits. The classified customer portrait data not only includes the customer's basic attribute information, but also describes the customer's behavior patterns, preferences, potential needs, etc. For example, for a customer group, their portrait may show that they mainly buy high-end products and frequently browse product pages, but actually buy less, indicating that this type of customer group is price sensitive and may need to focus on pushing promotional information to increase purchase conversion rates. The analysis results generate classified customer portrait data, which is used to formulate targeted operation strategies for each customer group.

[0099] Step S24: Performing intelligent analysis of differentiated customer maintenance operation management strategies based on the clustered customer behavior data to obtain differentiated customer maintenance operation management decisions.

[0100] In the embodiment of the present invention, the stickiness analysis model of the customer group is used to calculate the stickiness indicators such as customer loyalty, purchase potential and service demand of each customer group. The stickiness index is weighted by historical behavior data (such as customer interaction frequency, purchase interval time, and after-sales problem solving) to quantify the customer's dependence and loyalty to the enterprise. The intelligent analysis tool generates differentiated maintenance strategies through optimization algorithms such as linear programming. For example, for different customer groups, the system recommends efficient maintenance methods such as personalized customization recommendations. In the entire strategy generation process, multiple dimensions such as the stickiness, behavioral characteristics, and purchase cycle of the customer group are comprehensively considered to ensure the accuracy and personalization of management decisions.

[0101] Further, step S24 includes the following steps:

[0102] Step S241: performing customer group stickiness index analysis based on the clustered customer behavior data to generate customer group stickiness index data;

[0103] In the embodiment of the present invention, the customer group here refers to different categories of customer groups obtained through cluster analysis of customer behavior data. For each clustered customer group, the stickiness index data of the customer group is generated based on its behavioral indicators such as purchase frequency, consumption amount, interaction frequency, and duration of using product services.

[0104] Step S242: setting a customer group stickiness weight parameter of the customer group stickiness index data; designing a customer group stickiness evaluation model through the customer group stickiness weight parameter and the customer group stickiness index data;

[0105] In an embodiment of the present invention, weight parameters of different stickiness indicators are determined according to the business needs and expert knowledge of the enterprise. For example, if the purchase frequency and long-term loyalty of customers are more important to the enterprise, higher weights are set for these indicators. A weight value is assigned to each stickiness indicator using the analytic hierarchy process (AHP) or the entropy method. After the weight parameters are set, a customer group stickiness evaluation model is constructed based on these weight parameters and the previously calculated customer group stickiness indicator data. The evaluation model can adopt a weighted scoring model, which calculates the comprehensive stickiness score of each customer group by multiplying the value of each stickiness indicator by the weight parameter and summing them. The higher the stickiness score of a customer group, the stronger its stickiness.

[0106] Step S243: performing personalized customer behavior prediction processing based on the clustered customer behavior data to generate personalized customer behavior prediction data;

[0107] In the embodiment of the present invention, the historical behavior data of each customer group is probabilistically modeled. Probabilistic models such as Markov chains or Bayesian networks can be used to analyze the transfer status of customer behavior, predict the possibility of future behavior, and model the behavior pattern of each customer, including the purchase behavior, interaction behavior, and the possibility of abandoning the service. By training these probabilistic models, the probability of future behavior is predicted based on historical data, and personalized customer behavior prediction data is generated, including the type of behavior that each customer may perform in a certain period of time in the future and the probability of its occurrence, such as the probability of repeat purchases, the probability of churn, etc.

[0108] Step S244: using the customer group stickiness evaluation model to perform differentiated enterprise customer stickiness feature analysis processing on the personalized customer behavior prediction data to generate differentiated enterprise customer stickiness feature data;

[0109] In the embodiment of the present invention, the customer group stickiness evaluation model has been established based on customer behavior data, and a comprehensive stickiness evaluation of customers is performed through multiple behavioral stickiness indicators (such as purchase frequency, customer retention rate, recommendation willingness, etc.) combined with corresponding weight parameters. For each customer or customer group, the evaluation model will calculate based on its historical behavior pattern and future behavior prediction data to generate customer stickiness feature data. For example, if customer group A has a high purchase frequency and has made recommendations many times in the past year, and the behavior prediction model shows that there is a high intention to repurchase in the future, the model will evaluate the stickiness of the customer group as strong, and the stickiness feature data may include labels such as "high loyalty" and "strong recommendation willingness". For customer group B, if the interaction frequency has gradually decreased over the past period of time and it is predicted that there will be a high risk of churn in the future, the evaluation model will generate low stickiness feature data, marked as "high churn risk". In the process of calculating stickiness features, the evaluation model quantifies the customer's behavioral characteristics and extracts the corresponding quantitative feature indicators to generate differentiated customer stickiness feature data, including customer loyalty score, repurchase rate, churn risk, interaction frequency, recommendation behavior tendency, etc. These differentiated stickiness feature data can be used for subsequent operational decision-making to ensure that different strategies are adopted for customers with different levels of stickiness.

[0110] Step S245: Performing intelligent analysis of differentiated customer maintenance operation management strategies based on differentiated enterprise customer stickiness feature data to obtain differentiated customer maintenance operation management decisions.

[0111] In an embodiment of the present invention, based on differentiated customer stickiness feature data, the system automatically analyzes the customer's maintenance needs and operation strategies. For example, for a customer group with a stickiness feature of "high loyalty", the system will identify their high loyalty and high willingness to recommend, and the decision may include continuing to provide personalized value-added services or recommendation incentive programs to further enhance customer experience and loyalty. On the other hand, for a customer group whose stickiness feature data is marked as "high risk of churn", the system will identify the possible reasons for churn and automatically generate targeted retention strategies. In order to optimize the implementation of the strategy, the corresponding marketing resources and operation means are automatically allocated according to the characteristics of different customer groups. Finally, through intelligent analysis, the system formulates the optimal differentiated customer maintenance operation strategy for each type of customer based on the customer stickiness feature data. These decisions will be automatically executed on the enterprise operation platform to ensure that different types of customers can obtain targeted maintenance measures, thereby improving customer satisfaction, retention rate and the overall profitability of the enterprise.

[0112] Further, step S243 includes the following steps:

[0113] Define the probability coefficient of customer group behavior type based on clustered customer behavior data and generate the probability coefficient of customer group behavior type;

[0114] Perform customer behavior frequency feature analysis based on clustered customer behavior data to generate customer behavior frequency feature data;

[0115] Calculate the personalized customer behavior conditional probability based on the customer group behavior type probability coefficient and customer behavior frequency characteristic data to generate personalized customer behavior conditional probability data;

[0116] Perform personalized customer behavior prediction processing based on personalized customer behavior conditional probability data to generate personalized customer behavior prediction data.

[0117] In an embodiment of the present invention, clustered customer behavior data is obtained and processed. These data come from the results of the previous cluster analysis, in which the behavior characteristics of different customer groups are relatively consistent. For each customer group, the process of defining the probability coefficient of the customer group behavior type includes identifying the behavior pattern of the customer group and setting a corresponding probability coefficient for each behavior pattern. For example, common behaviors such as purchase behavior, interaction behavior, and recommendation behavior of a certain group can be analyzed. Assume that a customer group has the following types of behaviors: frequent purchases, high-frequency interactions, low-frequency recommendations, etc. Through historical data analysis, the probability of occurrence of each behavior type in the group can be determined. For example, the probability of purchase behavior of a certain customer group is 0.7, the probability of interaction behavior is 0.5, and the probability of recommendation behavior is 0.2. The probability coefficient of each behavior type is generated by statistical analysis of historical behavior data, and these coefficients will be used to represent the possibility of customer groups engaging in these behaviors in the future. The process of generating the probability coefficient of customer group behavior type is based on the behavior frequency in the historical data, and defines the type probability of different customer behaviors through statistical and probability analysis tools, such as statistical model libraries in Python (such as Scipy, Pandas, etc.). Detailed feature analysis is performed on the behavior frequency of each customer group. Specific operations include counting the number of occurrences of each behavior of the customer group within a certain period of time. Behavior frequency feature data indicates the frequency of a customer performing a specific behavior in multiple time periods. For example, if we analyze the purchase behavior of a customer group in the past six months, the purchase frequency of each month is 5 times, 4 times, 6 times, 3 times, 2 times, and 5 times respectively. These data can be used to calculate the frequency characteristics of customer behavior, including the average value, variance, and fluctuation of behavior frequency. Further feature analysis can be combined with different behavior types for comprehensive analysis, such as interaction frequency, website visit frequency, etc. The generated customer behavior frequency feature data will make a quantitative analysis of each customer's behavior in the time dimension. Use data processing tools (such as SQL or Python data processing libraries) for analysis, extract customer behavior frequency information, generate customer behavior frequency feature data, and form a visual report to intuitively display customer behavior frequency characteristics. Combine the probability coefficient of customer group behavior type with the customer behavior frequency feature data and calculate it through the conditional probability formula. The conditional probability here refers to the probability that a customer will perform a certain behavior (such as recommending others to buy) under a certain condition (such as when the purchase frequency is high). For example, assuming that the probability of a customer group's purchase behavior is 0.7, and the frequency characteristic of the recommendation behavior is once a month, the conditional probability of their recommendation behavior when the purchase frequency of this customer group is high can be calculated based on the conditional probability formula. If there is a certain logical dependency between the behaviors, for example, a customer's high purchase frequency is usually accompanied by a high recommendation frequency, the model can be further optimized to capture this dependency.Through the above process, the mathematical model of conditional probability is used for calculation to generate personalized customer behavior conditional probability data, which can help predict the probability of a specific customer performing certain behaviors under specific conditions in the future. Tools such as Bayesian formula, Markov chain model or statsmodels library in Python can be used for these calculations. The conditional probability data is input into the behavior prediction model, and the model infers the probability of each customer performing certain behaviors in the future based on these conditional probability data. For example, based on a customer's purchase frequency characteristics, behavior type probability and its conditional probability, the system can predict the customer's purchase probability, churn probability and recommendation behavior probability in the next month. Assuming that the conditional probability of a customer's purchase behavior is 0.6 and the conditional probability of the recommended behavior is 0.4, the system infers that the probability of the customer continuing to purchase in the future is 60%, and the probability of the recommended behavior is 40%. In addition, it is combined with supervised learning of machine learning, which can be implemented using Python's machine learning library (such as Scikit-learn) to perform supervised learning prediction on the overall probability prediction behavior. Supervised learning can automatically obtain the prediction trend of personalized customer behavior and output personalized customer behavior prediction data.

[0118] Further, as an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S3 in the embodiment, step S3 includes the following steps:

[0119] Step S31: collecting and classifying the corresponding enterprise contract data in the customer portrait data according to the associated enterprise transaction feature data;

[0120] In an embodiment of the present invention, characteristic data related to enterprise affairs are obtained from the enterprise management resource library. These data are collected through the enterprise's transaction processing system, including customer transaction records, contract information, project task data, etc. Based on these enterprise affairs characteristic data, enterprise contract data corresponding to the classified customer portrait is screened. This process ensures that the acquired data matches the customer portrait by setting specific contract screening conditions, such as attributes such as customer category, contract amount, and contract term. The classified customer portrait data here is generated through preliminary analysis, including information such as the customer's basic attributes, behavioral characteristics, and purchasing preferences. The collection of contract data is automatically completed through structured query language (SQL) or enterprise data management tools to ensure that the enterprise contract data can correspond one-to-one with the customer portrait data to form an association relationship.

[0121] Step S32: performing enterprise contract parsing processing on the enterprise contract data based on a preset semantic parsing model to generate enterprise contract parsing data; performing contract key clause attribute identification based on the enterprise contract parsing data to generate contract key clause attribute data;

[0122] In an embodiment of the present invention, a semantic parsing model based on natural language processing (NLP) is used to parse the acquired enterprise contract text data one by one, and the contract data is converted into a processable text format, such as converting PDF or other unstructured data formats into plain text. Then, a rule-based parsing model or a deep learning model (such as a BERT model) is used to segment and semantically understand the contract text, and parse the important terms and key information in the contract. The key contents of the analysis include the validity period, payment terms, liability for breach of contract, delivery time, etc. of the contract. Contract parsing data is generated by extracting keywords from the contract terms. These parsed data are annotated with predefined tags, such as "payment terms", "performance conditions", etc., to form structured contract key terms attribute data for subsequent processing.

[0123] Step S33: performing contract clause rule matching exception marking processing based on the contract key clause attribute data, and generating a contract key clause attribute exception node;

[0124] In an embodiment of the present invention, the system extracts relevant rule data from a contract terms rule knowledge base. The knowledge base stores preset contract terms rule sets, which are formulated based on industry standards, corporate practices, and legal requirements. For example, the payment period in the contract terms shall not exceed 30 days, the delivery date shall not be earlier than the contract signing date, etc. Then, the system compares and analyzes the parsed contract key terms attribute data with the terms in the rule knowledge base to identify anomalies. For example, if the payment period of a contract exceeds the preset rule threshold, the system will automatically mark the term as an abnormal node. The system executes this process through a precise rule matching algorithm (such as a condition-based Boolean logic matching algorithm) to generate a contract key terms attribute abnormal node containing abnormal information.

[0125] Step S34: Performing a feasibility analysis on the execution of a contract event based on the attribute data of the key contract terms, and generating feasibility data on the execution of the contract event;

[0126] In an embodiment of the present invention, historical data related to contract execution is obtained, including performance records of previous contracts, event progress, delivery status, etc. These historical data are stored in the contract execution database as the basis for analysis. Then, the system analyzes the logical relationship of contract performance based on the attribute data of key contract terms. For example, whether the delivery period and payment time involved in the contract match the actual project schedule. The system infers the terms and event dependencies in the contract by executing the logical relationship analysis model, and generates logical relationship data for the execution of contract events. Based on these logical relationships, the system further uses historical execution data to evaluate the feasibility of contract terms and generate contract event execution feasibility data. If the execution feasibility of a certain term is lower than the preset threshold, the system marks the term as infeasible or risky.

[0127] Step S35: when an abnormal node of a key contract clause attribute is identified or the contract event execution feasibility data is less than a preset contract event execution feasibility threshold, the enterprise contract data is marked as abnormal enterprise contract data; otherwise, the enterprise contract data is marked as valid enterprise contract data;

[0128] In the embodiment of the present invention, a comprehensive analysis is performed from the abnormal node information of the key terms of the contract obtained in the previous step and the feasibility data of the execution of the contract event. For each contract, the system detects whether there are nodes marked as abnormal in the key terms of the contract. For example, if the payment terms or delivery period terms of a contract contain abnormal nodes, it means that these terms do not match the enterprise standards or rules, and this contract has potential risks. At the same time, the system will also detect the feasibility data of the execution of the contract event. If the execution feasibility of the contract is lower than the preset threshold, for example, the threshold is set to 70%, and the execution feasibility of a contract is only 60%, it is considered that the contract may have a high risk of failure during the performance process. When any key terms of a contract are marked as abnormal, or the feasibility of the execution of the contract event is lower than the preset threshold, the system will automatically mark the contract as an "abnormal contract". This marking is achieved through a Boolean field in the contract database, for example, adding an "abnormal" attribute to the contract data, and the value is "True". The system automatically performs the marking process and classifies all contracts that meet the abnormal conditions into the "abnormal contract" list, which is convenient for further management and processing. If all key terms of the contract meet the standards and the feasibility of contract event execution is higher than the threshold, the contract is marked as a "valid contract". The marking process is achieved by modifying the "abnormal" attribute of the contract data to "False".

[0129] Step S36: Analyze the enterprise contract abnormality feedback strategy through the abnormal nodes of the contract key clause attributes corresponding to the abnormal enterprise contract data and the contract event execution feasibility data to obtain the enterprise contract abnormality feedback decision.

[0130] In an embodiment of the present invention, a comprehensive analysis is performed on the enterprise contract data that has been marked as abnormal, and all abnormal nodes and their corresponding abnormal information are summarized. The summary of abnormal nodes includes the abnormal types of contract clause attributes, such as payment terms overdue, delivery conditions not feasible, etc. Each abnormal node is marked with the nature of the abnormality and the degree of impact. For example, if the payment period of a certain clause exceeds the 30-day limit of the enterprise standard, its abnormal nature is "term exceeded" and the degree of impact is "high". Next, the system uses the contract event execution feasibility data for in-depth analysis. For example, the system focuses on analyzing those contract events with low execution feasibility and determines the key factors affecting the execution of the event, such as resource shortages, dependent task delays, etc. Then, the system combines the type of abnormal nodes and the feasibility issues of contract events to analyze the enterprise contract abnormality feedback strategy. The analysis process is executed through a rule engine (such as a Drools rule engine) or a decision tree model, and the system recommends corresponding processing strategies for each type of abnormality. For example, if the delivery time of a contract is abnormally delayed, and the feasibility analysis results of the contract show that resource shortage is the main factor, the feedback strategy generated by the system may include: increasing manpower or material investment, renegotiating with suppliers to ensure resource supply, etc. If the payment terms of a contract have a significant impact, the system may suggest renegotiating payment terms with the customer or increasing incentives for early payment to reduce the pressure on the company's capital turnover. The generated abnormal feedback strategy is output in the form of a report, including a detailed description of the problem, abnormal type, impact, and recommended feedback measures for each abnormal contract. The report is pushed to the contract management team through the enterprise management system so that the abnormal contract can be corrected and optimized. The entire analysis and strategy generation process is implemented through automated tools to ensure that the company can respond to contract abnormalities in a timely manner, reduce risks in contract performance, and ensure the stability of corporate operations.

[0131] Further, step S33 includes the following steps:

[0132] Acquire a contract terms and rules knowledge base; perform contract terms and rules matching anomaly analysis on the contract key terms and rules attribute data according to the contract terms and rules knowledge base; when it is identified that the contract key terms and rules attribute data do not match the contract terms and rules knowledge base, mark the corresponding contract key terms and rules attribute data as a contract key terms and rules attribute anomaly node.

[0133] In an embodiment of the present invention, contract terms and rules are obtained from a contract terms and rules knowledge base. The contract terms and rules knowledge base is a pre-established database that stores standardized rules and requirements for various terms in a contract. The knowledge base includes contract terms standards of laws and regulations, industry practices, and the company's own management regulations, with the goal of ensuring that all contracts comply with corporate and legal standards. The contract terms and rules knowledge base is stored and managed through a relational database (such as MySQL or PostgreSQL), and the system uses SQL statements to perform query operations to extract the required terms and rules. Each rule is labeled, including information such as the rule name, applicable scenario, and threshold, to facilitate subsequent matching. For example, for payment terms, the rule knowledge base stipulates that the payment period shall not exceed 30 days, and corresponding interest shall be paid for deferred payment, etc. All these rules are recorded in the knowledge base in a structured manner. After completing the acquisition of the contract terms and rules, the system matches and analyzes the contract key terms attribute data according to these rules. The contract key terms and attributes data are generated by the previous semantic parsing model and include important terms in the contract, such as payment terms, delivery date, breach of contract terms, etc. The system compares the attributes of each contract clause one by one by using an automated rule matching tool (such as the Drools rule engine) or a conditional judgment module based on Boolean logic. For example, for a payment clause in a contract, the attribute data shows that the payment period is 45 days, while the maximum payment period specified in the knowledge base is 30 days. The system will automatically determine that the clause does not comply with the rules and is an abnormal clause. During the matching process, the system compares the rules of each key clause of the contract one by one, and makes conditional judgments based on the actual attribute values ​​of the contract clauses and the rules in the knowledge base. If there is a deviation between the clause attribute data and the rules in the knowledge base, the clause is considered to be non-compliant with the standard regulations. During the matching process, a rule engine based on Boolean logic is applied to determine whether it complies with the preset clause rules. For each clause, the rule engine inputs the attribute value in the contract into the judgment module. If the judgment result is False, it means that the clause does not comply with the preset rules. When the attribute data of a contract clause does not match the preset rules in the contract clause rule knowledge base, the system will mark the clause as abnormal and record the abnormal attributes of the clause, such as the type of clause, attribute value, and content that does not match the rules. Then, the system marks it as an "abnormal node of the key contract clause attribute". This marking process is automatically completed through data annotation tools, such as adding a Boolean field of "abnormal tag" to the abnormal clause in the database, or adding an additional "abnormal" tag to the abnormal node in the XML or JSON format of the contract parsing data for subsequent processing. For example, when the delivery terms of a contract show that the delivery time is 60 days after the contract is signed, but the rule knowledge base stipulates that the delivery time shall not exceed 45 days, the system will mark the delivery terms as an abnormal node.During the marking process, the system will also generate an exception report, which details the abnormal situation of the clause and the specific rule information that does not match, so as to facilitate subsequent decision-making and contract amendment. The entire exception marking process is automatically executed without manual intervention, thus improving recognition efficiency and accuracy.

[0134] Further, step S34 includes the following steps:

[0135] Obtain the historical contract execution event library; perform contract execution logic relationship analysis on the key contract clause attribute data to obtain contract execution logic relationship data;

[0136] Perform contract event execution reasoning analysis based on contract execution logic relationship data to generate contract event execution reasoning data;

[0137] The contract event execution feasibility analysis is performed on the contract event execution reasoning data according to the historical contract execution event library to generate the contract event execution feasibility data.

[0138] In an embodiment of the present invention, a historical contract execution event library is obtained. The historical contract execution event library is a database containing all the past contract execution conditions of an enterprise. The database records the detailed performance information of each contract, such as the signing time, delivery time, payment progress, performance, breach of contract handling, delivery delay, etc. of the contract. The process of obtaining the event library uses a database query tool (such as an SQL query statement) to extract relevant historical data from the enterprise contract management system. The historical contract execution event library is stored in the form of a relational database (such as MySQL, PostgreSQL, etc.). The system uses structured query language (SQL) to perform data retrieval and obtains a series of information data related to contract performance, including contract execution status, execution events, execution time nodes, etc. The acquired historical data provides a basis for subsequent execution logic relationship analysis and feasibility evaluation. After completing the acquisition of the historical contract execution event library, the system performs a contract execution logic relationship analysis on the contract key terms attribute data to obtain the contract execution logic relationship data. The contract key terms attribute data includes important contents of the contract such as delivery period, payment terms, breach of contract terms, etc. The system needs to analyze the logical dependency relationship in its execution process based on these terms information. For example, in a certain contract, there is a clear logical dependency relationship between the delivery time node and the payment trigger condition - usually the payment condition can only be met after the delivery is completed and accepted. These clause data are analyzed through the preset logical relationship model. The model arranges the various clauses of the contract according to the execution before and after relationship by establishing an "event-dependency" structure to form a directed graph structure. Each node represents a contract event, and the directed edges between the nodes represent the dependency relationship of the events. For example, if delivery is a prerequisite for payment, there is a directed edge pointing to payment between the delivery node and the payment node. In this way, the system finally generates contract execution logical relationship data, which represents the dependency order and logical relationship of each key clause in the contract execution process. After the contract execution logical relationship data is obtained, the contract event execution reasoning analysis is further performed, and the contract execution logical relationship data is input into the reasoning engine for analysis. The reasoning engine can use a rule reasoning system (such as the Drools rule engine) or a reasoning model based on causal analysis to perform in-depth analysis of the logical relationship between each contract event. For example, if the delivery event and payment of a contract are reasoned and analyzed, if the delivery event is delayed, the system will infer that the corresponding payment time may also be delayed. The inference engine reads the logical relationship of contract execution, analyzes the dependency paths between events in turn, and infers the execution results and impacts under different circumstances. For example, if the delivery time node is postponed, the possible impact on the subsequent payment node and the execution of the breach of contract clause will be inferred to form the contract event execution inference data. The inference data is usually stored in the form of a tree structure or flowchart, which is convenient for visualizing the causal chain of contract events during the execution process.Following up on the inference data of contract event execution requires further feasibility analysis of contract event execution. The historical data in the historical contract execution event library is used to evaluate the feasibility of the inferred events. The purpose of feasibility evaluation is to determine the success probability of contract events in actual execution and the risks they may face. The system analyzes by matching events in the inference data with similar cases in the historical event library. For example, the system can find the same delivery conditions and payment arrangements in similar contracts in the past, evaluate the success rate and execution of these events, and especially the bottlenecks and problems encountered during their execution. In order to quantify feasibility, statistical analysis tools are used to count historical contract events, such as calculating the average execution time, success rate, and failure reasons of contract events under similar conditions, to generate quantitative indicators for contract event execution. Finally, the system compares these indicators with the inference data of the current contract event to obtain an execution feasibility score. Contract event execution feasibility data usually includes feasibility scores, potential execution risks, key factors for success and failure, etc. The generated feasibility data is displayed through the enterprise's intelligent management platform for decision makers to refer to, to help identify high-risk points in the contract execution process.

[0139] Further, as an embodiment of the present invention, refer to Figure 4 As shown, Figure 1 Detailed step flow diagram of step S4 in FIG. 1 , in this embodiment, step S4 includes the following steps:

[0140] Step S41: collecting and monitoring project task data corresponding to valid enterprise contract data based on the associated enterprise transaction feature data;

[0141] In an embodiment of the present invention, characteristic data of associated enterprise transactions are obtained from the enterprise management resource library. These characteristic data include relevant information of enterprise contracts, such as contract number, customer information, delivery content, signing time, etc. Then, based on these characteristic data, the system automatically queries and matches the corresponding project task data. For each valid enterprise contract, the system will obtain the corresponding project task list. These tasks include specific delivery activities, executors, time nodes, etc. The data collection process is carried out in the project management database through SQL query statements to obtain all relevant project tasks corresponding to the contract. Each task is marked with a unique identifier of the contract to ensure that the task data and contract data can be effectively associated. The result of this process is to establish a corresponding relationship between each valid contract and all its related project tasks, and save these task data to the project task data set for subsequent analysis and scheduling optimization.

[0142] Step S42: performing a project task scheduling requirement dependency analysis according to the project task data to generate project task scheduling requirement dependency analysis data, wherein the project task scheduling requirement dependency analysis data includes the project task scheduling requirement pre-dependency analysis data, scheduling requirement resource dependency analysis data, and scheduling requirement timing dependency analysis data;

[0143] In the embodiment of the present invention, after completing the project task data collection, the system performs a project task scheduling demand dependency analysis on these task data. The analysis process is divided into three parts: scheduling demand pre-dependency analysis, scheduling demand resource dependency analysis, and scheduling demand timing dependency analysis. First, the pre-dependency of the project task is analyzed to clarify the execution order of each task and its pre- and post-dependency relationship. For example, delivery A needs to be executed after task B is completed. This dependency is modeled through a directed graph to form pre-dependency data. Next, the system performs a resource dependency analysis on the data of the project task, and generates resource dependency data by identifying the resource types (such as manpower, materials, equipment) required for each task and the occupancy of the resources. Finally, the scheduling demand timing dependency analysis is performed to form timing dependency data by determining the earliest start time, the latest completion time, and the time constraints between tasks for each task. The entire analysis is implemented through graph theory algorithms and scheduling theory, and finally generates comprehensive project task scheduling demand dependency analysis data.

[0144] Step S43: performing a weight evaluation analysis of the project task scheduling requirements based on the project task scheduling requirements dependency analysis data to design a weight parameter of the project task scheduling requirements;

[0145] In an embodiment of the present invention, the number of dependencies of each task in the pre-dependency relationship is counted, and the more dependencies a task has, the higher its priority. Then, with respect to the resource dependency relationship, the key resources required for the task are evaluated, and tasks with tight key resources are assigned higher scheduling weights. Finally, in the timing dependency relationship, the time urgency of the task is also considered as an important factor. All these factors are comprehensively evaluated through the Analytic Hierarchy Process (AHP), and a weight parameter is assigned to each task. This weight parameter characterizes the importance and priority of the task in the overall scheduling, so that the needs of these high-weight tasks can be met first in the scheduling optimization process.

[0146] Step S44: performing a project task scheduling response characteristic analysis on the project task data according to the project task scheduling requirement dependency analysis data to generate project task scheduling response characteristic data;

[0147] In an embodiment of the present invention, the response characteristics of each task in the scheduling process are analyzed based on the dependency data. The schedulability of each task in the time dimension is evaluated, such as the task's start time window, execution time, and impact on other tasks. Next, the system analyzes the response characteristics in the resource dimension, including the difficulty of resource allocation, substitutability, etc. For example, if a task requires a specific device and resources are in short supply, the system will mark the task as "difficult to respond to resource scheduling." The results of the response characteristic analysis include the task's time response level, resource response level, and the degree of impact on other tasks. These characteristics are stored in the form of structured data to generate project task scheduling response characteristic data for subsequent deviation quantification analysis.

[0148] Step S45: performing a quantitative analysis of the project task scheduling deviation on the project task scheduling response characteristic data to generate quantitative data of the project task scheduling deviation;

[0149] In an embodiment of the present invention, the planned scheduling and the actual scheduling are compared to calculate the deviations of each task in terms of time, resources, etc. The time deviation includes the difference between the actual start time of the task and the planned start time, and the difference between the actual completion time and the planned completion time. The resource deviation includes the difference between the resource occupancy time and the actual demand time. The system generates deviation data by quantifying these deviations. For example, a task is scheduled to start on Monday, but is postponed to Wednesday due to unavailable resources. The time deviation is two days. The deviation quantification process is implemented through data analysis tools, and the deviation quantification data is used to measure the execution effect of the current scheduling strategy and the problems and bottlenecks in the scheduling.

[0150] Step S46: Performing a project task scheduling requirement optimization strategy analysis based on the project task scheduling requirement weight parameter and the project task scheduling deviation quantification data to obtain a project task scheduling requirement optimization decision.

[0151] In an embodiment of the present invention, cluster analysis is performed on the deviation data to identify the task types or resource bottlenecks that have more deviations in the scheduling process. Then, the scheduling of high-weight tasks is adjusted preferentially according to the weight parameters, for example, reallocating resources, modifying the time window for task execution, etc. The system uses a scheduling optimization model based on a heuristic algorithm (such as a genetic algorithm, an ant colony optimization algorithm) to solve the optimal task scheduling solution. The goal of the optimization process is to minimize scheduling deviations, improve resource utilization, and meet the dependencies between tasks, generate an optimized scheduling demand strategy, and provide a rescheduling and resource allocation plan for project tasks to ensure that project tasks can be completed efficiently and orderly. The optimization decision results will be fed back to the enterprise intelligent management platform for project managers to execute.

[0152] This specification provides an enterprise intelligent management system based on artificial intelligence, which is used to execute the enterprise intelligent management method based on artificial intelligence as described above. The enterprise intelligent management system based on artificial intelligence includes:

[0153] The associated enterprise affairs feature analysis module is used to obtain enterprise affairs data from the enterprise management resource library; perform feature analysis and processing on associated enterprise affairs based on the enterprise affairs data to generate associated enterprise affairs feature data;

[0154] The customer management module is used to extract customer-related features based on the characteristic data of related enterprise transactions to generate customer-related feature data; to analyze and process classified customer portraits based on the customer-related feature data to generate classified customer portrait data; to perform intelligent analysis of differentiated customer maintenance operation management strategies based on clustered customer behavior data to obtain differentiated customer maintenance operation management decisions;

[0155] The contract management module is used to collect and classify the corresponding enterprise contract data in the customer portrait data according to the characteristic data of the associated enterprise affairs; perform enterprise contract anomaly analysis based on the enterprise contract data, and generate abnormal enterprise contract data and valid enterprise contract data respectively, wherein the abnormal enterprise contract data includes abnormal nodes of key contract terms and contract event execution feasibility data; perform enterprise contract anomaly feedback strategy analysis through the abnormal nodes of key contract terms and contract event execution feasibility data corresponding to the abnormal enterprise contract data, so as to obtain enterprise contract anomaly feedback decision;

[0156] The project task management module is used to collect and monitor the project task data corresponding to the effective enterprise contract data based on the characteristic data of the associated enterprise affairs; analyze the project task scheduling demand optimization strategy based on the project task data to obtain the project task scheduling demand optimization decision;

[0157] The enterprise intelligent management module is used to execute intelligent enterprise management operations based on differentiated customer maintenance operation management decisions, enterprise contract exception feedback decisions, and project task scheduling demand optimization decision feedback to the terminal, wherein the intelligent enterprise management operations include enterprise customer maintenance operation management operations, enterprise contract exception feedback operations, and project task scheduling demand optimization operations.

[0158] The beneficial effect of this application is that the enterprise transaction data is obtained through the enterprise management resource library, and the multi-dimensional feature analysis and processing of these data are performed to automatically generate the characteristic data of related enterprise transactions. The whole process utilizes the data-driven artificial intelligence model, which can quickly and efficiently complete the processing and analysis of data, replacing the traditional manual method, thereby greatly improving the overall management efficiency of the enterprise, and the enterprise management becomes more refined and accurate, reducing redundant work and improving the response speed of decision-making. Through the analysis of the characteristic data of related enterprise transactions, the behavioral characteristics of customers are extracted, and then customer portraits are classified. Based on the cluster analysis of these portraits and customer behaviors, enterprises can better understand the needs and behavior patterns of different types of customers, and predict and classify them through customer stickiness evaluation models, which provides support for enterprises to formulate differentiated customer maintenance strategies. The intelligent customer management model enables enterprises to accurately identify customer groups, optimize resource allocation, and improve customer satisfaction and loyalty, thereby occupying a more advantageous position in market competition. In the process of enterprise contract data parsing, a contract clause rule matching mechanism based on semantic analysis is adopted, which can timely discover abnormal nodes of key clauses in the contract and perform intelligent reasoning analysis on the feasibility of contract execution. Once potential problems in contract execution are discovered, the system will provide feedback strategy suggestions so that the enterprise can take remedial measures, improve the accuracy of contract management, prevent contract disputes, and reduce the legal and financial risks of the enterprise. By using the characteristic data of related enterprise affairs and contract data, the dependencies, scheduling requirements and response characteristics of project tasks are analyzed. Through intelligent analysis of multi-dimensional weight evaluation and deviation quantification, an optimized project task scheduling strategy is designed, which improves the flexibility and adaptability of project management and reduces resource waste and time delays during project implementation. Through automatic scheduling optimization of artificial intelligence, enterprises can ensure that projects are completed efficiently under limited resources and time conditions and maximize business goals. Through artificial intelligence technology, a closed-loop management system with data as the core is constructed, involving all aspects of corporate affairs, customer management, contract management and project scheduling. The management content of multiple dimensions is highly integrated in the same system. The management decision of each link is fed back to the next stage through the analysis results of the previous stage, and connected with the terminal execution system to ensure the closed-loop implementation and continuous optimization of the decision, which can ensure the real-time and flexibility of enterprise management. It can also continuously improve and improve the operation of each link through the feedback mechanism, and promote the gradual transformation of enterprises to intelligent and data-driven. Therefore, the enterprise method based on artificial intelligence of the present invention ensures the high efficiency of customer maintenance management and the accuracy of customer needs by designing intelligent decisions for personalized customer maintenance operations for different users. By identifying abnormal clauses and executable risks in enterprise contracts through intelligent analysis means, the accuracy and security of contract management are improved, and the execution ability and management efficiency of enterprise contracts are guaranteed.Through multi-dimensional analysis of the dependencies of project task data, the scheduling requirements of project tasks are comprehensively evaluated, and according to the importance of scheduling requirements, the resource scheduling requirements of project tasks are intelligently adjusted to improve the follow-up execution efficiency of project tasks. Cross-domain data association analysis and decision optimization have achieved efficient and intelligent enterprise management, and integrated analysis has greatly reduced the island effect between various management systems. It has a high degree of adaptability and can flexibly respond to the management needs of enterprises of different sizes and types.

[0159] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0160] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. An enterprise intelligent management method based on artificial intelligence, characterized in that: The following steps are involved: Step S1: Acquire enterprise transaction data from the enterprise management resource library; Based on the enterprise transaction data, characteristic analysis and processing of the related enterprise transactions are performed to generate characteristic data of the related enterprise transactions; wherein step S1 includes: Step S11: Acquire enterprise transaction data from the enterprise management resource library; Step S12: Perform heterogeneous data analysis on the enterprise transaction data to generate enterprise transaction heterogeneous data; Step S13: extracting enterprise transaction traceability tags from enterprise transaction data to generate enterprise transaction traceability tag data; Step S14: establishing an enterprise transaction data matrix according to the enterprise transaction heterogeneous type data, and transferring the enterprise transaction data to the enterprise transaction data matrix through the enterprise transaction traceability tag data to perform enterprise transaction data association mapping to obtain an associated enterprise transaction data matrix; Step S15: performing correction processing on similar redundant data of associated nodes on the associated enterprise affairs data matrix to generate a corrected associated enterprise affairs data matrix; Step S16: Perform feature analysis processing according to the corrected associated enterprise transaction data matrix to generate associated enterprise transaction feature data; Step S2: extract customer-related features based on the associated enterprise transaction feature data to generate customer-related feature data; perform classified customer portrait analysis and processing based on the customer-related feature data to generate classified customer portrait data; perform differentiated customer maintenance operation management strategy intelligent analysis based on clustered customer behavior data to obtain differentiated customer maintenance operation management decisions; wherein step S2 includes: Step S21: extracting customer-related features based on the associated enterprise transaction feature data to generate customer-related feature data, wherein the customer-related feature data includes customer attribute label data and customer behavior data; wherein the associated enterprise transaction feature data is analyzed to extract feature data related to the customer, including identifying the customer's attribute label and behavior data from a large amount of transaction data through data mining technology, using an attribute recognition algorithm to convert unstructured customer-related data into a structured label format, and using a behavior pattern analysis technology to construct a customer behavior data model, and the generated data is organized into customer-related feature data, including the customer's static attributes and dynamic behavior patterns; Step S22: Performing behavior clustering analysis on the customer behavior data to generate clustered customer behavior data; Step S23: performing classified customer portrait analysis and processing based on the customer attribute label data and the clustered customer behavior data to generate classified customer portrait data; Step S24: Performing intelligent analysis of differentiated customer maintenance operation management strategies based on clustered customer behavior data to obtain differentiated customer maintenance operation management decisions; wherein step S24 includes: Step S241: performing customer group stickiness index analysis based on the clustered customer behavior data to generate customer group stickiness index data; Step S242: setting a customer group stickiness weight parameter of the customer group stickiness index data; designing a customer group stickiness evaluation model through the customer group stickiness weight parameter and the customer group stickiness index data; Step S243: performing personalized customer behavior prediction processing based on the clustered customer behavior data to generate personalized customer behavior prediction data; wherein step S243 includes: Define the probability coefficient of customer group behavior type based on clustered customer behavior data and generate the probability coefficient of customer group behavior type; Perform customer behavior frequency feature analysis based on clustered customer behavior data to generate customer behavior frequency feature data; Calculate the personalized customer behavior conditional probability based on the customer group behavior type probability coefficient and customer behavior frequency characteristic data to generate personalized customer behavior conditional probability data; Perform personalized customer behavior prediction processing based on personalized customer behavior conditional probability data to generate personalized customer behavior prediction data; Step S244: using the customer group stickiness evaluation model to perform differentiated enterprise customer stickiness feature analysis processing on the personalized customer behavior prediction data to generate differentiated enterprise customer stickiness feature data; Step S245: performing intelligent analysis of differentiated customer maintenance operation management strategies based on differentiated enterprise customer stickiness feature data to obtain differentiated customer maintenance operation management decisions; Step S3: Collect and classify the corresponding enterprise contract data in the customer portrait data according to the characteristic data of the associated enterprise transactions; perform enterprise contract anomaly analysis based on the enterprise contract data, and generate abnormal enterprise contract data and valid enterprise contract data respectively, wherein the abnormal enterprise contract data includes abnormal nodes of contract key terms attributes and contract event execution feasibility data; perform enterprise contract anomaly feedback strategy analysis through abnormal nodes of contract key terms attributes corresponding to the abnormal enterprise contract data and contract event execution feasibility data to obtain enterprise contract anomaly feedback decision; wherein step S3 includes: Step S31: collecting and classifying the corresponding enterprise contract data in the customer portrait data according to the associated enterprise transaction feature data; Step S32: performing enterprise contract parsing processing on the enterprise contract data based on a preset semantic parsing model to generate enterprise contract parsing data; performing contract key clause attribute identification based on the enterprise contract parsing data to generate contract key clause attribute data; Step S33: performing contract clause rule matching exception marking processing based on the contract key clause attribute data, and generating a contract key clause attribute exception node; Step S34: Performing a feasibility analysis on the execution of contract events based on the attribute data of the key contract terms, and generating feasibility data on the execution of contract events; wherein step S34 includes: Obtain the historical contract execution event library; perform contract execution logic relationship analysis on the key contract clause attribute data to obtain contract execution logic relationship data; Perform contract event execution reasoning analysis based on contract execution logic relationship data to generate contract event execution reasoning data; Conduct contract event execution feasibility analysis on contract event execution reasoning data based on the historical contract execution event database to generate contract event execution feasibility data; Step S35: when an abnormal node of a key contract clause attribute is identified or the contract event execution feasibility data is less than a preset contract event execution feasibility threshold, the enterprise contract data is marked as abnormal enterprise contract data; otherwise, the enterprise contract data is marked as valid enterprise contract data; Step S36: analyzing the enterprise contract abnormality feedback strategy through the abnormal nodes of the contract key clause attributes corresponding to the abnormal enterprise contract data and the contract event execution feasibility data to obtain the enterprise contract abnormality feedback decision; Step S4: collecting and monitoring project task data corresponding to effective enterprise contract data based on the associated enterprise transaction characteristic data; performing project task scheduling demand optimization strategy analysis based on the project task data to obtain project task scheduling demand optimization decisions; Step S5: Based on the differentiated customer maintenance operation management decisions, enterprise contract exception feedback decisions and project task scheduling demand optimization decisions, the intelligent enterprise management operations are fed back to the terminal, wherein the intelligent enterprise management operations include enterprise customer maintenance operation management operations, enterprise contract exception feedback operations and project task scheduling demand optimization operations.

2. The enterprise intelligent management method based on artificial intelligence according to claim 1 is characterized in that: Step S33 includes the following steps: Acquire a contract terms and rules knowledge base; perform contract terms and rules matching anomaly analysis on the contract key terms and rules attribute data according to the contract terms and rules knowledge base; when it is identified that the contract key terms and rules attribute data do not match the contract terms and rules knowledge base, mark the corresponding contract key terms and rules attribute data as a contract key terms and rules attribute anomaly node.

3. The enterprise intelligent management method based on artificial intelligence according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: collecting and monitoring project task data corresponding to valid enterprise contract data based on the associated enterprise transaction feature data; Step S42: performing a project task scheduling requirement dependency analysis according to the project task data to generate project task scheduling requirement dependency analysis data, wherein the project task scheduling requirement dependency analysis data includes the project task scheduling requirement pre-dependency analysis data, scheduling requirement resource dependency analysis data, and scheduling requirement timing dependency analysis data; Step S43: performing a weight evaluation analysis of the project task scheduling requirements based on the project task scheduling requirements dependency analysis data to design a weight parameter of the project task scheduling requirements; Step S44: performing a project task scheduling response characteristic analysis on the project task data according to the project task scheduling requirement dependency analysis data to generate project task scheduling response characteristic data; Step S45: performing a quantitative analysis of the project task scheduling deviation on the project task scheduling response characteristic data to generate quantitative data of the project task scheduling deviation; Step S46: Performing a project task scheduling requirement optimization strategy analysis based on the project task scheduling requirement weight parameter and the project task scheduling deviation quantification data to obtain a project task scheduling requirement optimization decision.

4. An enterprise intelligent management system based on artificial intelligence, characterized in that: Used to execute the enterprise intelligent management method based on artificial intelligence as claimed in claim 1, the enterprise intelligent management system based on artificial intelligence includes: The associated enterprise affairs feature analysis module is used to obtain enterprise affairs data from the enterprise management resource library; perform feature analysis and processing on associated enterprise affairs based on the enterprise affairs data to generate associated enterprise affairs feature data; The customer management module is used to extract customer-related features based on the characteristic data of related enterprise transactions to generate customer-related feature data; to analyze and process classified customer portraits based on the customer-related feature data to generate classified customer portrait data; to perform intelligent analysis of differentiated customer maintenance operation management strategies based on clustered customer behavior data to obtain differentiated customer maintenance operation management decisions; The contract management module is used to collect and classify the corresponding enterprise contract data in the customer portrait data according to the characteristic data of the associated enterprise affairs; perform enterprise contract anomaly analysis based on the enterprise contract data, and generate abnormal enterprise contract data and valid enterprise contract data respectively, wherein the abnormal enterprise contract data includes abnormal nodes of key contract terms and contract event execution feasibility data; perform enterprise contract anomaly feedback strategy analysis through the abnormal nodes of key contract terms and contract event execution feasibility data corresponding to the abnormal enterprise contract data, so as to obtain enterprise contract anomaly feedback decision; The project task management module is used to collect and monitor the project task data corresponding to the effective enterprise contract data based on the characteristic data of the associated enterprise affairs; analyze the project task scheduling demand optimization strategy based on the project task data to obtain the project task scheduling demand optimization decision; The enterprise intelligent management module is used to execute intelligent enterprise management operations based on differentiated customer maintenance operation management decisions, enterprise contract exception feedback decisions, and project task scheduling demand optimization decision feedback to the terminal, wherein the intelligent enterprise management operations include enterprise customer maintenance operation management operations, enterprise contract exception feedback operations, and project task scheduling demand optimization operations.

Images