Enterprise internal control digital management method based on AI large model

Abstract

The invention discloses an enterprise internal control digital management method based on an AI large model, relates to the technical field of enterprise data management, and is used for solving the problems that in the prior art, enterprise data is difficult to supervise when being abnormal, and the data interaction security between enterprises is poor. According to the method, an internal control data warehouse is created for each department in an enterprise, enterprise operation data of each department is obtained and subjected to digital processing, and digital data to be controlled of each enterprise is obtained and stored in each internal control data warehouse; the method comprises the following steps: constructing an AI large model for supervising enterprise operation data corresponding to all departments of an enterprise, analyzing to obtain abnormal data in the enterprise corresponding to each department based on the AI large model, carrying out internal management and control, and when data interaction is carried out between different enterprises, judging whether the two enterprises have direct handover authority or not; and selecting internal associated data interaction or external transfer data interaction based on a judgment result, and managing and controlling the interactive transfer data of the two enterprises of data interaction in the whole process.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise data management, and specifically to a digital management method for enterprise internal control based on an AI large model. Background Art

[0002] An AI large model refers to a deep learning model with a huge number of parameters, usually reaching the hundreds of millions or even trillions, trained based on massive amounts of data, and capable of realizing multi-task and multi-modal intelligent analysis and generation through a complex neural network architecture.

[0003] In existing enterprise management, the operation data corresponding to each department lacks effective supervision, and its security is difficult to guarantee. The discovery of problem data is not timely and accurate enough. The emergence of the AI large model provides an effective means for data supervision. How to achieve data control of different departments under each enterprise through the AI large model, and how to supervise the data interaction between different enterprises to ensure the safe progress of data interaction are problems that need to be solved urgently at present. Summary of the Invention

[0004] In order to solve the above problems, the purpose of the present invention is to provide a digital management method for enterprise internal control based on an AI large model.

[0005] The purpose of the present invention can be achieved through the following technical solutions: A digital management method for enterprise internal control based on an AI large model, including the following steps: Step S1: Create an internal control data warehouse for each department within the enterprise, obtain the enterprise operation data corresponding to each department, and perform digital processing on the enterprise operation data to obtain the digital data to be controlled corresponding to each department, and store it in its respective internal control data warehouse; Step S2: Construct an AI large model for supervising the enterprise operation data corresponding to all departments of the enterprise, analyze and obtain the abnormal data within each department corresponding to the enterprise based on the AI large model, and perform internal control on the abnormal data; Step S3: When data interaction occurs between different enterprises, judge whether there is a direct handover permission between the enterprises, select internal associated data interaction or external transfer data interaction based on the judgment result, and perform full-process control on the interaction transfer data of the two enterprises for data interaction.

[0006] Further, the process of creating an internal control data warehouse for each department within the enterprise includes: Number all the departments within the enterprise in sequence and denote them as i, i = 1, 2, 3,..., n, where n is a natural number greater than 0. Each department under the enterprise is associated with a corresponding authorization identification code, and the authorization identification code is used to judge the authorization status of the corresponding department; Build a data access layer, a data cleaning rule library, and a permission unit; The authorization status includes "authorized" and "unauthorized"; When a certain department of an enterprise is in the "authorized" state, an instruction for creating a database application carrying the total data storage volume corresponding to the current department is initiated by this department and transmitted to the data access layer. According to the total data storage volume of the data under the corresponding department, a corresponding memory space is created in the data access layer to construct a preliminary data warehouse for the corresponding enterprise; The preliminary data warehouse consists of several data blocks. Set the monitoring period for each data block, clean the environmental information of each data block under the monitoring period through the data cleaning rule library, perform the permission allocation operation of the permission unit on the data warehouse, allocate the internal control permissions for data interaction between different departments of the enterprise, associate the internal control permissions to the preliminary data warehouse corresponding to the corresponding department, and create the internal control data warehouse corresponding to each department; When a certain department of an enterprise is in the "unauthorized" state, no operation is performed.

[0007] Furthermore, the process of obtaining the enterprise operation data corresponding to each department, digitally processing the enterprise operation data to obtain the digitally controlled data corresponding to each department, and storing it in the internal control data warehouse of each department includes: Set the data entry time period for each department of the enterprise. Each department obtains its own enterprise operation data. The enterprise operation data includes the financial data, transaction data, and inventory data of the finance department, the document data, image data, and audio - video data of the sales department, the equipment operation data, business indicator data, and server log data corresponding to the technical operation and maintenance department, and also includes the customer data and supplier data corresponding to the operation department; Configure digital processing tools. The digital processing tools digitally process the enterprise operation data of each department of the enterprise according to the configured data processing format, data processing rules, and verification rules to obtain the digitally controlled data corresponding to each department. Store the digitally controlled data of each department in the internal control data warehouse constructed corresponding to each department, and monitor the storage environment. When the storage environment is abnormal, stop the storage of the digitally controlled data corresponding to the corresponding department.

[0008] Furthermore, the process of building an AI large - model for supervising the enterprise operation data corresponding to all departments of the enterprise includes: Build a hierarchical model architecture corresponding to the AI large - model for supervising the enterprise operation data of each department. The specific composition of the hierarchical model architecture includes a multi - source data input layer, a modality encoding layer, a cross - modality alignment layer, a domain knowledge injection layer, a task decision layer, and a feedback optimization layer; Model fusion is carried out according to the hierarchical model architecture to construct the final AI large model for supervising the enterprise operation data of each department in the enterprise, obtain the historical operation data corresponding to each department, divide the training set, validation set and test set, and then train the AI large model until the model prediction rate of the AI large model meets the preset compliance threshold.

[0009] Furthermore, the process of analyzing the abnormal data in each department's corresponding enterprise based on the AI large model and performing internal control on the abnormal data includes: The AI large model after construction analyzes the digital data to be controlled in each department in turn. If abnormal data is found in the digital data to be controlled in a certain department, the storage path and abnormal items of the abnormal data are further located. An internal control unit is constructed to perform internal control on the abnormal data located in each department in turn. The internal control includes traversing the data from the root node of the abnormal data in the storage path, marking the abnormal positions of all the traversed abnormal items on the storage path, creating an internal control sub-node at each abnormal position on the storage path, editing an abnormal repair script, and injecting it into each internal control sub-node on the storage path in a coded manner. The internal control sub-node modifies the corresponding abnormal items at each abnormal position according to the abnormal repair script. If the modification of the abnormal items cannot be completed, the corresponding abnormal items are excluded.

[0010] Furthermore, when data interaction occurs between different enterprises, the process of judging whether there is a direct transfer permission between the enterprises and selecting internal associated data interaction or external transfer data interaction based on the judgment result includes: When two enterprises conduct data interaction, each of the two enterprises constructs an interaction request. An interaction rule matching library is constructed. The interaction rule matching library stores several enterprise pairing relationship groups. The enterprise pairing relationship group is used to include two enterprises with business interaction requirements and the data interaction between the respective departments of the two enterprises. There is a direct transfer permission between the two enterprises in the enterprise pairing relationship group, and the interaction requests of the two enterprises in the enterprise pairing relationship group are associated in the interaction rule matching library. When it is judged that the two enterprises conducting data interaction have a direct transfer permission, internal associated data interaction is performed on the data interaction of the two enterprises. When the two enterprises do not have a direct transfer permission, external transfer data interaction is performed on the data interaction between the two enterprises.

[0011] Furthermore, internal associated data interaction and external transfer data interaction each include: The internal associated data interaction is as follows: construct an internal data transfer channel between enterprises. When data interaction occurs between departments of the same type under different enterprises, the departments directly conduct data interaction through the internal data transfer channel; The external transfer data interaction is as follows: construct an intermediate cache component within the internal data transfer channel. When data interaction occurs between departments of different types under two enterprises, each department transfers the relevant data that needs to be interacted to the intermediate cache component. The intermediate cache component serves as the data transfer party to conduct data interaction on the relevant data of the two enterprises respectively, and saves all the data of the data interaction at the intermediate cache component.

[0012] Furthermore, the process of fully controlling the interaction transfer data of the two enterprises in data interaction includes: When two enterprises conduct data interaction, map the interaction transfer data synchronously generated during data interaction to a pre-constructed control chain, construct a transfer time axis for the transfer process of the interaction transfer data between the two enterprises. The transfer time axis includes several transfer time nodes; Mark the time stamps of the interaction transfer data under each transfer time node, analyze whether there are abnormal data transfer operations under each time stamp. If so, suspend the data interaction between the two enterprises, analyze the reasons for the abnormal data transfer operations, and obtain the troubleshooting measures corresponding to the abnormal data transfer operations. If not, do not perform any operations.

[0013] Compared with the prior art, the beneficial effects of the present invention are: by creating corresponding internal control data warehouses for each department within the enterprise, digitizing the enterprise operation data of each department to obtain their respective digital data to be controlled and storing them in their respective internal control data warehouses, constructing an AI large model for supervising the enterprise operation data, analyzing and obtaining the abnormal data of each department within the enterprise, and conducting internal control. When data interaction occurs between different enterprises, judge whether there is a direct transfer permission between the enterprises, select to conduct internal associated data interaction or external transfer data interaction based on the judgment result, and fully control the interaction transfer data, solving the problems that it is difficult to supervise the enterprise data when it is abnormal and the data interaction security between enterprises is poor. Brief Description of the Drawings

[0014] Figure 1 It is a flow chart of the present invention. Detailed Embodiment

[0015] As Figure 1 shown, an enterprise internal control digital management method based on an AI large model includes the following steps: Step S1: Create an internal control data warehouse for each department within the enterprise, obtain the enterprise operation data corresponding to each department, perform digital processing on the enterprise operation data to obtain the digital data to be controlled corresponding to each department, and store it in their respective internal control data warehouses; Step S2: Build an AI large model for supervising the enterprise operation data corresponding to all departments of the enterprise. Based on the analysis of the AI large model, obtain the abnormal data within the enterprise corresponding to each department, and conduct internal control on the abnormal data; Step S3: When data is exchanged between different enterprises, determine whether there is a direct handover permission between the enterprises. Based on the judgment result, select to conduct internal associated data exchange or external transfer data exchange, and conduct full-process control on the interactive transfer data of the two enterprises involved in the data exchange.

[0016] It should be further noted that in the specific implementation process, the process of creating an internal control data warehouse for each department within the enterprise includes: Number all the departments within the enterprise in sequence, and record the number as i, where i = 1, 2, 3,..., n, and n is a natural number greater than 0. Each department under the enterprise is associated with a corresponding authorization identification code, and the authorization identification code is used to judge the authorization status of the corresponding department; Build a data access layer, a data cleaning rule library, and a permission unit; Starting from the label i = 1, traverse the authorization identification codes of each department in sequence. According to the authorization identification code, obtain whether the corresponding department belongs to a preset permission form. If the authorization identification code is within the permission form, judge the authorization status of the corresponding department as "authorized"; if the authorization identification code is not within the permission form, judge the authorization status of the corresponding department as "unauthorized"; When a certain department under the enterprise is in the "authorized" authorization status, the department initiates a database creation application instruction carrying the total data storage volume corresponding to the current department, and transmits the database creation application instruction to the data access layer. After receiving the database creation application instruction, the data access layer creates a corresponding memory space in the data access layer according to the total data storage volume of the data corresponding to the corresponding department, and then constructs a preliminary data warehouse corresponding to the enterprise; The preliminary data warehouse consists of several data blocks. Set the monitoring period corresponding to each data block, clean the environmental information of each data block under the monitoring period through the data cleaning rule library, screen out and eliminate the alienation interference information corresponding to each data block. The existence of the alienation interference information will have a negative impact on the storage environment of the corresponding data block and interfere with the normal progress of the data storage work; After eliminating the alienation interference information of all data blocks, continue to perform the permission allocation operation of the permission unit on the data warehouse, allocate the internal control permissions for data interaction between different departments under the enterprise, associate the internal control permissions to the corresponding preliminary data warehouses of the respective departments, and then create the internal control data warehouses corresponding to each department; The internal control permission is interpreted as follows: when data is interacted between different departments under the enterprise, the interaction channel for data interaction is within the internal interaction channels preset by the enterprise, that is, data interaction between departments is carried out through the internal interaction channels; When a certain department under the enterprise is in the "unauthorized" authorization state, no operation is performed.

[0017] It should be further noted that in the specific implementation process, obtaining the enterprise operation data corresponding to each department, digitally processing the enterprise operation data to obtain the digital data to be controlled corresponding to each department, and storing it in the respective internal control data warehouses includes: Set the data entry time periods corresponding to each department under the enterprise. During their respective data entry time periods, each department obtains its corresponding enterprise operation data. The enterprise operation data includes the financial data, transaction data, and inventory data corresponding to the finance department, the document data, image data, and audio - video data of the sales department, the equipment operation data, business indicator data, and server log data corresponding to the technical operation and maintenance department, and also includes the customer data and supplier data corresponding to the operation department; The descriptions of the data of the finance department are as follows: The financial data includes, but is not limited to, accounting vouchers (debit / credit), general ledger details, and balance sheets; the transaction data includes, but is not limited to, sales orders (SKU / quantity / amount), purchase contracts (terms / delivery date), and transaction contracts (transaction date / transaction party); the inventory data includes, but is not limited to, real - time inventory levels (warehouse / shelf level), turnover rate, and safety inventory thresholds; The descriptions of the data of the sales department are as follows: The document data includes, but is not limited to, PDF contracts, Word - based technical agreements, and Excel quotes; the image data includes, but is not limited to, product design drawings, quality inspection photos, and sales vouchers; the audio - video data includes, but is not limited to, customer service call recordings and production site monitoring videos; The descriptions of the data of the technical operation and maintenance department are as follows: The equipment operation data includes, but is not limited to, equipment operating temperature, equipment vibration frequency, and equipment working energy consumption; the business indicator data includes, but is not limited to, real - time website UV, real - time website PV, and overall equipment efficiency; the server log data includes, but is not limited to, the concurrent data volume of server access and server response time; The descriptions of various data of the operation department are as follows: The customer data includes, but is not limited to, basic information such as customer classification (Level A, Level B, and Level C), credit limit, and historical transaction volume, and also includes behavioral data such as website click-through rate and customer service consultation subject. The supplier data includes, but is not limited to, the qualified rate of product quality delivered, on-time delivery rate, and cost competitiveness rate of the delivered products; Configure digital processing tools, and the digital processing tools perform digital processing on the enterprise operation data corresponding to each department of the enterprise according to the configured data processing format, data processing rules, and verification rules, so as to obtain the digital data to be controlled corresponding to each department; Store the digital data to be controlled of each department into the internal control data warehouse constructed respectively, and monitor the storage environment. When the storage environment is abnormal, stop the storage of the digital data to be controlled corresponding to the corresponding department.

[0018] It should be further noted that in the specific implementation process, the process of constructing an AI large model for supervising the enterprise operation data corresponding to all departments of the enterprise includes: Construct a hierarchical model architecture corresponding to the AI large model for supervising the enterprise operation data of each department. The specific composition of the hierarchical model architecture includes a multi-source data input layer, a modality encoding layer, a cross-modal alignment layer, a domain knowledge injection layer, a task decision layer, and a feedback optimization layer; The multi-source data input layer is used for inputting the digital data to be controlled corresponding to the internal control data warehouse of each department; The modality encoding layer consists of a text encoder, a time series encoder, and a graph structure encoder; The text encoder performs long text modeling on the digital data to be controlled of each department; The time series encoder analyzes the time series dependence relationship of the digital data to be controlled of each department; The graph structure encoder analyzes the graph structure of the digital data to be controlled of each department; The cross-modal alignment layer is used to set cross-modal interaction interfaces for the modality encoding layers corresponding to different departments respectively, and then align the modality encoding layers corresponding to each two departments for data interaction to the same dimension level to complete the data interaction between the corresponding departments; The domain knowledge injection layer is used to inject the domain knowledge graph corresponding to each department. The domain knowledge graph is used as the supervision index corresponding to each department, and the unqualified data within each department is located through the supervision index; The task decision layer is used to make decision analysis on the data anomaly level corresponding to each department; Construct activation functions and loss functions by the feedback optimization layer, train data for other architecture layers in the hierarchical model architecture except itself, and stop data training after each architecture layer reaches its respective preset architecture metrics to complete the optimization of each architecture layer. Perform model fusion according to the hierarchical model architecture, and then construct an AI large model that finally monitors the enterprise operation data of each department in the enterprise, obtain the historical operation data corresponding to each department, divide the training set, validation set, and test set, and then train the AI large model until the model prediction rate of the AI large model meets the preset standard threshold.

[0019] It should be further noted that in the specific implementation process, the process of analyzing the abnormal data corresponding to each department in the enterprise based on the AI large model and performing internal control on the abnormal data includes: The constructed AI large model analyzes the digital data to be controlled for each department in turn. If abnormal data is found in the digital data to be controlled for a certain department, further locate the storage path and abnormal items of the abnormal data. Construct an internal control unit, and the internal control unit performs internal control on the abnormal data located for each department in turn. The internal control includes traversing the data from the root node of the abnormal data in the storage path, marking the abnormal positions of all abnormal items obtained by the traversal on the storage path, creating an internal control sub-node at each abnormal position on the storage path, editing an abnormal repair script, and injecting the abnormal repair script code into each internal control sub-node on the storage path. The internal control sub-node modifies the corresponding abnormal item at each abnormal position according to the abnormal repair script. If the modification of the abnormal item cannot be completed, the corresponding abnormal item is excluded.

[0020] It should be further noted that in the specific implementation process, when data is exchanged between different enterprises, the process of judging whether there is a direct handover permission between enterprises and selecting internal associated data exchange or external transfer data exchange based on the judgment result includes: When two enterprises exchange data, each of the two enterprises constructs an interaction request. Construct an interaction rule matching library, which stores several enterprise pairing relationship groups. The enterprise pairing relationship group is used to include two enterprises with business interaction requirements and the data exchange between their respective departments under the two enterprises. There is a direct handover permission between the two enterprises under the enterprise pairing relationship group, and the respective interaction requests of the two enterprises under the enterprise pairing relationship group are associated with requests in the interaction rule matching library. The enterprise pairing relationship group is used to pair the relationships between the same departments under different enterprises. When it is determined that there is a direct handover permission between two enterprises conducting data interaction, internal associated data interaction is performed on the data interaction between the two enterprises. When it is determined that there is no direct handover permission between two enterprises conducting data interaction, external transfer data interaction is performed on the data interaction between the two enterprises; The content of the internal associated data interaction is: constructing an internal data transfer channel between enterprises. When data interaction occurs between departments of the same type under different enterprises, the departments directly conduct data interaction through the internal data transfer channel; The content of the external transfer data interaction is: constructing an intermediate cache component within the internal data transfer channel. When data interaction occurs between departments of different types under two enterprises, each department transfers the relevant data that needs to be interacted to the intermediate cache component. The intermediate cache component serves as the data transfer party to conduct data interaction on the relevant data of the two enterprises respectively, and saves all the data of the data interaction at the intermediate cache component.

[0021] It should be further noted that in the specific implementation process, the process of fully controlling the interaction transfer data of two enterprises conducting data interaction includes: When data interaction occurs between two enterprises, the interaction transfer data synchronously generated during the data interaction is mapped to a pre-constructed control chain, and a corresponding transfer time axis for the transfer process of the interaction transfer data between the two enterprises is constructed in the control chain; The transfer time axis includes several transfer time nodes; Mark the time stamps of the interaction transfer data under each transfer time node, and analyze whether there are data transfer abnormal operations under each time stamp. The data transfer abnormal operations include data throughput abnormality, data interruption abnormality, and data transmission delay abnormality; If so, suspend the data interaction between the two enterprises, analyze the reasons for the data transfer abnormal operations, construct a corresponding data log form, upload the data log form to a pre-constructed troubleshooting database, and obtain the troubleshooting measures corresponding to the data transfer abnormal operations between the enterprises currently conducting data interaction from the troubleshooting database until the full control of each time stamp of the interaction transfer data under the data transfer abnormal operations is completed; If not, no operation is performed.

[0022] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A digital management method for enterprise internal control based on AI big model, characterized by: The following steps are involved: Step S1: Create an internal control data warehouse for each department in the enterprise, obtain the enterprise operation data corresponding to each department, and digitize the enterprise operation data to obtain the digital data to be controlled corresponding to each department, and store them in their respective internal control data warehouses; Step S2: Build an AI big model to supervise the enterprise operation data of all departments of the enterprise, analyze the abnormal data of each department in the enterprise based on the AI ​​big model, and perform internal control on the abnormal data; Step S3: When data is exchanged between different enterprises, determine whether there is direct handover authority between the enterprises, and choose to conduct internal related data interaction or external transit data interaction based on the judgment result, and perform full-process management and control of the interactive flow data of the two enterprises interacting with the data.

2. According to the method of digital management of enterprise internal control based on AI big model in claim 1, it is characterized by: The process of creating an internal control data warehouse for each department within the enterprise includes: All departments in the enterprise are numbered in sequence and recorded as i, i = 1, 2, 3, ..., n, where n is a natural number greater than 0. Each department under the enterprise is associated with a corresponding authorization identification code, which is used to determine the authorization status of the corresponding department; Build data access layer, data cleaning rule base and permission unit; The authorization status includes "authorized" and "unauthorized"; When a department of an enterprise is in the "authorized" state, the department initiates a database construction application instruction with the total data storage of the current department and transmits it to the data access layer. According to the total data storage of the data under the corresponding department, the corresponding memory space is created in the data access layer to build a preliminary data warehouse for the corresponding enterprise. The preliminary data warehouse consists of several data blocks. The monitoring cycle of each data block is set. The environmental information of each data block in the monitoring cycle is cleaned through the data cleaning rule library. The permission allocation operation of the permission unit is performed on the data warehouse. The internal control permissions for data interaction between different departments of the enterprise are allocated, and the internal control permissions are associated with the preliminary data warehouse corresponding to the corresponding department. The internal control data warehouse corresponding to each department is created. When a department under an enterprise is in "unauthorized" status, no operation will be performed.

3. According to claim 2, a digital management method for internal control of an enterprise based on an AI big model is characterized in that: The process of obtaining the enterprise operation data corresponding to each department, digitizing the enterprise operation data, obtaining the digital data to be controlled corresponding to each department, and storing them in their respective internal control data warehouses includes: Set the data entry period for each department under the enterprise, and each department obtains its own enterprise operation data, which includes the financial data, transaction data and inventory data of the finance department, the document data, image data and audio and video data of the sales department, the equipment operation data, business indicator data and server log data of the technical operation and maintenance department, and the customer data and supplier data of the operation department; Configure digital processing tools, and use them to digitally process the enterprise operation data of each department under the enterprise according to the configured data processing format, data processing rules and verification rules to obtain the digital data to be controlled corresponding to each department. The digital data to be controlled of each department is stored in the corresponding internal control data warehouse, and the storage environment is monitored. When an abnormality occurs in the storage environment, the storage of the digital data to be controlled corresponding to the corresponding department is stopped.

4. According to claim 3, a digital management method for internal control of an enterprise based on an AI big model is characterized in that: The process of building an AI big model to supervise the enterprise operation data of all departments of the enterprise includes: Construct a hierarchical model architecture corresponding to the AI ​​big model that supervises the enterprise operation data of each department. The specific components of the hierarchical model architecture include multi-source data input layer, modal encoding layer, cross-modal alignment layer, domain knowledge injection layer, task decision layer and feedback optimization layer; Model fusion is performed according to the hierarchical model architecture to build an AI big model that ultimately supervises the enterprise operation data of each department under the enterprise. The historical operation data corresponding to each department is obtained, and the AI ​​big model is trained after dividing the data into training sets, validation sets, and test sets until the model prediction rate of the AI ​​big model meets the preset threshold.

5. According to claim 4, a digital management method for internal control of an enterprise based on an AI big model is characterized in that: The process of analyzing abnormal data in each department of the enterprise based on the AI ​​big model and conducting internal control of abnormal data includes: The constructed AI model analyzes the digital data to be controlled in each department in turn. If the analysis shows that there are abnormal data in the digital data to be controlled in a certain department, the storage path and abnormal items of the abnormal data are further located; Construct an internal control unit and perform internal control on the abnormal data located by each department in turn. The internal control includes traversing the data from the root node of the abnormal data in the storage path, marking the abnormal positions of all abnormal items traversed on the storage path, creating an internal control sub-node at each abnormal position on the storage path, editing the abnormal repair script, and injecting it into each internal control sub-node on the storage path in the form of code. The internal control sub-node executes the modification of the corresponding abnormal item at each abnormal position according to the abnormal repair script. If the modification of the abnormal item cannot be completed, the corresponding abnormal item will be eliminated.

6. According to claim 5, a digital management method for internal control of an enterprise based on an AI big model is characterized in that: When different enterprises interact with each other, the process of determining whether there is direct handover authority between the enterprises and selecting internal related data interaction or external transfer data interaction based on the determination result includes: When two enterprises interact with each other, they each construct an interaction request. Construct an interaction rule matching library, which stores a number of enterprise pairing relationship groups. The enterprise pairing relationship group is used to include two enterprises with business interaction needs, and data interaction between the respective departments of the two enterprises; There is direct handover authority between the two enterprises in the enterprise pairing relationship group, and the interaction requests of the two enterprises in the enterprise pairing relationship group are associated in the interaction rule matching library; When it is determined that two enterprises performing data interaction have direct handover authority, internal associated data interaction is performed for the data interaction between the two enterprises. When the two enterprises do not have direct handover authority, external transit data interaction is performed for the data interaction between the two enterprises.

7. According to claim 6, a digital management method for internal control of an enterprise based on an AI big model is characterized in that: Internal related data interaction and external transit data interaction each include: Internal related data interaction is to build internal data flow channels between enterprises. When departments of the same type in different enterprises interact with each other, the departments interact with each other directly through the internal data flow channels. External transit data interaction is: build an intermediate cache component in the internal data flow channel. When different types of departments under two enterprises interact with each other, each department will transfer the relevant data that needs to be interacted with to the intermediate cache component. The intermediate cache component acts as a data transfer party to interact with the relevant data of the two enterprises and save all the data of the data interaction at the intermediate cache component.

8. According to claim 7, a digital management method for internal control of an enterprise based on an AI big model is characterized in that: The process of full control over the interactive data flow between two enterprises includes: When two enterprises interact with each other, the interactive flow data generated synchronously during the data interaction is mapped to the pre-built control chain, and a flow timeline is constructed for the flow process of the interactive flow data between the two enterprises. The flow timeline includes several flow time nodes. Mark the timestamp of the interactive flow data at each flow time node, analyze whether there is any abnormal data flow operation at each timestamp, if so, suspend the data interaction between the two enterprises, analyze the cause of the abnormal data flow operation, and obtain the corresponding troubleshooting measures for the abnormal data flow operation, if not, do not perform any operation.

Images