Evaluation account making and checking method based on AI Agent digital employee

Abstract

The invention discloses an AI Agent-based digital employee valuation account making and reconciliation method, which comprises the following steps of: dividing digital employees deployed on a virtual machine into a digital check post, a daily end clearing post and a reconciliation post according to an valuation account making and reconciliation process; and an AI Agent digital employee is constructed, the AI Agent digital employee manages and distributes a digital check post, a daily end clearing post and an account checking post, and the whole process of automatic value estimation, account making and account checking is completed. According to the method, the efficiency and quality problems in the valuation account making process can be effectively solved, the AI Agent robot can dynamically and reasonably distribute product batches, the AI Agent robot is a planner and a manager of the whole work process, the digital employee robot of each post can complete valuation account making work more accurately and efficiently, and the work efficiency is improved. Operators concentrate on exception handling and management of the business process, time and labor cost are saved for enterprises, and accuracy and normalization of valuation accounting work are improved.

Description

Technical Field

[0001] The present invention relates to the field of valuation and liquidation, and in particular to a valuation and accounting reconciliation method based on AI Agent digital employees. Background Art

[0002] With the continuous increase in the volume of fund asset custody and outsourced valuation business, it is becoming increasingly difficult to complete all valuation and liquidation business accurately and on time with limited manpower. Introducing more professional securities accounting talents also has problems such as high cost, lack of talent, and low efficiency. In the actual manual valuation operation of fund assets, there are factors such as batch copying, fuzzy batch running, running again after failed running, no quality inspection, inability to accurately locate the specific product execution steps and status, and no detailed comparison process for custody outsourcing. There is an overall lack of quality control, and it is often necessary to wait until the running fails to find the cause. The time cost of re-running is large, which affects the valuation output ratio and timeliness. Summary of the invention

[0003] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and provide a valuation accounting method based on AI Agent digital employees.

[0004] The purpose of the present invention is achieved through the following technical solutions:

[0005] A valuation and accounting reconciliation method based on AI Agent digital employees, comprising the following steps:

[0006] S1. According to the valuation and reconciliation process, the digital employees deployed on the virtual machines are divided into digital inspection posts, end-of-day settlement posts, and reconciliation posts;

[0007] Among them, the digital inspection post is used for the on- and off-site transaction data inspection link; the end-of-day settlement post is used for the accounting link; the reconciliation post is used for the statement data inspection link, internal reconciliation link, and custody outsourcing reconciliation link;

[0008] S2. Build AI Agent digital employees, who manage and assign digital inspection posts, end-of-day settlement posts, and reconciliation posts, completing the entire process of automated valuation and reconciliation;

[0009] The AI ​​Agent digital employee processing logic is as follows:

[0010] Store the historical valuation processing time of several fund product pool products. According to the concurrent processing capacity of the Hang Seng system, it is divided into four levels: fast account, relatively fast account, slow account, and relatively slow account, and the corresponding number of concurrent accounts for each level is set;

[0011] AI Agent plans the products to be processed every day, dynamically groups them according to product processing time, self-evaluates historical grouping data and processing efficiency before grouping, self-optimizes grouping planning, and plans the best planning combination for the day;

[0012] AI Agent monitors the status of digital employees on virtual machines, whether they are idle or working; when it monitors that a digital employee on a virtual machine is idle, AI Agent digital employees will automatically assign batches to the idle digital employees.

[0013] The calculation process of the number of products that should be allocated to the four levels of the AI ​​Agent digital employee in the dynamic grouping is as follows:

[0014] First, calculate the number of products that the four levels can theoretically process per unit time, set as R fast , R faster , R slow , R slower , the calculation formula is as follows:

[0015] R fast =C fast ×T;

[0016] R faster =C faster ×T;

[0017] R slow =C slow ×T;

[0018] R slower =C slower ×T;

[0019] Among them, C fast , C faster , C slow , C slower They are the concurrency numbers corresponding to the four levels of fast account, relatively fast account, slow account, and relatively slow account; T represents the total processing time;

[0020] Then, according to the historical grouping data and processing efficiency, the proportion of each level of product quantity is self-evaluated and optimized to obtain the proportion of the product quantity of four levels: fast account, relatively fast account, slow account, and relatively slow account. fast , P faster , P slow , P slower ;

[0021] Finally, calculate the actual number of products that should be allocated to each level, set as n fast 、n faster 、n slow 、n slower , the calculation formula is as follows:

[0022] n fast =N×P fast ;

[0023] n faster =N×P faster ;

[0024] n slow =N×P slow ;

[0025] n slower =N×P slower ;

[0026] Where N represents the total number of products that need to be processed.

[0027] The AI ​​Agent takes into account the fluctuation of processing time, and the calculation process of the actual number of products to be allocated to the four levels is as follows:

[0028] (1) Calculate the minimum value R of the number of products processed per unit time for the four levels min-fast , R min-faster , R min-slow , R min-slower and the maximum value R max-fast , R max-faser , R max-slow , R max-slower :

[0029]

[0030] Among them, t fast-min ,t fast-max is the minimum and maximum processing time of the quick account product, t faster-min ,t faster-max is the minimum and maximum processing time of the faster account product, t slow-min ,t slow-max The minimum and maximum processing time of slow account products, t slower-min ,t slower-max Minimum and maximum processing times for slower account products;

[0031] (2) Combined with the historical processing efficiency weights, calculate the number of products processed per unit time R for the four levels com-fast , R com-faster , R com-slow , R com-slower :

[0032]

[0033] Among them, W fast , W faster , W slow , Wslower The processing efficiency weights for the four levels are obtained based on historical data;

[0034] (3) Recalculate product quantity allocation:

[0035] The proportion of new products in four levels com-fast , P com-faster , P com-slow , P com-slower Calculated as follows:

[0036]

[0037] (4) The final number of products in the four grades n com-fast 、n com-faster 、n com-slow 、n com-slower The calculation is as follows:

[0038] n com-fast =N×P com-fast ;

[0039] n com-faster =N×P com-faster ;

[0040] n com-slow =N×P com-slow ;

[0041] n com-slower =N×P com-slower .

[0042] After calculating the actual number of products to be allocated to the four levels, the AI ​​Agent groups the products according to the calculated number of products, or groups them according to the processing time, or groups them in combination with the product type, or adjusts the groups in real time.

[0043] At the same time, the present invention provides:

[0044] A server comprises a processor and a memory, wherein the memory stores at least one program, and the program is loaded and executed by the processor to implement the valuation and accounting method based on AI Agent digital employees.

[0045] A computer-readable storage medium, wherein at least one program is stored in the storage medium, and the program is loaded and executed by a processor to implement the valuation and accounting method based on AI Agent digital employees.

[0046] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0047] 1. The present invention can effectively solve the efficiency and quality problems in the valuation and accounting process by constructing a valuation and accounting method based on AI Agent digital employees. The AI ​​Agent robot can dynamically and reasonably allocate product batches and is the planner and manager of the entire workflow. The digital employee robots in various positions can complete the valuation and accounting work more accurately and efficiently. The operators can focus on the exception handling and management of the business process, saving time and labor costs for the enterprise and improving the accuracy and standardization of the valuation and accounting work.

[0048] 2. The present invention uses AI Agent-type digital employees to automate valuations, which can greatly ensure the accuracy, stability and timeliness of daily valuations of all current products, and ensure that the daily valuation of each fund asset is traceable, manageable and controllable. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 Flowchart for reconciliation of accounts for valuation.

[0050] Figure 2 Schematic diagram of batch processing time.

[0051] Figure 3 Flowchart for automatic valuation and accounting reconciliation of digital employees on virtual machines.

[0052] Figure 4 Schematic diagram of implementing batch processing time for digital workers on virtual machines.

[0053] Figure 5 This is the AI ​​Agent digital employee architecture diagram.

[0054] Figure 6 Schematic diagram of the work of AI Agent digital employees. DETAILED DESCRIPTION

[0055] The present invention is further described in detail below in conjunction with embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0056] Valuation and reconciliation process Figure 1 As shown, it mainly includes five major links: on-site and off-site transaction data inspection, accounting link, statement data inspection, internal reconciliation, and custody outsourced reconciliation. Among them, the accounting link includes five parts: public type detection, day-end settlement, voucher generation, valuation table generation and error prevention type detection.

[0057] Valuation accounting is done in the Hang Seng system in batches, with each batch containing concurrent accounting for multiple products. Only when each product is accounted for can the current batch end and the next batch be processed. This rule also applies to each node in the process. Figure 2shown.

[0058] like Figures 3 to 6 The present invention provides a valuation and accounting reconciliation method based on AI Agent digital employees, comprising the following steps:

[0059] S1. According to the valuation and reconciliation process, the digital employees deployed on the virtual machines are divided into digital inspection posts, end-of-day settlement posts, and reconciliation posts;

[0060] Among them, the digital inspection post is used for the on- and off-site transaction data inspection link; the end-of-day settlement post is used for the accounting link; the reconciliation post is used for the statement data inspection link, internal reconciliation link, and custody outsourcing reconciliation link;

[0061] S2. Build AI Agent digital employees, who manage and assign digital inspection posts, end-of-day settlement posts, and reconciliation posts, completing the entire process of automated valuation and reconciliation;

[0062] like Figure 3 The daily valuation and reconciliation are carried out by digital employees in three positions: data inspection, end-of-day settlement, and reconciliation. The daily cut operation will be carried out at 2:30 a.m. every trading day. The daily cut operation will initialize the data to be processed in the resource pool, perform operation detection on the resource pool robot, synchronize the valuation and accounting system product accounting number and other key operations.

[0063] The specific implementation process of the three robots: digital inspection, end-of-day settlement, and reconciliation:

[0064] a) Data inspection post: Data inspection mainly involves file-type data such as product transaction data and information-type data such as statement fund data.

[0065] File data is checked by RPA robots at regular intervals against pre-configured paths and file name rules to monitor data availability. Based on business logic, each product will have multiple fund accounts, and each fund account will have two paths for storing managed and outsourced transaction data files. Different file name matching rules are configured based on the inherent properties of the fund account. When all fund accounts under a product have completed data availability, the current product data is available, and the data checker notifies the end-of-day clearing post.

[0066] The RPA robot regularly checks the availability of the statement funds data and automatically imports the available data. Based on the same business logic as above, the data checker notifies the reconciliation post when the statement funds data of all fund accounts are available and imported.

[0067] b) Day-end settlement post: RPA robots are divided into two categories: one is the task allocation machine and the other is the task processing machine. Each functional node has and can only have one corresponding task allocation machine in the same application. There are multiple task processing machines, each of which can receive different tasks issued by different allocation machines and perform different process processing according to different tasks. The day-end settlement post here mainly refers to the RPA robot that plays the role of the task processor. After receiving relevant task instructions, such as the day-end settlement processing instruction, the RPA robot will execute the pre-designed day-end settlement process to complete the machine status pre-detection, data acquisition, accounting, and accounting end detection. When acquiring data, it will ensure that only one processor accesses the resource pool to be processed at the same time to avoid excessive consumption. Pre-detection, accounting, and accounting end detection are developed in detail according to the actual needs of the current process. After the day-end settlement post completes all node processes, it will send a notification to the reconciliation post.

[0068] c) Reconciliation post: After receiving the notification of arrival of reconciliation statements from the data check post and the notification of completion of accounting from the end-of-day settlement post, the reconciliation post will perform the corresponding internal reconciliation operation. Internal reconciliation is the comparison of data within the trusteeship and the outsourcing. When the internal data reconciliations of both parties are consistent, an automated trusteeship outsourcing reconciliation can be performed. At this time, if the reconciliation results are inconsistent or there is data that needs to be adjusted, the business personnel will initiate a re-check request.

[0069] like Figure 4 , use AI Agent digital employees to achieve load balancing of business processes. The business needs to automatically do accounting and reconciliation for 5,000+ products every day. As mentioned above, products are processed in batches. The distribution of processing time of products in a batch seriously affects the execution efficiency of each batch, and the efficiency of each batch affects the overall efficiency of the business. In order to better solve this problem, it is necessary to use AI Agent robots to reasonably plan 5,000+ products into different product groups, and the processing time of the same product group is concentrated, which reduces waiting time and improves the processing efficiency of each batch and the overall processing efficiency.

[0070] like Figure 5 , the AI ​​Agent digital employee architecture is:

[0071] a) Memory: AIAgent digital staff stores the historical valuation processing time of all 5,000+ products, which are divided into four different levels according to the concurrent processing capacity of the Hang Seng system: fast account, relatively fast account, slow account, and relatively slow account. The corresponding concurrency numbers for each level are 5, 10, 30, and 60 respectively.

[0072] b) Planning: AI Agent plans the products to be processed daily, dynamically groups them according to product processing time, self-evaluates historical grouping data and processing efficiency before grouping, self-optimizes grouping planning, and plans the best planning combination for the day.

[0073] c) Tools: AI Agent monitors the status of digital employees (data checker, end-of-day clearing, and reconciliation) on virtual machines, whether they are idle or working.

[0074] d) Execution: When it is detected that a digital employee is idle, the AI ​​Agent digital employee will automatically assign batches to the idle digital employees.

[0075] The calculation process of the number of products that should be allocated to the four levels of the AI ​​Agent digital employee in the dynamic grouping is as follows:

[0076] First, calculate the number of products that the four levels can theoretically process per unit time, set as R fast , R faster , R slow , R slower , the calculation formula is as follows:

[0077] R fast =C fast ×T;

[0078] R faster =C faster ×T;

[0079] R slow =C slow ×T;

[0080] R slower =C slower ×T;

[0081] Among them, C fast , C faster , C slow , C slower The numbers of concurrent requests corresponding to the four levels of fast account, relatively fast account, slow account and relatively slow account are 5, 10, 30 and 60 respectively; T represents the total processing time (which can be determined according to the actual operation of the Hang Seng system, such as the working hours of a day, etc.);

[0082] Then, according to the historical grouping data and processing efficiency, the proportion of each level of product quantity is self-evaluated and optimized to obtain the proportion of the product quantity of four levels: fast account, relatively fast account, slow account, and relatively slow account. fast , P faster , P slow , P slower (The sum of these proportions should be 1);

[0083] Finally, calculate the actual number of products that should be allocated to each level, set as n fast 、n faster 、n slow 、n slower , the calculation formula is as follows:

[0084] n fast =N×P fast ;

[0085] n faster =N×P faster ;

[0086] n slow =N×P slow ;

[0087] n slower =N×P slower ;

[0088] Wherein, N represents the total number of products that need to be processed (here, more than 5,000 products).

[0089] The AI ​​Agent takes into account the fluctuation of processing time, and the calculation process of the actual number of products to be allocated to the four levels is as follows:

[0090] (1) Calculate the minimum value R of the number of products processed per unit time for the four levels min-fast , R min-faster , R min-slow , R min-slower and the maximum value R max-fast , R max-faser , R max-slow , R max-slower :

[0091]

[0092]

[0093] Among them, t fast-min ,t fast-max is the minimum and maximum processing time of the quick account product, t faster-min ,t faster-max is the minimum and maximum processing time of the faster account product, t slow-min、 t slow-max The minimum and maximum processing time of slow account products, t slower-min ,t slower-max Minimum and maximum processing times for slower account products;

[0094] (2) Combined with the historical processing efficiency weights, calculate the number of products processed per unit time R for the four levels com-fast , Rcom-faster , R com-slow , R com-slower :

[0095]

[0096] Among them, W fast , W faster , W slow , W slower are the processing efficiency weights of the four levels based on historical data (the sum of the weights is 1);

[0097] (3) Recalculate product quantity allocation:

[0098] The proportion of new products in four levels com-fast , P com-faster , P com-slow , P com-slower Calculated as follows:

[0099]

[0100] (4) The final number of products in the four grades n com-fast 、n com-faster 、n com-slow 、n com-slower The calculation is as follows:

[0101] n com-fast =N×P com-fast ;

[0102] n com-faster =N×P com-faster ;

[0103] n com-slow =N×P com-slow ;

[0104] n com-slower =N×P com-slower .

[0105] After calculating the actual number of products to be allocated to the four levels, the AI ​​Agent groups the products according to the calculated number of products, or groups them according to the processing time, or groups them in combination with the product type, or adjusts the groups in real time.

[0106] 1. Allocate products according to calculated quantities

[0107] According to the number of products calculated above, the products are allocated to the four groups of fast account, faster account, slow account, and slower account. For example, from the list of all products, first select n com-fast products into the quick account group, then select n com-fasterThe products are placed in the faster account group, and so on.

[0108] 2. Consider product characteristics to assist in grouping

[0109] Sort and group by processing time: If you can get a more accurate estimate of the processing time for each product, you can sort all products by processing time from smallest to largest. Then, based on the previously calculated range of product quantities for each level, start with the product with the shortest processing time. For example, the product with the shortest processing time is grouped into the fast group, and the following products are grouped into the faster group, the slower group, and the slower group.

[0110] Grouping by product type: If different product types have different processing time patterns, you can first classify them by product type. For example, products are divided into types A, B, C, etc. It is known that most of the A-type products belong to the faster processing level in history. Then when grouping, the A-type products that meet the faster processing time range should be placed in the faster group.

[0111] 3. Adjust grouping in real time

[0112] During the grouping process, if it is found that the actual processing time of a product does not match the expected processing time range of the group it belongs to, or if the system load and other conditions change (for example, a group's concurrent processing tasks are backlogged), the product can be adjusted to another more suitable group in real time. For example, if a product originally assigned to the fast account group is found to have a processing time longer than the upper limit of the fast account group during actual processing, it can be adjusted to the slow account group.

[0113] The following is a Python sample code framework for implementing the function of grouping products based on the above grouping ideas. Assuming that there is already a list products containing all product information, and each product has an attribute (such as processing_time) to indicate its processing time, the sample code is as follows:

[0114] #Define the number of concurrency corresponding to each level.

[0115] CONCURRENCY_FAST = 5;

[0116] CONCURRENCY_FASTER = 10;

[0117] CONCURRENCY_SLOW = 30;

[0118] CONCURRENCY_SLOWER = 60;

[0119] #Total processing time (assuming an example value is set here, the actual value needs to be determined according to the situation).

[0120] TOTAL_PROCESSING_TIME=3600 # in seconds, assuming 1 hour here;

[0121] #Processing time range for each level (example value, need to be adjusted according to actual situation);

[0122] FAST_MIN_TIME,FAST_MAX_TIME=10,30;

[0123] FASTER_MIN_TIME,FASTER_MAX_TIME=30,60;

[0124] SLOW_MIN_TIME,SLOW_MAX_TIME=60,120;

[0125] SLOWER_MIN_TIME,SLOWER_MAX_TIME=120,240;

[0126] #The processing efficiency weights of each level based on historical data (example values, to be determined based on actual conditions);

[0127] WEIGHT_FAST=0.3;

[0128] WEIGHT_FASTER=0.3;

[0129] WEIGHT_SLOW = 0.2;

[0130] WEIGHT_SLOWER = 0.2;

[0131] def calculate_group_sizes(products):

[0132] """

[0133] Calculate the size of each group based on the product list;

[0134] """

[0135] num_products=len(products).

[0136] #Calculate the minimum and maximum number of products processed per unit time at each level.

[0137] R_MIN_FAST=TOTAL_PROCESSING_TIME / FAST_MAX_TIME*CONCURRENCY_FAST;

[0138] R_MAX_FAST = TOTAL_PROCESSING_TIME / FAST_MIN_TIME * CONCURRENCY_FAST;

[0139] R_MIN_FASTER = TOTAL_PROCESSING_TIME / FASTER_MAX_TIME * CONCURRENCY_FASTER;

[0140] R_MAX_FASTER = TOTAL_PROCESSING_TIME / FASTER_MIN_TIME * CONCURRENCY_FASTER;

[0141] R_MIN_SLOW = TOTAL_PROCESSING_TIME / SLOW_MAX_TIME * CONCURRENCY_SLOW;

[0142] R_MAX_SLOW = TOTAL_PROCESSING_TIME / SLOW_MIN_TIME * CONCURRENCY_SLOW;

[0143] R_MIN_SLOWER = TOTAL_PROCESSING_TIME / SLOWER_MAX_TIME * CONCURRENCY_SLOWER;

[0144] R_MAX_SLOWER = TOTAL_PROCESSING_TIME / SLOWER_MIN_TIME * CONCURRENCY_SLOWER;

[0145] # Calculate the number of products processed per unit time comprehensively.

[0146] R_COM_FAST = (WEIGHT_FAST * ((R_MIN_FAST + R_MAX_FAST) / 2) + (1 - WEIGHT_FAST) * ((R_MIN_FAST + R_MAX_FAST) / 2));

[0147] R_COM_FASTER = (WEIGHT_FASTER * ((R_MIN_FASTER + R_MAX_FASTER) / 2) + (1 - WEIGHT_FASTER) * ((R_MIN_FASTER + R_MAX_FASTER) / 2));

[0148] R_COM_SLOW = (WEIGHT_SLOW * ((R_MIN_SLOW + R_MAX_SLOW) / 2) + (1 - WEIGHT_SLOW) * ((R_MIN_SLOW + R_MAX_SLOW) / 2));

[0149] R_COM_SLOWER = (WEIGHT_SLOWER * ((R_MIN_SLOWER + R_MAX_SLOWER) / 2) + (1 - WEIGHT_SLOWER) * ((R_MIN_SLOWER + R_MAX_SLOWER) / 2));

[0150] # Calculate the proportion of the quantity of each grade of products.

[0151] total_com = R_COM_FAST + R_COM_FASTER + R_COM_SLOW + R_COM_SLOWER;

[0152] P_COM_FAST = R_COM_FAST / total_com;

[0153] P_COM_FASTER = R_COM_FASTER / total_com;

[0154] P_COM_SLOW = R_COM_SLOW / total_com;

[0155] P_COM_SLOWER = R_COM_SLOWER / total_com;

[0156] # Calculate the quantity of products in each group.

[0157] n_com_fast = int(num_products * P_COM_FAST);

[0158] n_com_faster = int(num_products * P_COM_FASTER);

[0159] n_com_slow = int(num_products * P_COM_SLOW);

[0160] n_com_slower = int(num_products * P_COM_SLOWER);

[0161] return n_com_fast, n_com_faster, n_com_slow, n_com_slower;

[0162] def group_products(products):

[0163] """

[0164] Group the products according to the calculated group size;

[0165] """

[0166] n_com_fast,n_com_faster,n_com_slow,n_com_slower=calculate_group_sizes(products);

[0167] fast_group=products[:n_com_fast];

[0168] faster_group=products[n_com_fast:n_com_fast+n_com_faster];

[0169] slow_group=products[n_com_fast+n_com_faster:n_com_fast+n_com_faster+n_com_slow];

[0170] slower_group=products[n_com_fast+n_com_faster+n_com_slow:];

[0171] return fast_group,faster_group,slow_group,slower_group.

[0172] #Assume that there is a sample product list here, which needs to be replaced with real data in actual applications.

[0173] products_example=[{"processing_time":20},{"processing_time":40},{"processing_time":80};

[0174] {"processing_time":150},{"processing_time":200}];

[0175] fast_group,faster_group,slow_group,slower_group=group_products(products_example);

[0176] print("Number of products in fast account group:",len(fast_group));

[0177] print("Number of products in the faster group:",len(faster_group));

[0178] print("Number of products in the slow account group:",len(slow_group));

[0179] print("Number of products in the slower group:",len(slower_group)).

[0180] In the above code:

[0181] The calculate_group_sizes function calculates the number of products that should be allocated to each group based on the product list and the set parameters of each level (number of concurrency, processing time range, processing efficiency weight, etc.).

[0182] The group_products function performs the actual grouping operation on the input product list according to the group sizes calculated by the calculate_group_sizes function.

[0183] The present invention implements a method for global human-machine collaborative automated valuation and reconciliation using an AI Agent written in Phthon. AI Agents written in other languages ​​are also within the scope of protection of this case.

[0184] The present invention combines AI Agent digital employee management and allocation, automatically divides 5000+ products into multiple groups with a small number of product groups, runs them on all accounting robots as much as possible, and uses a sound monitoring and recovery mechanism to make full use of load balancing, high concurrency, high reliability and high scalability to improve accounting efficiency and quality. Through file inspection and risk control of each reconciliation link, and complete separation of accounting steps, the overall accounting efficiency, stability and success rate are improved.

[0185] At the same time, the present invention provides:

[0186] A server comprises a processor and a memory, wherein the memory stores at least one program, and the program is loaded and executed by the processor to implement the valuation and accounting method based on AI Agent digital employees.

[0187] A computer-readable storage medium, wherein at least one program is stored in the storage medium, and the program is loaded and executed by a processor to implement the valuation and accounting method based on AI Agent digital employees.

[0188] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A valuation and accounting reconciliation method based on AI Agent digital employees, characterized in that: The following steps are involved: S1. According to the valuation and reconciliation process, the digital employees deployed on the virtual machines are divided into digital inspection posts, end-of-day settlement posts, and reconciliation posts; Among them, the digital inspection post is used for the on- and off-site transaction data inspection link; the end-of-day settlement post is used for the accounting link; the reconciliation post is used for the statement data inspection link, internal reconciliation link, and custody outsourcing reconciliation link; S2. Build AI Agent digital employees, who manage and assign digital inspection posts, end-of-day settlement posts, and reconciliation posts, completing the entire process of automated valuation and reconciliation; The AI ​​Agent digital employee processing logic is as follows: Store the historical valuation processing time of several fund product pool products. According to the concurrent processing capacity of the Hang Seng system, it is divided into four levels: fast account, relatively fast account, slow account, and relatively slow account, and the corresponding number of concurrent accounts for each level is set; AIAgent plans the products to be processed every day, dynamically groups them according to the product processing time, self-evaluates the historical grouping data and processing efficiency before grouping, self-optimizes the grouping plan, and plans the best planning combination for the day; AI Agent monitors the status of digital employees on virtual machines, whether they are idle or working; when it monitors that a digital employee on a virtual machine is idle, AI Agent digital employees will automatically assign batches to the idle digital employees.

2. The valuation and accounting reconciliation method based on AI Agent digital employees according to claim 1 is characterized in that: The calculation process of the number of products that should be allocated to the four levels of the AI ​​Agent digital employee in the dynamic grouping is as follows: First, calculate the number of products that the four levels can theoretically process per unit time, set as R fast , R faster , R slow , R slower , the calculation formula is as follows: R fast =C fast ×T; R faster =C faster ×T; R slow =C slow ×T; R slower =C slower ×T; Among them, C fast , C faster , C slow , C slower They are the concurrency numbers corresponding to the four levels of fast account, relatively fast account, slow account, and relatively slow account; T represents the total processing time; Then, based on the historical grouping data and processing efficiency, the product quantity proportion of each level is self-evaluated and optimized to obtain the product quantity proportion P of four levels: fast account, relatively fast account, slow account, and relatively slow account. fast , P faster , P slow , P slower ; Finally, calculate the actual number of products that should be allocated to each level, set as n fast 、n faster 、n slow 、n slower , the calculation formula is as follows: n fast =N×P fast ; n faster =N×P faster ; n slow =N×P slow ; n slower =N×P slower ; Where N represents the total number of products that need to be processed.

3. The valuation and accounting reconciliation method based on AI Agent digital employees according to claim 2 is characterized in that: The AI ​​Agent takes into account the fluctuation of processing time, and the calculation process of the actual number of products to be allocated to the four levels is as follows: (1) Calculate the minimum value R of the number of products processed per unit time for the four levels min-fast , R min-faster , R min-slow , R min-slower and the maximum value R max-fast , R max-faser , R max-slow , R max-slower : Among them, t fast-min ,t fast-max is the minimum and maximum processing time of the quick account product, t faster-min ,t faster-max is the minimum and maximum processing time of the faster account product, t slow-min、 t slow-max The minimum and maximum processing time of slow account products, t slower-min、 t slower-max Minimum and maximum processing times for slower account products; (2) Combined with the historical processing efficiency weights, calculate the number of products processed per unit time R for the four levels com-fast , R com-faster , R com-slow , R com-slower : Among them, W fast , W faster , W slow , W slower The processing efficiency weights for the four levels are obtained based on historical data; (3) Recalculate product quantity allocation: The proportion of new products in four levels com-fast , P com-faster , P com-slow , P com-slower Calculated as follows: (4) The final number of products in the four grades n com-fast 、n com-faster 、n com-slow 、n com-slower The calculation is as follows: n com-fast =N×P com-fast ; n com-faster =N×P com-faster ; n com-slow =N×P com-slow ; n com-slower =N×P com-slower 。 4. The valuation and accounting reconciliation method based on AI Agent digital employees according to claim 2 or 3 is characterized in that: After calculating the actual number of products to be allocated to the four levels, the AI ​​Agent groups the products according to the calculated number of products, or groups them according to the processing time, or groups them in combination with the product type, or adjusts the groups in real time.

5. A server, comprising a processor and a memory, characterized in that: At least one program is stored in the memory, and the program is loaded and executed by the processor to implement the valuation and accounting method based on AI Agent digital employees as described in any one of claims 1 to 3.

6. A computer-readable storage medium, wherein at least one program is stored in the storage medium, characterized in that: The program is loaded and executed by the processor to implement the valuation and accounting method based on AI Agent digital employees as described in any one of claims 1 to 3.

Images