An automatic management method, device and system for clearing accounts receivable

Abstract

The present application relates to the field of receivables management, in particular to a kind of receivables automatic clearing management method, the receivables automatic clearing management method is by obtaining the data of the proceeds of the project pre-processing, and according to the matching rule of multilevel is matched to obtain data matching result, according to scene matching condition, the data matching result is matched transaction scene to avoid artificial transaction scene judgment, the data matching result of transaction scene carries out data check, and through clearing condition identification data check result, the proceeds of the project is cleared step, and using confidence level grading condition is graded and data confirmation, and the proceeds of the project after data confirmation generates accounting voucher, simultaneously marks its reconciliation state and generates matching number, through data check, the association between receivables information and proceeds information, avoid artificial repetitive operation account reconciliation, so as to avoid reconciliation delay and improve reconciliation efficiency.

Description

Technical Field

[0001] This invention relates to the field of accounts receivable management, and in particular to a method, apparatus and system for automatic accounts receivable settlement management. Background Technology

[0002] In the process of corporate asset operation and management, reconciling and settling accounts receivable and credit funds is a crucial and demanding task.

[0003] In the current technology, most enterprises' asset operation involves a variety of transaction scenarios, such as large single transactions, multiple small transactions, advance payments, and cross-border transactions including tax and fee deductions. However, because the credit system, trade finance system and core banking system are independent of each other, accounts receivable information and cash flow information cannot be automatically linked. It is necessary to rely on manual judgment of different transaction scenarios and perform a large number of repetitive manual operations to reconcile accounts, resulting in reconciliation delays and low efficiency. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, device, and system for automatic accounts receivable settlement management to address the above-mentioned problems.

[0005] This invention is implemented as follows: an automatic accounts receivable settlement management method, the method comprising:

[0006] S1: Obtain project payment data and preprocess the project payment data;

[0007] S2: Perform hierarchical data matching on the preprocessed project payment data according to multi-level matching rules to obtain the data matching results;

[0008] S3: Match the data matching results to the transaction scenario based on the scenario matching conditions, and perform data verification on the data matching results of the transaction scenario;

[0009] S4: Identify the verification results of the transaction scenario based on the settlement conditions. If the settlement conditions are met, settle the project payment.

[0010] S5: If the settlement conditions are not met, the data for project receipts will be supplemented. For the processing results after data supplementation, S3-S4 will be executed.

[0011] S6: Classify and confirm the settled project receivables according to the confidence level classification conditions, update the settlement status of project receivables and generate accounting vouchers.

[0012] S7: Mark the reconciliation status of the project receipts for which accounting vouchers are generated, generate a matching number, and synchronize it to the first platform.

[0013] In one embodiment, the present invention provides an automatic accounts receivable settlement management device, the automatic accounts receivable settlement management device comprising:

[0014] The preprocessing module acquires project payment data and preprocesses it.

[0015] The hierarchical matching module performs hierarchical data matching on the preprocessed project payment data according to multi-level matching rules to obtain the data matching results.

[0016] The transaction scenario matching module matches the data matching results with transaction scenarios based on the scenario matching conditions, and performs data verification on the data matching results of the transaction scenarios;

[0017] The settlement module identifies the verification results of the transaction scenario based on the settlement conditions. If the settlement conditions are met, the project payment is settled.

[0018] If the settlement conditions are not met, the data completion module will complete the data for project payments. The processing results after data completion will be processed in steps S3-S4.

[0019] The voucher generation module classifies settled project receipts according to confidence level criteria and confirms data, updates the settlement status of project receipts, and generates accounting vouchers.

[0020] The data synchronization module marks the reconciliation status of the project receipts that generate accounting vouchers, generates a matching number, and synchronizes it to the first platform.

[0021] In one embodiment, the present invention provides an automatic accounts receivable settlement management system, the automatic accounts receivable settlement management system comprising: a management server and an ERP system;

[0022] The management server is used to execute the automatic accounts receivable settlement management method described above;

[0023] The ERP system communicates with the management server to receive data from the management server.

[0024] The automatic accounts receivable settlement management method provided in this invention preprocesses the project's payment data and performs hierarchical matching based on multi-level matching rules to obtain data matching results. Then, it matches the data matching results to transaction scenarios based on scenario matching conditions. The data matching results for the transaction scenarios are then verified. The verification results are identified through settlement conditions. If the settlement conditions are met, the project's payments are settled. Confidence level grading conditions are used for grading and data confirmation. For the confirmed project payments, accounting vouchers are generated, and their reconciliation status is marked with a matching number. By matching the data matching results to transaction scenarios and verifying the data matching results for the transaction scenarios, the association between accounts receivable information and payment information can be obtained, avoiding repetitive manual reconciliation operations, thereby preventing reconciliation delays and improving reconciliation efficiency. Attached Figure Description

[0025] Figure 1 This is a flowchart of an automatic accounts receivable settlement management method in one embodiment;

[0026] Figure 2 This is a flowchart illustrating data matching in an automatic accounts receivable settlement management method in one embodiment.

[0027] Figure 3 This is a structural block diagram of an automatic accounts receivable settlement management device in one embodiment;

[0028] Figure 4 This is a block diagram of an automatic accounts receivable settlement management system in one embodiment;

[0029] Figure 5 This is a block diagram of the internal structure of a computer device in one embodiment. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0031] It is understood that the terms "first," "second," etc., used in this invention may be used to describe various elements herein, but unless specifically stated otherwise, these elements are not limited by these terms. These terms are used only to distinguish one element from another. For example, without departing from the scope of this invention, a first script may be referred to as a second script, and similarly, a second script may be referred to as a first script.

[0032] Figure 1 Here is a flowchart of an automatic accounts receivable settlement management method provided in one embodiment, such as... Figure 1 As shown, in one embodiment, an automatic accounts receivable settlement management method is proposed. Specifically, it may include the following steps:

[0033] S1: Obtain project payment data and preprocess the project payment data;

[0034] S2: Perform hierarchical data matching on the preprocessed project payment data according to multi-level matching rules to obtain the data matching results;

[0035] S3: Match the data matching results to the transaction scenario based on the scenario matching conditions, and perform data verification on the data matching results of the transaction scenario;

[0036] S4: Identify the verification results of the transaction scenario based on the settlement conditions. If the settlement conditions are met, settle the project payment.

[0037] S5: If the settlement conditions are not met, the data for project receipts will be supplemented. For the processing results after data supplementation, S3-S4 will be executed.

[0038] S6: Classify and confirm the settled project receivables according to the confidence level classification conditions, update the settlement status of project receivables and generate accounting vouchers.

[0039] S7: Mark the reconciliation status of the project receipts for which accounting vouchers are generated, generate a matching number, and synchronize it to the first platform.

[0040] In this embodiment of the invention, the acquired project payment data can be from all online banking systems under the enterprise's name. The online banking systems can include, but are not limited to, banks and cross-border payment institutions. By automatically downloading bank transaction details related to the enterprise's projects from the associated online banking systems, project payments can be understood as the enterprise's details of receipts for the project, or invoices for related cash flow projects. Preprocessing can be done by processing some invalid transactions in the automatically downloaded bank transaction details, so that the preprocessed data is all payment data related to the same project.

[0041] In this embodiment of the invention, the multi-level matching rule can be to first perform hierarchical data matching on the preprocessed project payment data through a hierarchical matching model of customer name, transaction amount, and invoice number. The resulting data matching result can be at least two of the above-mentioned data results. It can be understood that after the obtained project payment data is preprocessed and subjected to the multi-level matching rule, the customer name and transaction amount data are obtained, but the invoice number data is missing. This is one level of matching result. Alternatively, the data of the project payment customer name, transaction amount, and invoice number can be obtained. This is another level of matching result. The purpose of matching the data through the multi-level matching rule is to create a hierarchical structure based on the data matching results of different levels after matching, so as to facilitate different settlement orders for different transaction scenarios.

[0042] In this embodiment of the invention, the scenario matching condition can be the classification of several levels of data matching results. Different levels of data matching results are matched with different transaction scenarios to determine whether the project payment belongs to different transaction scenarios. Different transaction scenarios can include, but are not limited to, large single transaction scenarios, small multiple transaction scenarios, prepayment scenarios, and cross-border scenarios with tax and fee deductions. Avoiding the process of manually judging transaction scenarios can improve efficiency. The data matching results of different transaction scenario levels are verified. The data verification process can be the comparison and verification of the project payment data with the outstanding payments in the project accounts receivable data table. The data verification process for different transaction scenarios is related to the data matching results of the different levels.

[0043] In this embodiment of the invention, the settlement condition can be that the customer name, transaction amount, and invoice number in the project payment data correspond to the information of the outstanding payments in the project receivables data table, and are correct and without any missing or ambiguous information. If all the project payment data meets the settlement condition, the settlement procedure is automatically executed, and the current project payment is marked as "settled".

[0044] In this embodiment of the invention, if any item in the project receivables data—customer name, transaction amount, or invoice number—does not match, is incorrect, or contains ambiguous information in the project receivables data table, and this defect cannot be overcome after data verification, then the relevant financial personnel need to be notified to supplement or confirm the missing or incorrect information in the project receivables. After the data is supplemented or confirmed, the transaction scenario is matched again and the data is verified. Then, the results of the data verification are identified according to the settlement conditions. If the conditions are met, the settlement continues; if the conditions are not met, the data needs to be supplemented again until the project receivables are settled.

[0045] In this embodiment of the invention, the confidence level grading condition can be to score and grade the project receipts that have been automatically settled after matching in multiple dimensions, or to score them in different steps. This can be used to confirm data in different ways through confidence level grading. If the confidence level grading scores of several receipts of the current project are high, and the matching of each receipt and the multiple dimensions of the automatic settlement process meet the grading requirements, the settlement status of the project receipts is updated after the data confirmation of multiple outstanding receipts of the project is completed. At the same time, the settlement time and operator are updated. The system can also automatically mark "AI matching" and mark it as "final settlement". This can be understood as a further confirmation of the settled project receipts, so as to avoid reconciliation errors and reduce efficiency. After updating the settlement status, standardized accounting vouchers can be generated for the project receipts. The vouchers can also be accompanied by screenshots of the corresponding matching basis, such as screenshots of bank details and accounts receivable details. The accounting vouchers can include: the name of the accounting voucher: indicating the voucher number, usually composed of letters and numbers. Date of Voucher Preparation: Indicates the date the transaction occurred, usually in the order of year, month, and day. Voucher Number: Indicates the voucher number. Summary of Economic Transaction: Provides a brief description of the transaction; it should be concise, accurate, and clear. Debit and Credit Accounts: Indicates the names and amounts of the accounting accounts involved in the transaction, including debit and credit entries. Amount: Indicates the amount involved in the transaction. Number of Attached Original Documents: Indicates the number of documents attached to the voucher, such as invoices and receipts. Signatures or Seals of Voucher Preparer, Auditor, Bookkeeper, Head of Accounting Department, and Accounting Supervisor: Indicates the names of the personnel who prepared the voucher, reviewed the voucher, and recorded the voucher in the ledgers. Specific content and format can be adjusted according to actual circumstances and management needs.

[0046] In this embodiment of the invention, the bank transaction details corresponding to the generated standardized accounting vouchers are marked as "reconciled" and a unique matching number is generated based on the accounting vouchers. The matching number can be, for example, AM-20250927-001, to facilitate subsequent tracing based on the matching number and simultaneous synchronization to the ERP system of the first platform. The first platform can be the enterprise's management platform or a client connected to facilitate financial reconciliation inquiries and data supplementation to the client.

[0047] The automatic accounts receivable settlement management method provided in this invention preprocesses the project's payment data and performs hierarchical matching based on multi-level matching rules to obtain data matching results. Then, it matches the data matching results to transaction scenarios based on scenario matching conditions. The data matching results for the transaction scenarios are then verified. The verification results are identified through settlement conditions. If the settlement conditions are met, the project's payments are settled. Confidence level grading conditions are used for grading and data confirmation. For the confirmed project payments, accounting vouchers are generated, and their reconciliation status is marked with a matching number. By matching the data matching results to transaction scenarios and verifying the data matching results for the transaction scenarios, the association between accounts receivable information and payment information can be obtained, avoiding repetitive manual reconciliation operations, thereby preventing reconciliation delays and improving reconciliation efficiency.

[0048] In one embodiment of the present invention, the step of performing hierarchical data matching on the preprocessed project payment data according to multi-level matching rules to obtain several data matching results includes:

[0049] Obtain project payment data through timed trigger mode and / or event trigger mode;

[0050] The system filters invalid transactions in project payments through a natural language processing module and performs word segmentation on the project payment data. From the segmented words, it identifies words that are associated with at least one of the following: customer name, transaction amount, summary text, and transaction time.

[0051] The related words and their corresponding synonyms, near-synonyms or antonyms are merged into a unified information identifier to obtain at least one of the following information: customer name, transaction amount, summary text and transaction time;

[0052] Information is supplemented by linking to the customer information database to complete missing elements;

[0053] The information after information completion is matched hierarchically according to the multi-level matching rules, and the data matching result is obtained which includes at least two of the following combinations: customer information, amount information, and identifier information.

[0054] In this embodiment of the invention, the data for obtaining project payments can be obtained through an API interface encapsulated using the national cryptographic algorithm SM4, connecting to all online banking systems (including banks and cross-border payment institutions) under the enterprise's name. A dual-mode approach is adopted: "timed triggering (executed every 1 / 30 minutes) + event triggering (instant push when a new loan transaction is added in online banking)," automatically downloading "bank transaction details" (including payer name, transaction amount, summary, transaction time, payer account, and remarks fields). The entire data transmission process is encrypted, complying with the requirements of the Personal Information Protection Law and the Data Security Law. The national cryptographic algorithm SM4 encryption refers to a symmetric encryption algorithm promulgated by the State Cryptography Administration of China, used to protect the confidentiality of data. This algorithm provides efficient encryption performance while maintaining high security. The encryption and decryption process of the SM4 algorithm is reversible; the same key can be used for both encryption and subsequent decryption, thereby ensuring data integrity and confidentiality.

[0055] In this embodiment of the invention, the Natural Language Processing (NLP) module is invoked to automatically filter invalid transactions (such as bank fees, internal account transfers, and refund reversals); data consistent with project payment receipts can be obtained, such as: customer name xx company, transaction amount xxxx.xx yuan, payment for xx project in the summary text, invoice number 123456789, transaction date xxxx year xx month xx day, etc.; the Natural Language Processing (NLP) module enables various theories and methods for effective communication between humans and computers using natural language. Natural Language Processing is a science that integrates linguistics, computer science, and mathematics.

[0056] In this embodiment of the invention, a "customer name standardization engine" (a word vector model trained on historical transaction data) is used to merge synonyms such as "XX Co., Ltd.", "XX Company", and "XX Corp" into a unified customer information identifier. This can be achieved by searching a customer information database using acquired keywords, thereby filling in missing customer information and facilitating the identification of the project and company name to which the payment belongs. For example, after unifying the customer identifier for "XX Company", "XX Co., Ltd." is obtained. The completed information ensures the customer name is in full, and the amount information can be standardized to two decimal places. The word vector model trained on historical transaction data is a technique that maps each word in the vocabulary to a vector space. The word vector model primarily performs text classification, helping to map words to the vector space and calculate the vector representation for each category to classify the test text.

[0057] In this embodiment of the invention, missing fields (such as ambiguous payer names) are filled in by associating the payment account with the "customer information database." The cleaned data is then stored in a distributed financial database, supporting full-chain traceability of operation logs (retained for 10 years). Information can also be supplemented for abbreviations in the summary regarding project receipts. For example, if the note is "receipt 'a' for project A," the full name of both the relevant project and the full name of the receipt are supplemented for unified identification, so that it can be registered after settlement and the generation of accounting vouchers.

[0058] In this embodiment of the invention, the information after completion is subjected to hierarchical data matching. After completion, customer information (customer name and / or project name), amount information (transaction amount), and identifier information (invoice number in the project payment details) can be obtained. Hierarchical data matching can be based on any combination of the number of completed information items, resulting in a hierarchical classification. For example, project payment data may include customer information, amount information, and identifier information, and all three items must be complete and accurate; this constitutes one level of matching. Alternatively, if the amount information or identifier information is inconsistent or incorrect, this also constitutes another level of matching. At least two types of information combinations will be included. When retrieving bank transaction data from the online banking system, the transfer transaction details should at least include customer information and amount information, such as the customer's name and the transaction amount. While customer and amount information should not be missing, there may be instances where customer information is abbreviated or the amount does not match the outstanding amount in the project's accounts receivable data table. Identifier information, such as the invoice number, may be missing or unclear. A missing invoice number might indicate a transfer was made before an invoice was issued, but an invoice needs to be issued afterward. Because the details are retrieved in real-time using a timed trigger mode, situations may arise where project payments have been received but identifier information is missing.

[0059] As one embodiment of the present invention, such as Figure 2 As shown, the step of matching the data matching results with transaction scenarios based on scenario matching conditions includes:

[0060] The data matching results for each level are: first level, second level, and third level.

[0061] If the first level of customer information, identifier information, and amount information are correct and complete, then it is matched as an automatic settlement transaction scenario;

[0062] The second level is the case where the customer information is correct and not missing, the identifier information is incorrect or missing, and the amount information is not missing but is incorrect, which is matched as a transaction scenario where the amount information is split and / or totaled.

[0063] The third level includes scenario one, where customer information is incorrect but identifier and amount information are correct and complete; scenario two, where customer information and amount information are correct and complete but identifier information is correct but partially missing; scenario three, where customer information and identifier information are correct and complete but amount information is incorrect; and scenario four, where abnormal transactions occur. Scenario one, scenario two, scenario three, and scenario four all match as abnormal transaction scenarios.

[0064] In this embodiment of the invention, the project payment data can be classified into three levels after hierarchical matching. The first level is the project payment data after the customer information database is completed, which obtains customer information with a unified identifier, namely the customer name and / or project name, or contract project (8% error tolerance, supports recognition of Chinese and English abbreviations and space differences). The corresponding customer information is locked, and the amount information obtained, namely the transaction amount, is accurately compared with the outstanding amount in the project receivables data table (allowing an error of ±0.01 yuan, to deal with bank rounding differences). The identifier information in the transaction summary, namely the invoice number (supports recognition of multiple formats such as "INV-12345" and "invoice 12345"), is obtained and matched with the invoice number of the outstanding amount in the project receivables data table. This is the data matching result of the first level, and it is matched with the automatic settlement transaction scenario. The automatic settlement transaction scenario can automatically perform subsequent settlement, generate accounting vouchers, and other procedures, avoiding manual repetitive reconciliation and thus improving reconciliation efficiency.

[0065] In this embodiment of the invention, the second level is the customer information obtained after the project payment data is supplemented by the customer information database. This information has a unified identifier, namely the customer name and / or project name, which can also be the contract project (with an 8% error tolerance rate and support for Chinese and English abbreviations and space differences). The corresponding customer information is locked. However, if the amount information does not match the amount of outstanding payments in the project receivables data table, and the identifier information is mismatched or missing, this is the data matching result of the second level. If the amount of project payment is greater than the amount of outstanding payments, it can be considered as a case of multiple small transactions at the same time. Conversely, it can be considered as a case of cross-border transactions including tax and fee deductions. It is necessary to split and / or sum the amount information in order to further compare the amount information and match this case as a transaction scenario of splitting and / or summing the amount information.

[0066] In this embodiment of the invention, the third level refers to the situation where, after the project payment data has been supplemented in the customer information database, customer information with multiple customer names is obtained and further splitting and matching comparison is required. At this time, both customer information and amount information are abnormal; or the customer information and amount information are not missing and are correct, but the identifier information, i.e., the invoice number, is only partially or is blurred and cannot be fully identified; or the current project payment data is an abnormal transaction such as payment initiated by an unfamiliar account (not a customer-registered account), a large amount of payment (≥500,000 yuan), or a payment from a high-risk area. All of these can be matched as abnormal transaction scenarios to achieve full scenario coverage, so that each payment can fall into the matching transaction scenario for processing, avoiding missed reconciliation that affects the reconciliation effect and efficiency.

[0067] As one embodiment of the present invention, such as Figure 2 As shown, the transaction scenarios involving the splitting and / or totaling of the amount information include:

[0068] When the amount of project receivables is greater than the amount of a single outstanding receivable in the project receivables data table and the identifier is missing, including scenario A where the total amount of multiple outstanding receivables is present, scenario B where the customer supplier portal executes payment steps, and scenario C where there is no outstanding receivables.

[0069] When the amount of project receipts is less than the amount of a single outstanding payment in the accounts receivable data table and the identifier information is missing, this includes scenario D with deduction of withholding income tax and scenario E without deduction of withholding income tax.

[0070] When the amount of project receivables equals the sum of multiple outstanding receivables in the accounts receivable data table and is missing identifier information, it is scenario F where multiple outstanding receivables are settled simultaneously.

[0071] When the amount of project receivables equals the sum of the amount of receivables in multiple accounts receivable data tables and the identifier information is missing, it is scenario G, which is the splitting of receivables amount information.

[0072] The abnormal transaction scenarios include: scenario H with ambiguous identifier information, scenario I with refund association, scenario J with abnormal transaction anti-fraud, and scenario K with multi-currency transactions.

[0073] In this embodiment of the invention, the transaction scenarios of the amount information matching at the second level are mainly classified into several sub-scenarios based on the comparison between the amount information of project receipts and the single outstanding amount information in the project receivables data table. When the transaction amount of project receipts is greater than the single outstanding amount, it can be understood that the project receipts are multiple outstanding amounts traded together at the same time, or it can be a payment from other supplier portals, or it can be a pre-transaction, i.e., it includes scenario A, scenario B, and scenario C.

[0074] In this embodiment of the invention, when the transaction amount of project receipts is less than the amount of a single outstanding payment, it may be due to the actual amount received being less than the amount of outstanding payments caused by the deduction of taxes included in the cross-border payment, or it may be due to a partial underpayment that causes the transaction amount of receipts to be less than the amount of outstanding payments, i.e., it includes scenario D and scenario E.

[0075] In this embodiment of the invention, when the transaction amount of the project payment is equal to the sum of the amounts of multiple outstanding payments, it can be understood that the amounts of multiple outstanding payments are received at the same time. In this case, the transaction amount of the project payment needs to be split and settled separately, which is scenario F.

[0076] In this embodiment of the invention, when the transaction amount of project payment is equal to the total receivables of different projects, it can be understood that a transaction amount contains multiple customer information, and each customer information corresponds to a payment amount. It is necessary to split the transaction amount and assign it to different customer information, which is scenario G.

[0077] In this embodiment of the invention, the third level of abnormal transaction scenarios may include, but is not limited to, scenario H, scenario I, scenario J, and scenario K. Abnormal transaction scenarios mainly involve special processing of transaction details that do not conform to the first level or the second level of other abnormal situations.

[0078] As one embodiment of the present invention, such as Figure 2 As shown, data verification is performed on the data matching results for transaction scenarios involving total and / or split amounts, including:

[0079] In scenario A, a multi-item combination matching algorithm is used to combine the remaining outstanding receivables details obtained through traversal and search. The total result is then compared with the amount of project receivables, and the identifier information of the total amount combination is recorded.

[0080] Scenario B: The payment notification is retrieved through the key management module. The amount and identifier information on the payment notification are obtained and compared with the amount of outstanding receivables in the project receivables data table. If they match the amount of outstanding receivables, S4 is executed; otherwise, S5 is executed.

[0081] In scenario C, the real-time accounts receivable balance verification module confirms that there are no outstanding project receivables, executes S6-S7, and records the missing identifier information.

[0082] In scenario D, the withholding income tax rate is retrieved from the tax database, the calculation result is compared with the amount of receivable. If the result matches the amount of receivable, the difference is recorded and S4 is executed. If the result does not match the amount of receivable, it is marked and S5 is executed.

[0083] Scenario E: Mark and record project payments and add auxiliary information, then execute S5;

[0084] In scenario F, the exception rules module prioritizes matching the specified amount of outstanding receivables for data verification, then sorts the amount of outstanding receivables in ascending order by date, and matches the amount of receivables for the remaining projects for data verification.

[0085] In scenario G, multiple customer information is identified and obtained through the identification and / or splitting module. The amount information of project payments is split according to the identifier information, and the amount information of each split project payment is checked against the amount of outstanding receivables in the accounts receivable data table.

[0086] In this embodiment of the invention, when matching to scenario A (multiple item total matching scenario), the "multiple item combination matching algorithm" is automatically triggered. It iterates through all outstanding accounts receivable details under the customer's name and calculates all possible "multiple item amount combinations" through dynamic programming algorithm. If there is a combination that is completely consistent with the bank transaction amount (supporting 2-5 item combination combinations), the data is automatically checked according to the "combination details" and the settlement operation of S4 can continue to be executed. At the same time, the combination logic (such as "invoice 12345 + invoice 12346, total amount XX yuan") is recorded in the "settlement log" for easy financial verification.

[0087] The multi-part combination matching algorithm, also known as the many-to-many combination matching algorithm, is an algorithm used to efficiently process and combine large amounts of data, and can handle many-to-many matching needs more flexibly and efficiently. Dynamic programming is an algorithmic paradigm that achieves efficient solution in both time and space by decomposing the problem, storing solutions to subproblems, and avoiding redundant computation. Dynamic programming (DP) is a problem-solving algorithmic paradigm with wide applications in many fields. Its core idea is to decompose the problem into subproblems and store the solutions to the solved subproblems to avoid redundant computation and improve efficiency. The success of dynamic programming algorithms is based on two fundamental principles: optimal substructure and overlapping subproblems. Optimal substructure means that the optimal solution to a problem can be derived from the optimal solutions to its subproblems; overlapping subproblems refer to problems that can be decomposed into several overlapping subproblems, which may be solved multiple times. To avoid redundant computation, memoization is used to store the solutions to the solved subproblems for direct use later, improving efficiency.

[0088] In this embodiment of the invention, when matching scenario B (the processing scenario where the customer has a supplier portal), the encrypted login credentials of the customer's supplier portal (such as Ariba, Coupa, SAPAriba Network) are retrieved from the key management module (non-plaintext storage) of the "customer information database". The "headless browser + automatic CAPTCHA recognition" technology (connected to a third-party AI CAPTCHA recognition interface with an accuracy of ≥99%) is used to simulate manual login and circumvent the portal's anti-crawling mechanism.

[0089] If the "Payment Notification" corresponding to the payment in the portal is structured data (such as Excel or JSON format), extract multiple invoice numbers directly. If it is unstructured data (such as PDF scans or images), call the OCR text recognition module (supporting multiple languages, with a recognition accuracy of ≥98.5%) to extract the invoice number. Then, verify the "Invoice Number - Amount" against the outstanding amount information in the Accounts Receivable data table, and execute the settlement operation in S4 after verification. If the amount information still does not match, execute S5, mark it as "Not Reconciled," and push the pending task to the financial middle platform (with a link to the customer portal, three historical transaction records, and customer contact information). At the same time, set a 48-hour timeout reminder to avoid reconciliation delays. After the information is complete, re-match the transaction scenario and verify the data. After verification is correct, execute the settlement operation in S4.

[0090] The key management module is a system or service responsible for generating, storing, and managing the keys used in the encryption and decryption process. It provides functions for key creation, storage, updating, and deletion, as well as support for secure access and permission management. Different key management modules may have different implementations and functionalities, but their core objective is to ensure the security and availability of keys. The "headless browser + automatic CAPTCHA recognition" technology combines the automated operation capabilities of a headless browser with CAPTCHA recognition technology to achieve automated data collection and processing. A headless browser is a browser without a visual interface, operated through programming. It can simulate human actions, including opening web pages, filling out forms, and clicking buttons, thereby automating web page operations. Automatic CAPTCHA recognition uses image recognition and other technologies to automatically identify CAPTCHAs on web pages, bypassing website access restrictions. The entire access and data collection process is encrypted, and key management information is provided to the client. The operations performed are standardized and comply with the requirements of the Personal Information Protection Law and the Data Security Law.

[0091] In this embodiment of the invention, when matching scenario C (the scenario where the customer has no outstanding accounts receivable), the "Real-time Verification of Customer Accounts Receivable Balance" module confirms that the customer has no outstanding accounts receivable details and automatically determines it as "early collection". S6-S7 can be executed directly to generate an accounting voucher (Debit: Bank Deposit - XX Account, Credit: Accounts Receivable - XX Customer (Early Collection)) and mark the bank transaction as "Reconciled". At the same time, a CRM system notification is triggered (pushed to the corresponding sales representative, with "early collection amount + customer name") to remind the sales representative to follow up on the customer's subsequent orders and ensure that the early collection is associated with the subsequent invoice to avoid confusion in fund ownership.

[0092] The Real-Time Customer Receivable Balance Verification module is part of the Accounts Receivable (AR) module and is primarily responsible for handling customer-related business, especially customer balance inquiries and management. This module is a sub-module of the Financial Accounting (FI) module and is mainly used to handle customer-related business, such as customer master data maintenance, invoice processing, collection processing, miscellaneous business processing, customer balance inquiries, and periodic transaction processing.

[0093] In this embodiment of the invention, when matching to scenario D (a scenario where the customer has a "WHT deduction" identifier), the withholding tax (WHT) rate for the customer's country is retrieved from the "Global Tax Database" (connected to the official tax API, with tax rates updated in real time at a frequency of ≤1 hour). The system automatically calculates the matching accounts receivable amount using the "Tax-inclusive Amount Calculation Formula" (Tax-inclusive receivable amount = Bank receipt amount / (1 - WHT tax rate)). If the calculated "Tax-inclusive receivable amount" matches the amount of an accounts receivable under the customer's name (allowing an error of ±0.02 yuan), after data verification, the system automatically completes the tax difference record (with the note "WHT deduction: XX"). If the tax rate is set to XX% (e.g., RMB), and the verification is completed, the settlement operation in step S4 is executed. If the tax rate update is delayed or the calculated amount does not match, the system marks it as "not reconciled" and includes a "tax rate query link + calculation process" in the pending task to assist in manual verification by finance personnel. Step S5 is then executed to complete the data and re-perform scenario matching and data verification. After the verification is completed again and the amount matches, the settlement operation in step S4 is executed again. The global tax database is a platform that provides global tax information and analysis tools to help users understand and compare tax policies. For example, tax policies, tax rates, regulations, and tax reports are the most authoritative and reliable sources of tax information.

[0094] In this embodiment of the invention, when matching to scenario E (a scenario where the customer does not have the "deduct WHT" identifier), the payment for the project is marked as "not reconciled". Execute S5 to generate a pending task in the financial middle platform (priority set to "medium"), with auxiliary information such as "whether the customer has deduction records in the past", "details of the most recent transaction (including whether other fees were deducted)", and "customer's financial contact information". At the same time, it supports financial personnel to initiate "customer reconciliation inquiry" with one click (the system automatically generates a standardized inquiry email, including bank transaction screenshots), improving communication efficiency. After the data is completed, data matching and transaction scenario matching are performed again, and the completed data is verified. After the data verification is completed and the settlement conditions are met, execute S4 to perform the settlement operation.

[0095] Customer reconciliation refers to the process by which businesses reconcile accounts with customers to ensure the accuracy of their financial records. Reconciliation is a crucial step in ensuring financial clarity and accuracy, helping businesses to promptly identify and correct financial errors and mitigate potential risks.

[0096] In this embodiment of the invention, when matching scenario F (the scenario where the project's receipt amount matches the total of multiple transactions and there is no invoice number), the system automatically sorts the receivables in ascending order by "date of occurrence" (strictly following the first-in, first-out principle), prioritizing the settlement of the earliest outstanding receivables until the total amount matches the bank's receipt amount. If the customer information database contains a special agreement for "specifying the settlement of a specific item" (e.g., the customer has previously requested in writing to prioritize the settlement of a certain overdue invoice), the system automatically triggers the "exception rule engine," prioritizing the matching of the specified item, and then performing data verification on the remaining amount according to the first-in, first-out principle. After the data verification and matching are completed, the settlement operation of S4 is executed. After settlement, a "multiple settlement report" (including the invoice number, occurrence date, settlement amount, and settlement order of each item) is generated and synchronized to the credit control department for updating the customer's credit rating. The exception rule engine module is a special rule engine that allows, based on existing protection strategies, to specify that normal requests meeting certain characteristics can skip the scanning of certain protection modules or rules. The main purpose of the exception rules engine module is to set exceptions in the protection policy, so that certain normal, high-frequency or requests containing specific content can bypass the regular protection checks, thereby avoiding false blocking or false alarms, and allowing accounts receivable that need to be settled first to undergo data verification and settlement operations.

[0097] In this embodiment of the invention, when matching scenario G (splitting of collected payments scenario), the project payment refers to the same payment corresponding to multiple customers (e.g., the summary indicates "collection of payments from Company A + Company B"). The system uses the "keyword recognition + amount splitting" module to first extract the names of the collecting customers (supporting simultaneous recognition of 2-3 customers), and then automatically splits the amount to the corresponding customer's accounts receivable according to the splitting amount indicated in the summary (e.g., "A: 1000 yuan, B: 2000 yuan") or the historical transaction ratio. Data verification is performed on each customer and its corresponding amount. After verification, the settlement operation of S4 is performed to generate a "split settlement voucher" (recording the loan information of each customer separately), and the bank transaction is marked as "reconciled (splitting of collected payments)". At the same time, reconciliation details are sent to each customer to clarify the ownership of the collected amount.

[0098] The "keyword recognition + amount splitting" module typically refers to a functional unit that combines Natural Language Processing (NLP) technology and data processing algorithms. The main purpose of this module is to extract key information from unstructured text data and to rationally split and process the monetary data. Keyword recognition is an important component of NLP, involving finding words or phrases with specific meanings or representing specific concepts from text. In financial data processing, keyword recognition can help the system automatically identify key information such as "company name," "amount," "invoice number," and "date," for example, identifying Company A and Company B. Amount splitting refers to dividing the extracted monetary data according to certain rules. This may involve allocating the total amount according to different categories or items, such as splitting the total amount on an invoice into the amount of goods, taxes, and shipping costs. In some cases, amount splitting may also involve distributing a large amount to multiple recipients according to a certain proportion or rule, such as the amount splitting in a red envelope distribution algorithm, which can split the monetary information into amount A and amount B. Amount A may correspond to the project receipt of Company A, while amount B corresponds to the project receipt of Company B. The sum of amount A and amount B is the project receipt received.

[0099] As one embodiment of the present invention, such as Figure 2 As shown, data verification is performed on the data matching results for abnormal transaction scenarios, including:

[0100] In scenario H, a fuzzy search algorithm is used to perform a fuzzy search on the identifier information of the project payment, and several complete identifiers are matched and a matching suggestion list is generated for personnel confirmation. The confirmed identifier information is then checked against the data, and S4 is executed.

[0101] Scenario I: The refund transaction identification module identifies the details of the refunded project receivables, generates a reverse settlement voucher, restores the pending receivable status of the project receivables, marks the restored pending receivable status, and performs data verification.

[0102] Scenario J connects to the anti-fraud database model. After obtaining project payments, anti-fraud verification is performed. If at least one abnormal information is detected, such as unfamiliar account information or payments from high-risk areas, an alert is triggered and the matching process is stopped. If no abnormality is found, subsequent steps are performed. If an abnormality is found, it is marked and a refund application is generated.

[0103] In scenario K, the real-time exchange rate module converts the amount of payments received for projects with multiple currencies. The converted result is then compared with the amount to be received. If there is an exchange rate difference, the difference is recorded and included in the exchange gain or loss. Then, S4 is executed.

[0104] In this embodiment of the invention, when a scenario H (fuzzy invoice number matching scenario) in an abnormal transaction scenario is matched, if the bank transaction summary only contains partial identifier information, i.e., the invoice number (such as "INV-123" or "12345", the complete version should be "INV-12345"), the system calls the "fuzzy retrieval algorithm" (based on the edit distance algorithm, supporting character missing and sequence difference recognition). Combining this with the format of outstanding invoice numbers under the customer's name (such as a fixed prefix "INV-" and a fixed number of 8 digits), the system associates and matches the most likely complete invoice number, generating a "matching suggestion list" (arranged in descending order of matching similarity, with similarity scores such as 98% and 92%) for financial personnel to select and confirm before executing the settlement operation in S4. The fuzzy retrieval algorithm is based on text similarity calculation and aims to identify incompletely matching text information. The core of the fuzzy retrieval algorithm is to find records with high similarity to the query text in the text data. The algorithm typically uses string edit distance, cosine similarity, Jaccard similarity, and other methods to measure the similarity between texts.

[0105] In this embodiment of the invention, when matching to Scenario I (refund-related settlement scenario), and a refund application occurs (such as invoice cancellation or order cancellation), the "Refund Transaction Identification" module (based on the summary keywords "refund," "reversal," or transaction type code) automatically associates the originally settled accounts receivable details, generating a "reverse settlement voucher" (Debit: Accounts Receivable - XX Customer, Credit: Bank Deposit - XX Account), and simultaneously restores the original invoice's "pending payment" status. If the original invoice is associated with subsequent orders, the system synchronously notifies the CRM sales representative to avoid conflicts between order and fund ownership. The refund transaction is marked as "refunded - pending reconciliation" until the original invoice is resettled or confirmed as void. If the pending payment status is restored, the project payment data corresponding to the pending payment is obtained again, and data matching, scenario matching, and data verification are performed again. The refund transaction identification module is an important component of the transaction system, responsible for identifying and processing refund transactions. In modern e-commerce and payment systems, refunds are a common business scenario involving reverse fund flows and order status updates. The refund transaction recognition module ensures the accuracy and efficiency of the refund process.

[0106] In this embodiment of the invention, when matching scenario J (abnormal transaction anti-fraud scenario), the system connects to the "enterprise anti-fraud database" (including abnormal payment accounts, suspicious transaction amount ranges, and IP addresses from high-risk areas). After obtaining bank transaction details, an "anti-fraud verification" is first performed: if anomalies such as "large payments (≥500,000 RMB) from unfamiliar payment accounts (non-customer-registered accounts)" or "payments initiated from IP addresses from high-risk areas" are detected, an alert is automatically triggered (pushed to the financial risk control department, along with the reason for the anomaly and a screenshot of the transaction), suspending automatic data matching and other transaction scenario operations. After the financial risk control department verifies the authenticity of the transaction (e.g., contacting the customer for confirmation, verifying the ownership of the payment account), if no risk is confirmed, the matching process is manually triggered; if a risk is confirmed (e.g., incorrect receipt of payment), the system marks it as "abnormal transaction - pending refund" and generates a refund application to avoid financial losses. The enterprise anti-fraud database is a database system used to help enterprises detect and prevent fraudulent activities. It can monitor account information, transaction amounts, and transaction IP data in real time and perform data analysis to react quickly and reduce losses when fraud occurs.

[0107] In this embodiment of the invention, when matching scenario K (multi-currency transaction matching scenario), for multi-currency payments (such as USD and EUR) from cross-border customers, the system automatically converts the bank receipt amount into the enterprise's base currency (such as RMB) through the "real-time exchange rate matching" module (connected to the central bank's exchange rate midpoint API, with an exchange rate update frequency ≤ 5 minutes), and then matches it with the base currency amount in the "Accounts Receivable Data Table". If there is an exchange rate difference (≤ 0.03 RMB), the system automatically generates an "exchange rate difference record" (note "exchange rate XX, difference XX RMB"), which is recorded in "financial expenses - exchange gains and losses". After the total amount of the receipt and the difference in the data verification project are matched with the amount to be receivable, the settlement operation of S4 is executed to ensure the accuracy of multi-currency transaction matching. The real-time exchange rate matching module refers to a system module that is implemented through an API interface and can obtain and update exchange rate data in real time.

[0108] In one embodiment of the present invention, the step of classifying and confirming the settled project receivables according to confidence level grading conditions, updating the settlement status of project receivables, and generating accounting vouchers includes:

[0109] The confidence level criteria are calculated using a confidence scoring model to generate a confidence score for the project's receipts.

[0110] When the confidence score is ≥95%, batch confirmation of data and generation of confirmation logs are supported;

[0111] When 90% ≤ confidence score < 95%, generating a detailed list of settled project receipts facilitates personnel review and confirmation.

[0112] When the confidence score is less than 90%, generating a detailed list of matching discrepancies in settled project receivables facilitates personnel review and confirmation.

[0113] After batch confirmation and / or personnel confirmation, project receivables are marked as finally settled and accounting vouchers are generated.

[0114] In this embodiment of the invention, a confidence score model is used to score the project's payment collection process and then classify the resulting confidence scores to obtain corresponding confidence levels. For automatically matched settlement items, a confidence score is generated using an "AI confidence score model" (calculated based on 10 dimensions including matching dimension completeness, historical matching success rate, and data source credibility).

[0115] Confidence level ≥ 95% (high confidence): Supports batch confirmation by finance personnel (up to 100 transactions can be selected at a time), and the system automatically records batch confirmation logs;

[0116] 90% ≤ confidence level < 95% (medium confidence): Finance personnel need to review key information (bank details, accounts receivable details, matching basis) one by one before confirming;

[0117] Confidence level < 90% (low confidence): "Matching discrepancies" will be forcibly displayed (e.g., discrepancies in customer names, similar but unequal amounts, partially blurred invoice numbers). Manual verification by finance personnel is required (verification documents can be uploaded, such as screenshots of customer confirmation emails). Once finance personnel have performed batch confirmation or confirmed key or suspicious information one by one, the system will automatically update the accounts receivable status to "Final Settlement" and simultaneously lock the corresponding details (prohibiting duplicate processing). Unconfirmed items will be marked as "Pending Review" and displayed at the top of the finance dashboard, supporting filtering by "Confidence Level," "Customer Name," and "Duration of Unreview," improving review efficiency.

[0118] In one embodiment of the present invention, the generation of confidence scores includes:

[0119] The process of obtaining the data matching results in S2 is subjected to the first matching integrity A. x First match success rate A y The credibility of the primary data source (A) z The dimensional scoring is used to calculate the score A=M1A. x +M2A y +M3A z ;

[0120] The transaction scenario matching process of the data matching results in S3 is subjected to the second matching integrity B. x Second match success rate B y Credibility of the second data source (B)z The dimensional scoring is calculated to obtain the score B=M4B. x +M5B y +M6B z ;

[0121] The data verification process for the data matching results in the transaction scenario in S3 is performed using the third matching integrity C. x Third-party matching success rate C y Credibility of the third data source (C) z The dimensional scoring is calculated to obtain a score C=M7C. x +M8C y +M9C z ;

[0122] The identification process of the transaction scenario verification results in S4 is subjected to the fourth matching completeness D. x Fourth Match Success Rate D y Fourth, the credibility of the data source (D) z The dimensional scoring is used to calculate the score D=M. 10 D x +M 11 D y +M 12 D z ;

[0123] The process of completing the project receipt data in S5 is performed using the fifth matching integrity E. x Fifth match success rate E y Fifth, the credibility of the data source (E) z The dimensional scoring is used to calculate the score E=M. 13 E x +M 14 E y +M 15 E z ;

[0124] Confidence score = (A+B+C+D+E) / n;

[0125] Among them, M1-M 15 The weights for the dimension scores are: M1 + M2 + M3 = 1, where M1, M2, and M3 are all positive integers, and M1 > M2 ≥ M3; M4 + M5 + M6 = 1, where M4, M5, and M6 are all positive integers, and M5 > M4 ≥ M6; M7 + M8 + M9 = 1, where M7, M8, and M9 are all positive integers, and M9 > M7 ≥ M8; M 10 +M 11 +M 12 =1 and M 10 M 11 M 12 All are positive integers, M10 >M 11 ≥M 12 M 13 +M 14 +M 15 =1 and M 13 M 14 M 15 All are positive integers, M 15 >M 13 ≥M 14 n is the number of terms within the parentheses.

[0126] In this embodiment of the invention, the scoring of matching completeness, matching success rate, and data source credibility can include, but is not limited to, these three dimensions. The weight of each dimension varies across different processes, but the sum of the weights for each dimension is 1, and each weight is a positive integer. The weight ratio between M1:M2:M3 can be 40%:30%:30%; the weight ratio between M4:M2:M6 can be 30%:40%:30%; the weight ratio between M7:M8:M9 can be 30%:30%:40%; and the weight ratio between M... 10 M 11 M 12 The weighting ratio between them can be 40%:30%:30%; regarding M 13 M 14 M 15 The weighting ratio between these weights can be 30%:30%:40%. For example, if the data matching result for project payment is at the first level and matches an automatically settled transaction scenario, the process of obtaining the data matching result in S2 is scored, and the resulting score A = 40%. 96+30% 95+30% 97 = 96. The score for the transaction scenario matching process in S3 is B = 30%, based on the data matching results. 95+40% 95+30% 95 = 95. The data verification process for the data matching results in the transaction scenario of S3 is scored, resulting in a score of C = 30%. 97+30% 97+40% 97 = 97. The score for the identification process of the transaction scenario verification results in S4 is D = 40%. 96+30% 95+30% 97 = 96. The process of completing the project payment data in S5 is scored, resulting in a score E = 30%. 97+30% 97+40% 97=97, the confidence score of the project receipt is (96+95+97+96+97) / 5=96.2>95, so the confidence level of this project receipt is high confidence and can be included in the batch data confirmation step.

[0127] In this embodiment of the invention, the weights of the dimensional scoring are related to their respective processes. Specifically, in the process of obtaining data matching results in S2, more emphasis is placed on the matching process of data matching completeness; in the process of matching transaction scenarios with data matching results in S3, more emphasis is placed on the matching success rate of transaction scenarios; in the process of verifying data matching results of transaction scenarios in S3, more emphasis is placed on the process of verifying the credibility of the data source; in the process of identifying the verification results of transaction scenarios in S4, more emphasis is placed on the matching completeness of the verification results and the process of identifying settlement conditions; and in the process of supplementing project payment data in S5, more emphasis is placed on the process of supplementing the data source credibility.

[0128] In this embodiment of the invention, when a project payment is matched to the first level, a score of 95 or higher can be obtained. Alternatively, each item of the project payment data can be scored individually. For example, if customer information is complete and the name is the full name after unified identification, a score of 95 or higher can be obtained. If the identifier information is complete and matches the identifier information of the outstanding receivables in the accounts receivable data table, a score of 95 or higher can also be obtained. If the identifier information is only partially present, the score will be lower than 95. When a project payment is matched to the second level, a score of less than 95 can be obtained. Specifically, the score can be broken down into the sum of scores for each item of information. If the identifier information is missing but can be obtained directly or recovered through other means, a score between 90 and 95 can be obtained. If it cannot be obtained directly through other means and requires inquiry by financial personnel, the score will be lower than 90.

[0129] In this embodiment of the invention, during the transaction scenario matching process, a transaction scenario with automatic settlement can obtain a score of 95 or higher. When a transaction scenario involving the splitting and / or totaling of amount information is matched, if the amount information can be reconciled during data verification after splitting and / or totaling, and the identifier information can be matched, a score less than 95 but greater than 90 can be obtained, such as scenarios A, C, F, and G. If the reconciliation still fails and the identifier information is still missing, requiring data completion, a score less than 90 will be obtained, such as partial operations in scenario B, scenario D, and scenario E. When an abnormal transaction scenario is matched, scenarios H, I, and K can obtain a score less than 95 but greater than 90, while scenario J obtains a score less than 90.

[0130] In this embodiment of the invention, the confidence score of the project receipts can be obtained by adding up the obtained scores and taking their average. The confidence level is then classified according to the scores. The data confirmation process is different for different levels of data. When the score reaches a high confidence level, batch confirmation can be performed to improve the reconciliation efficiency.

[0131] In one embodiment of the present invention, the automatic accounts receivable settlement management method further includes an active reconciliation mechanism and an incentive mechanism, wherein the active reconciliation mechanism includes:

[0132] Generate a monthly reconciliation statement for project receipts on the set date;

[0133] The online reconciliation module sends the project payment details to the client for confirmation.

[0134] Synchronize the customer's confirmed results and / or feedback information to the first platform;

[0135] The incentive mechanism includes:

[0136] The behavior analysis module identifies and counts the completeness and quantity of identifier information for project payments.

[0137] A reconciliation reward mechanism is triggered for customers who provide complete identifier information N times consecutively.

[0138] For customers who provide incomplete and / or missing identifier information M times consecutively, send an identifier information prompt.

[0139] In this embodiment of the invention, the proactive reconciliation mechanism can be set to automatically generate "customer personalized reconciliation details" (including outstanding invoices, settled invoices, early payments, and WHT deduction records, supporting download in both PDF and Excel formats) at midnight on the 1st of each month, and send them to the customer's finance email address via encrypted email (using SMTP protocol for encrypted transmission); at the same time, an "online reconciliation link" (based on HTTPS protocol, allowing customer finance personnel to log in with their account and password, view details in real time, confirm reconciliation results, and the feedback information is immediately synchronized to the enterprise's financial system); for customers who have not confirmed within 72 hours, the system automatically sends a reconciliation reminder email.

[0140] In this embodiment of the invention, the reward mechanism mainly uses the "Customer Behavior Analysis" module to automatically identify whether a customer's "payment summary contains a complete invoice number." For customers who provide a complete invoice number for N consecutive times (3 or more consecutive times), the "Incentive Rule Engine" is triggered: 1. A 1.5% discount is given on the next quarter's orders (the system automatically updates the discount rate for this customer in the ERP pricing module, valid for 90 days); 2. The payment period is extended by 5 days (the customer's credit policy is updated synchronously, and "Incentive Payment Period Extension" is marked in the "Customer Information Database"); 3. An electronic reward notification (including reward details, validity period, usage method, and an official electronic signature from the company) is sent to the customer's finance manager via email and SMS. For customers who fail to provide an invoice number for M consecutive times (5 or more consecutive times), the system automatically pushes a "Guidance Email" (with a case study explaining that "providing an invoice number can speed up reconciliation and reduce payment disputes"), reducing the difficulty of subsequent matching.

[0141] like Figure 3 As shown, this embodiment of the invention also provides an automatic accounts receivable settlement management device, which includes:

[0142] The preprocessing module acquires project payment data and preprocesses it.

[0143] The hierarchical matching module performs hierarchical data matching on the preprocessed project payment data according to multi-level matching rules to obtain the data matching results.

[0144] The transaction scenario matching module matches the data matching results with transaction scenarios based on the scenario matching conditions, and performs data verification on the data matching results of the transaction scenarios;

[0145] The settlement module identifies the verification results of the transaction scenario based on the settlement conditions. If the settlement conditions are met, the project payment is settled.

[0146] If the settlement conditions are not met, the data completion module will complete the data for project payments. The processing results after data completion will be processed in steps S3-S4.

[0147] The voucher generation module classifies settled project receipts according to confidence level criteria and confirms data, updates the settlement status of project receipts, and generates accounting vouchers.

[0148] The data synchronization module marks the reconciliation status of the project receipts that generate accounting vouchers, generates a matching number, and synchronizes it to the first platform.

[0149] In this embodiment of the invention, the above-mentioned modules are modularized as the accounts receivable automatic settlement management method provided by the present invention. For the explanation of each module, please refer to the content of the accounts receivable automatic settlement management method section of the present invention. This embodiment will not repeat it here.

[0150] like Figure 4 As shown, this embodiment of the invention also provides an automatic accounts receivable settlement management system, which includes: a management server and an ERP system;

[0151] The management server is used to execute the automatic accounts receivable settlement management method described above;

[0152] The ERP system communicates with the management server to receive data from the management server.

[0153] In this embodiment of the invention, preferably, the test server may include, but is not limited to, a system or virtual server within a computer device that executes the automatic accounts receivable settlement management method. The ERP system may be an internal financial management system of an enterprise, a system platform that receives data from the management server to facilitate data confirmation of project receipts and data traceability based on matching numbers.

[0154] like Figure 5 The diagram shown is an internal structural diagram of the computer device in this embodiment. Figure 5 As shown, the computer device includes a processor, memory, network interface, input device, and display screen connected via a system bus. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and may also store a computer program. When executed by the processor, the computer program enables the processor to implement the visual detection method provided in this embodiment of the invention. The internal memory may also store a computer program. When executed by the processor, the computer program enables the processor to execute the accounts receivable automatic settlement management method provided in this embodiment of the invention. The display screen of the computer device can be a liquid crystal display screen or an e-ink display screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, a trackball, or a touchpad provided on the casing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0155] Those skilled in the art will understand that Figure 5 The structures shown are merely block diagrams of some structures related to the present invention and do not constitute a limitation on the computer devices on which the present invention is applied. Specific computer devices may include more or fewer components than those shown in the figures, or combine certain components, or have different component arrangements.

[0156] In one embodiment, the accounts receivable automatic settlement management device provided by this invention can be implemented as a computer program, which can be implemented in the form of, for example... Figure 4 The computer device shown operates on this device. The computer device's memory can store the various program modules that make up the automatic accounts receivable settlement management device, for example, Figure 3 The diagram shows a preprocessing module, a hierarchical matching module, a transaction scenario matching module, a settlement module, a data completion module, a voucher generation module, and a data synchronization module. The computer program comprised of these modules enables the processor to execute the steps of the automatic accounts receivable settlement management method described in the various embodiments of the present invention.

[0157] For example, Figure 5 The computer device shown can be used as follows Figure 3 The preprocessing module in the automatic accounts receivable settlement management device shown executes step S1; the computer device can execute step S2 through the hierarchical matching module; the computer device can execute step S3 through the transaction scenario matching module; the computer device can execute step S4 through the settlement module; the computer device can execute step S5 through the data completion module; the computer device can execute step S6 through the voucher generation module; and the computer device can execute step S7 through the data synchronization module.

[0158] In one embodiment, a computer device is provided, the computer device including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, performs the following steps:

[0159] S1: Obtain project payment data and preprocess the project payment data;

[0160] S2: Perform hierarchical data matching on the preprocessed project payment data according to multi-level matching rules to obtain the data matching results;

[0161] S3: Match the data matching results to the transaction scenario based on the scenario matching conditions, and perform data verification on the data matching results of the transaction scenario;

[0162] S4: Identify the verification results of the transaction scenario based on the settlement conditions. If the settlement conditions are met, settle the project payment.

[0163] S5: If the settlement conditions are not met, the data for project receipts will be supplemented. For the processing results after data supplementation, S3-S4 will be executed.

[0164] S6: Classify and confirm the settled project receivables according to the confidence level classification conditions, update the settlement status of project receivables and generate accounting vouchers.

[0165] S7: Mark the reconciliation status of the project receipts for which accounting vouchers are generated, generate a matching number, and synchronize it to the first platform.

[0166] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which, when executed by a processor, causes the processor to perform the following steps:

[0167] S1: Obtain project payment data and preprocess the project payment data;

[0168] S2: Perform hierarchical data matching on the preprocessed project payment data according to multi-level matching rules to obtain the data matching results;

[0169] S3: Match the data matching results to the transaction scenario based on the scenario matching conditions, and perform data verification on the data matching results of the transaction scenario;

[0170] S4: Identify the verification results of the transaction scenario based on the settlement conditions. If the settlement conditions are met, settle the project payment.

[0171] S5: If the settlement conditions are not met, the data for project receipts will be supplemented. For the processing results after data supplementation, S3-S4 will be executed.

[0172] S6: Classify and confirm the settled project receivables according to the confidence level classification conditions, update the settlement status of project receivables and generate accounting vouchers.

[0173] S7: Mark the reconciliation status of the project receipts for which accounting vouchers are generated, generate a matching number, and synchronize it to the first platform.

[0174] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0175] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0176] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0177] The embodiments described above are merely examples of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention.

Claims

1. An automatic settlement management method of receivables, characterized by, The automatic clearing management method of the receivables comprises: S1: obtaining data of project collection items and preprocessing the data of the project collection items; S2: performing hierarchical data matching on the preprocessed data of the project collection items according to multi-level matching rules to obtain data matching results; S3: matching the data matching results with transaction scenarios according to scenario matching conditions and performing data verification on the data matching results of the transaction scenarios; S4: identifying the verification results of the transaction scenarios according to clearing conditions, and if the clearing conditions are met, clearing the project collection items; S5: if the clearing conditions are not met, completing the data of the project collection items, and performing S3-S4 on the processing results after the data completion; S6: classifying and data confirming the cleared project collection items according to confidence level classification conditions, updating the clearing status of the project collection items, and generating accounting vouchers; S7: marking the project collection items with generated accounting vouchers with account status, generating matching numbers, and synchronizing to the first platform. The matching of the data matching results with transaction scenarios according to scenario matching conditions comprises: The hierarchical data matching results are respectively a first level, a second level, and a third level; The first level is a case where customer information, identifier information, and amount information are correct and have no missing, which is matched as an automatic clearing transaction scenario; The second level is a case where customer information is correct and has no missing, identifier information is incorrect or missing, and amount information has no missing but is incorrect, which is matched as an amount information splitting and / or totaling transaction scenario; The third level is a case one where customer information is incorrect, identifier information and amount information are correct and have no missing, a case two where customer information and amount information are correct and have no missing, and identifier information is correct but partially missing, a case three where customer information and identifier information are correct and have no missing, and amount information is incorrect, and a case four of abnormal transaction, which are all matched as abnormal transaction scenarios.

2. The method of claim 1, wherein the step of automatically clearing the receivables is performed by a computer. The hierarchical data matching of the preprocessed data of the project collection items according to multi-level matching rules to obtain data matching results comprises: Obtaining the data of the project collection items through a timed trigger mode and / or an event trigger mode; Filtering invalid transactions in the project collection items through a natural language processing module, and dividing words in the data of the project collection items, determining words associated with at least one of customer name, transaction amount, abstract text, and transaction time from the divided words; Merging the associated words and the synonymous words, the near-synonymous words, or the antonyms corresponding to the associated words into unified information identifiers to obtain information of at least one of the customer name, the transaction amount, the abstract text, and the transaction time; Completing information of missing elements through an associated customer information library; Performing hierarchical data matching on the information after the information completion according to multi-level matching rules, and obtaining data matching results containing combinations of at least two of customer information, amount information, and identifier information.

3. The method of claim 1, wherein the step of automatically clearing the receivables is performed by a computer system. The amount information splitting and / or totaling transaction scenario comprises: When the amount information of the project payment is greater than the single pending payment amount information in the project receivable data table and the identifier information is missing, including scenario A of multiple pending payment amount information, scenario B of customer supplier portal payment step, and scenario C of no pending payment amount information; When the amount information of the project payment is less than the single pending payment amount information in the receivable data table and the identifier information is missing, including scenario D of deducting pre-tax, and scenario E of no deducting pre-tax; When the amount information of the project payment is equal to the total of multiple pending payment amount information in the receivable data table and the identifier information is missing, it is scenario F of simultaneous settlement of multiple pending payment amount information; When the amount information of the project payment is equal to the total of the pending payment amount information in multiple receivable data tables and the identifier information is missing, it is scenario G of splitting the amount information of the payment; The abnormal transaction scenarios include scenario H of ambiguous identifier information, scenario I of refund association, scenario J of abnormal transaction anti-fraud, and scenario K of multi-currency transaction.

4. The method of claim 3, wherein the step of automatically clearing the receivables is performed by a computer. The data matching results of the amount information combination and / or splitting transaction scenarios are data checked, including: Scenario A, through the multiple combination matching algorithm, the remaining pending payment details obtained by the traversal search are combined to obtain the total result, which is data checked with the amount information of the project payment, and the identifier information of the amount information combination is recorded; Scenario B, through the key management module, the payment notification is called to obtain the amount information and the identifier information on the payment notification and data check with the pending payment amount information of the project receivable data table. If it is consistent with the pending payment amount information, S4 is executed, and if it is inconsistent with the pending payment amount information, S5 is executed; Scenario C, through the receivable balance real-time checking module, it is confirmed that there is no pending project receivable, S6-S7 are executed, and the missing identifier information is recorded; Scenario D, through the tax database, the pre-tax tax rate is called for calculation, and the calculation result is data checked with the pending payment amount information. If it is consistent with the pending payment amount information, the difference is filled and S4 is executed, and if it is inconsistent with the pending payment amount information, the mark is recorded and S5 is executed; Scenario E, the project payment is marked and auxiliary information is added, and S5 is executed; Scenario F, through the exception rule module, the specified pending payment amount information is preferentially matched for data checking, and the remaining project payment amount information is matched and data checked according to the ascending order of the date of the pending payment amount information; Scenario G, through the identification and / or splitting module, multiple customer information is identified and obtained, the amount information of the project payment is split according to the identifier information, and the amount information of each item of the project payment after splitting is data checked with the pending payment amount information in the receivable data table.

5. The method of claim 3, wherein the step of automatically clearing the receivables is performed by a computer. The data matching results of the abnormal transaction scenarios are data checked, including: Scenario H, the identifier information of the project collection is fuzzy searched by the fuzzy search algorithm, a number of complete identifiers are matched and matched suggestion list is generated to facilitate personnel confirmation, the confirmed identifier information is checked with data, and S4 is executed; Scenario I, the project receivables details of the refund transaction are identified by the refund transaction identification module, the reverse clearing voucher is generated, and the project receivables are restored to the state of receivables, the state of the restored receivables is marked and data checking is performed; Scenario J, link the anti-fraud database model, and perform anti-fraud verification after obtaining the project collection, if at least one of the abnormal information of the stranger account information and the high-risk area payment is detected, trigger the early warning and stop the matching process, if no abnormality is confirmed, the subsequent step is performed, if the abnormality is confirmed, the mark is generated and the refund application form is generated; Scenario K, the amount information of the project collection of multi-currency payment is converted by the real-time exchange rate interface module, the converted result is data accounted with the receivable amount information, if there is exchange rate difference, the difference is recorded and included in the exchange loss and gain, and S4 is executed.

6. The method of claim 1, wherein the step of automatically clearing the receivables is performed by a computer. The project collection is classified and data confirmed according to the confidence level classification condition, the clearing state of the project receivables is updated, and the accounting voucher is generated, including: The confidence level classification condition is calculated by the confidence score model to generate a confidence score; When the confidence score is greater than or equal to 95%, the batch confirmation of the data is supported and the confirmation log is generated; When the confidence score is between 90% and 95%, the details of the cleared project collection are generated to facilitate personnel review and confirmation; When the confidence score is less than 90%, the matching suspicious point details in the cleared project collection are generated to facilitate personnel review and confirmation; The project collection after batch confirmation and / or personnel confirmation is marked as the final clearing state of the project receivables and the accounting voucher is generated.

7. The method of claim 6, wherein the step of automatically clearing the receivables is performed by a computer. The confidence score is generated, including: The process of matching the data obtained in S2 is scored in the dimensions of the first matching integrity A x , the first matching success rate A y , and the first data source credibility A z , and the score A=M1A x +M2A y +M3A z is calculated. The transaction scenario matching process of the data matching result in S3 is respectively scored in the dimensions of second matching integrity B x , second matching success rate B y , and second data source credibility B z , and the score B is calculated as M4B x +M5B y +M6B z . Data checking process on data matching result of transaction scenario in S3 respectively carries out third matching integrity C x , third matching success rate C y , and third data source credibility C z dimension score, and the score C=M7C x +M8C y +M9C z is calculated. The identification process of the checking result of the transaction scenario in S4 is respectively scored in the fourth matching integrity D x , the fourth matching success rate D y , the fourth data source credibility D z dimension score, and the score D=M 10 D x +M 11 D y +M 12 D z is calculated. The process of data completion of the project collection in S5 is respectively scored in the dimensions of the fifth matching integrity E x , the fifth matching success rate E y , the fifth data source credibility E z , and the score E=M 13 E x +M 14 E y +M 15 E z is calculated. The confidence score=(A+B+C+D+E) / n; M1-M 15 M1+M2+M3=1 and M1, M2, M3 are positive integers, M1>M2≥M3; M4+M5+M6=1 and M4, M5, M6 are positive integers, M5>M4≥M6; M7+M8+M9=1 and M7, M8, M9 are positive integers, M9>M7≥M8; M 10 +M 11 +M 12 =1 and M 10 , M 11 , M 12 are positive integers, M 10 >M 11 ≥M 12 ; M 13 +M 14 +M 15 =1 and M 13 , M 14 , M 15 are positive integers, M 15 >M 13 ≥M 14 ; n is the number of items in the parentheses.

8. An automatic settlement management device for receivables, characterized by comprising: The automatic clearing management device for receivables includes: A preprocessing module acquires data of project collection and preprocesses the data of project collection; A hierarchical matching module performs hierarchical data matching on the preprocessed data of project collection according to multi-level matching rules to obtain data matching results; A transaction scenario matching module matches the data matching results according to the scenario matching condition, and performs data checking on the data matching results of the transaction scenario; A clearing module identifies the checking results of the transaction scenario according to the clearing condition, and clears the project collection if the clearing condition is met; A data completion module performs data completion of the project collection if the clearing condition is not met, and executes S3-S4 on the processing results after data completion; A voucher generation module classifies and data confirms the cleared project collection according to the confidence level classification condition, updates the clearing state of the project collection, and generates an accounting voucher; A data synchronization module marks the project collection generating the accounting voucher as the account state, generates a matching number, and synchronizes to the first platform.

9. An automatic write-off management system for receivables, characterized by The automatic clearing management system for receivables includes a management server and an ERP system; The management server is used for executing the automatic clearing management method of receivables as claimed in any one of claims 1-7. The ERP system communicates with the management server to receive data of the management server.

Images