Banking business-oriented financial asset intelligent redemption method
Through the AI engine training of the redemption logic of financial management products, the problems of complex channel access and user operations under diversified financial management products are solved, low-coupling and efficient intelligent redemption is achieved, operation difficulty and user operation complexity are reduced, and new rules are supported quickly expanding.
Patent Information
- Application Number
- CN202510452744.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-11
- Publication Date
- 2025-08-08
AI Technical Summary
When facing diversified wealth management products, the existing wealth management agency platform has high channel access and maintenance costs, user redemption operations are complex, operational efficiency is low, and new rules or restrictions require long-term development and testing.
The AI engine is used to train the redemption logic of financial products, and the AI engine is used to train the element information and custom configuration rules of financial products to realize intelligent redemption, including the calculation of the redemption business date, confirmation date and arrival date, as well as whether the current product can be redeemed, the redemption upper limit and the minimum holding requirements, etc. After the user enters the redemption application, the AI engine filters the products that do not meet the conditions and sorts them according to priority, and automatically allocates the redemption amount.
It reduces channel access costs, simplifies user operations, improves operational efficiency, reduces user operation time, and supports rapid expansion of new rules without code modification.
Smart Images
Figure CN120450869A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of financial asset redemption, and in particular to an intelligent redemption method for financial assets oriented to banking business. Background Art
[0002] Since the launch of the wealth management distribution platform, the wealth management products it supports have expanded from a single type to multiple, and the number of partnering wealth management subsidiaries has increased from one to multiple, resulting in a continuous increase in both product variety and quantity. With the diversification of wealth management products, redemption rules for different products have become increasingly distinct. For example, redemption availability, redemption confirmation, and redemption receipt times vary. Furthermore, some products may have logical restrictions such as daily user redemption limits, daily total redemption limits, and even redemption suspensions. To accommodate these differences, the platform has developed multiple redemption interfaces for different products since its inception. However, this also requires product terminals (channels) to connect to different redemption service interfaces for each product, leading to increasing channel access and maintenance costs.
[0003] At the same time, different wealth management products have varying yields, and some may have short-term redemption restrictions. To ensure financial security and stable returns, users often prefer to hold multiple wealth management products to mitigate risk through diversification. However, when users need to redeem funds, they must comprehensively consider multiple factors, including the yield profile, product status, and redemption quotas of all products held, individually reviewing and determining which products are currently redeemable. For example, to maximize returns, users may prefer to redeem products with lower recent yields while avoiding products with quota restrictions or those that are temporarily unavailable for redemption. These complex procedures undoubtedly increase the user burden.
[0004] To maintain a healthy business relationship with wealth management subsidiaries, wealth management distribution platforms need to coordinate user holdings of different products and maintain a dynamic balance. However, the platform cannot directly intervene in user redemption behavior. Operations staff can adjust the order in which products are displayed (for example, by placing certain products at the top) to increase their likelihood of selection, but this approach is generally limited in effectiveness. Furthermore, each new rule or restriction requires code development and release, which often takes a long time from development to testing and launch. Summary of the Invention
[0005] The purpose of the present invention is to provide a banking-oriented intelligent redemption method for financial assets in response to the deficiencies in the prior art.
[0006] To achieve the above objectives, the present invention provides a banking-oriented intelligent redemption method for wealth management assets, comprising:
[0007] Input the essential information of the wealth management product and the custom configured rules into the AI engine for training, so that the AI engine can understand the business logic of redemption of various wealth management products. The business logic includes the calculation method of the redemption business date, confirmation date and arrival date, as well as whether the current wealth management product is redeemable at the current time, the redemption limit, whether the remaining wealth management shares meet the minimum holding requirements, and whether the redeemed shares meet the minimum redemption standards;
[0008] When a user submits a redemption application to the wealth management front-end component, it queries the specified type or all wealth management products of the user based on the redemption application information entered by the user, and calls the AI engine to filter out wealth management products that do not meet the conditions. The details of the redeemable wealth management products are then sorted from high to low according to priority. The AI engine calculates the redemption amount allocation based on the redemption application information and the holdings of the redeemable wealth management products, generates business data for each redemption, generates the final redemption transaction flow, and saves it.
[0009] Furthermore, the essential information of the product includes TA information, product information under TA, opening period information of each product, redemption rule information, and working day information of TA and product.
[0010] Furthermore, the custom configuration rules include any combination of one or more of the following:
[0011] Configure several financial products and specify the redemption order, or sort them according to rules;
[0012] Configure several product types, and sort them according to rules;
[0013] Configure several wealth management subsidiaries and specify the redemption order;
[0014] As a default rule, redemptions are made from low to high product yields;
[0015] Configure the daily redemption quota for a specific TA, product type or product.
[0016] Furthermore, when training the AI engine, input rules limit the number of redemptions by the same user within a unit of time to control abnormal redemption behavior.
[0017] Furthermore, the redemption application information includes the redeemer's unique identity identifier, redemption shares, and redemption funds repayment bank account information.
[0018] Beneficial Effects: 1. The present invention can encapsulate the redemption service into an independent component for different types of financial products, supports subsequent expansion, and provides a unified API interface to reduce channel access costs. The component is designed with low coupling and high cohesion, and can be independently iterated and upgraded. The component's redemption calculation is based on AI, and by training rules on financial product data, it can achieve comprehensive support for product redemption calculation. For the redemption rules of newly added products, only new rules need to be provided for AI training and verification, without the need for repeated code development.
[0019] 2. This invention can automatically obtain the assets held by the user and intelligently allocate the redemption amount based on the redemption amount entered by the user. It can identify redeemable products and prioritize redemption of products with lower yields. At the same time, the component can automatically calculate the redemption amount for each held product based on established rules (such as quota limits), significantly reducing the complexity of user operations and reducing operation time from 5-10 minutes to less than 1 minute.
[0020] 3. The automatic redemption calculation function is based on the rule configuration of the components, and can prioritize the redemption of designated products, thereby helping the platform to efficiently manage product holdings and reduce operational difficulties. Due to the use of AI calculation, when adding new rules, only additional training and verification of the AI is required, without the need to modify the code and republish it, which greatly improves operational efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] Figure 1 is a schematic diagram of a smart redemption component according to an embodiment of the present invention;
[0022] Figure 2 It is a schematic diagram of the implementation class of the smart redemption component of an embodiment of the present invention. DETAILED DESCRIPTION
[0023] The present invention will be further illustrated below with reference to the accompanying drawings and specific embodiments. These embodiments are implemented based on the technical solutions of the present invention. It should be understood that these embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.
[0024] An embodiment of the present invention provides a banking-oriented intelligent redemption method for wealth management assets, comprising:
[0025] Input the essential information of the wealth management products and the custom configured rules into the AI engine for training, so that the AI engine can master the business logic of redemption of various wealth management products. The business logic includes the calculation method of the redemption business date, confirmation date and arrival date, as well as whether the current wealth management product can be redeemed at the current time, the redemption upper limit, whether the remaining wealth management shares meet the minimum holding requirements, and whether the redeemed shares meet the minimum redemption standards. Specifically, the essential information of the above products includes TA information, product information under TA, opening period information of each product, redemption rule information, TA and product working day information. Simply input these rules into the AI, and after the training is completed, the AI can calculate and output the redemption trial data. When adding a new product type or redemption rule, simply input the new rule into the AI for training, and the new rule can be adapted and passed the verification test without modifying the code or republishing.
[0026] For example:
[0027] Currently, wealth management products are mainly divided into three types: 1. Closed-end net value products; 2. Currency net value products; and 3. Cyclical net value products. Cyclical net value products are further divided into fixed-opening period products and minimum holding period products.
[0028] Closed net value products: do not support redemption transactions, only redemption upon maturity, and can be ignored during AI engine training.
[0029] For products with a net asset value (NAV) structure, there are no restrictions on redemption days. The only requirement is to verify that the TA and product status meet normal redemption conditions. Based on the financial management business day and the product's market closing time, if the current time has exceeded the current market closing time, the redemption transaction date is the next business day; otherwise, the current business day is used. The redemption confirmation date is calculated based on the redeem_cfm_n parameter configured in the product table: the redemption transaction date + redeem_cfm_n business days. The redemption funds arrival date is calculated based on the redeem_pay_period parameter configured in the product table: the redemption transaction date + redeem_pay_period business days.
[0030] For periodic net asset value products: The redemption opening date must be calculated using the product period opening information table, which contains the product code, redemption start date, redemption end date, and redemption confirmation date. The AI must find the opening date interval within which the current time falls (i.e., redemption start date ≤ current date ≤ redemption end date). The business date is the end date of this interval, and the redemption confirmation date is the corresponding value in the table. The redemption funds arrival date is still calculated based on the redeem_pay_period, which is the redemption business date + redeem_pay_period business days. If the current time does not match the opening date interval, the product is not in the redemption opening period and cannot be redeemed. Additionally, the TA status and product status must be verified to ensure that they meet the requirements.
[0031] Minimum Holding Period Products: These products are periodic net asset value products with no fixed opening date. Redemption conditions are based on the user's holding period. After confirming a purchase, users must meet the minimum holding period. After reaching this period, they can initiate redemption at any time. Subsequent processing rules are the same as for monetary net asset value products.
[0032] New product support: If the system adds a new product type or feature, it can be adapted by simply training the AI with additional training data without modifying the code.
[0033] The above custom configuration rules include any combination of one or two or more of the following:
[0034] 1) Products with priority for redemption: You can configure several products and specify the redemption order, or sort them according to rules (such as the most recent yield).
[0035] 2) Product types for priority redemption: Several product types can be configured, and they can be sorted within each type based on rules (such as yield). For example, products of wealth management subsidiary A can be prioritized for redemption, and they can also be sorted from low to high by yield.
[0036] 3) Priority redemption of designated wealth management subsidiaries: You can configure several wealth management subsidiaries and specify the order of redemption.
[0037] 4) Redemption in ascending order of yield: As a default rule, redemption is carried out in ascending order of product yield.
[0038] 5) Daily redemption quota configuration: You can configure a daily redemption quota for a specific TA, product type, or product. When using the AI engine to allocate redemption amounts, ensure that the configured quota is not exceeded.
[0039] When users submit redemption requests to the wealth management front-end component through the banking app, the Star Chart Financial banking zone, or the WeChat mini-program, the component queries the specified type of wealth management products or all of the user's wealth management products based on the redemption request information entered. The AI engine is then used to filter out ineligible wealth management products. Key criteria include: wealth management product status (must meet redemption requirements), wealth management subsidiary status (not deactivated), normal wealth management account status, normal status of associated bank accounts, product within the redemption period, supported redemption channels, and a remaining redemption limit greater than 0. Filtering criteria vary by TA or product type. The AI engine is trained to understand the redemption requirements of various products and quickly eliminates ineligible products.
[0040] The AI engine then sorts the details of redeemable wealth management products from high to low priority, for example, sorting them in ascending order of yield, prioritizing products with lower yields. The system also supports the configuration of prioritized or lower-priority products for redemption, enabling operational control of product holdings through redemption priority design. Specific rules include prioritizing products for redemption, prioritizing product types, prioritizing wealth management subsidiaries, and sorting products from lowest to highest yield, and these rules can be expanded based on business needs.
[0041] The AI engine calculates the redemption amount allocation based on the redemption application information and the holdings of the redeemable wealth management product, generates business data for each redemption, generates the final redemption transaction flow and saves it. Specifically, the AI engine calculates the redemption information based on the sorted redeemable holding details, and deducts the required redemption assets one by one. First, the purchase amount of the day is deducted. If it is insufficient, the confirmed shares are deducted until the full deduction is completed, and the calculation result is returned. If the redeemable shares are less than the user's applied shares, the actual redeemable shares are returned and the user is prompted. In addition, for the redemption of products with different holding details, the redemption business data is calculated according to the respective product types, such as the business date, confirmation date, arrival date, etc.
[0042] When training the AI engine, we input rules to limit the number of redemptions a user can make within a specific timeframe to control abnormal redemption behavior. These extended features can be implemented through AI training, eliminating the need for additional code development. Simply input the rule data and complete the training and verification process.
[0043] In the specific implementation, the above method is completed through the developed intelligent redemption component. The intelligent redemption component mainly designs and implements two modules: API service module and rule management configuration module. The API service module provides a unified redemption API interface to the outside world and is responsible for receiving and processing redemption applications. This module relies on the AI engine and obtains the final redemption data by calling the AI engine interface. The rule management configuration module is mainly responsible for the configuration of special redemption rules, and inputs the rule data into the AI engine, which is responsible for learning and applying it. The AI engine obtains the necessary business data for learning by accessing the agency wealth management business database. The dependency relationship of the intelligent redemption component is as follows: Figure 1 To enhance the versatility of the smart redemption component, the redemption interface supports multiple selection modes.
[0044] The input must:
[0045] Specify the unique identity of the redeemer: for example, the customer number, which is used to uniquely identify the redeemer.
[0046] Specify redemption shares: specify the specific shares or quantity to be redeemed.
[0047] Redemption Fund Return Bank Account Information: Provide the bank account details to which the redemption funds will be returned.
[0048] Optional inputs:
[0049] Redemption product type: Specify the product category to be redeemed (such as money market funds, bond funds, etc.).
[0050] Redemption product code: Use the product code to identify the redeemed product (this method is not recommended as it may be inconvenient to use due to code changes or complexity).
[0051] If a redemption request doesn't specify a specific product, the Smart Redemption component automatically retrieves all the user's holdings across all product types and allocates the redemption amount to each product based on pre-set priority rules. If the redemption product types are restricted, the Smart Redemption component only retrieves the user's holdings of a specific product type and allocates the redemption amount. The component also supports specifying a specific product for redemption, eliminating the need for automatic allocation and effectively executing a manually specified redemption.
[0052] See also Figure 2 The following is a detailed description of the implementation class and core parameters of the smart redemption component:
[0053] The TDServiceBase redemption API exposes the interface, encapsulating the request parameters in the parameter object TRrequest. This includes the core parameters: customer number custNo, redemption request amount redeemMoney, product type prodType, and product code prodCode.
[0054] The TS510052Service implementation class of the TDServiceBase interface supports fast and intelligent redemption applications, primarily involving advance payment. It ensures real-time deposit of the user's redemption amount within the system's configured limit. This generally applies to product types, supporting only products with a net asset value of 1. Because advance payment amounts must be specified, only wealth management products with a net asset value of 1 are supported. This implementation relies on TdFastRedeemTrialServiceImpl, an implementation class of the TdRedeemTrialService trial calculation interface, to register redemption transactions based on trial calculation results.
[0055] The TS510054Service implementation class of the TDServiceBase interface supports standard smart redemption applications, including redemption applications for more products and larger financial assets. This implementation relies on the TdRedeemTrialServiceImpl implementation class of the trial calculation interface to register redemption transactions based on the trial calculation results.
[0056] TdRedeemTrialService is the redemption trial service interface, which includes TdFastRedeemTrialServiceImpl, a fast smart redemption trial implementation, and TdRedeemTrialServiceImpl, a standard smart redemption trial implementation. The TdFastRedeemTrialServiceImpl class relies on the RedeemTriaIntf trial interface to implement the FastRedeemTriaImpl class. The TdRedeemTrialServiceImpl class relies on the RedeemTriaIntf trial interface to implement the RedeemTriaImpl class.
[0057] CustVolService is an asset query interface. It has an asset query implementation class, CustVolServiceImpl, which contains asset query methods: getCustVols(String custNo), getCustVols(String custNo, String prodType), and getCustVolByProdCode(String custNo, String prodCode). Other methods for querying specific assets can be extended here as needed.
[0058] The AssetSorting class is responsible for sorting the queried asset holdings according to the sorting priority rules configured in the component. It supports multiple sorting rules and can be upgraded based on business needs.
[0059] RedeemTriaIntf is a trial calculation interface, which includes the fast intelligent transfer trial calculation implementation class FastRedeemTriaImpl and the common intelligent transfer trial calculation implementation class RedeemTriaImpl. <custvol>custVols) performs redemption calculations and allocations on the details of different financial assets held by users and returns the calculation results.
[0060] The above description is merely a preferred embodiment of the present invention. It should be noted that any other aspects not specifically described are considered prior art or common knowledge to those skilled in the art. Improvements and modifications may be made without departing from the principles of the present invention, and such improvements and modifications are also within the scope of protection of the present invention.< / custvol>
Claims
1. A smart redemption method for financial assets for banking business, characterized in that: include: Input the essential information of the wealth management product and the custom configured rules into the AI engine for training, so that the AI engine can understand the business logic of redemption of various wealth management products. The business logic includes the calculation method of the redemption business date, confirmation date and arrival date, as well as whether the current wealth management product is redeemable at the current time, the redemption limit, whether the remaining wealth management shares meet the minimum holding requirements, and whether the redeemed shares meet the minimum redemption standards; When a user submits a redemption application to the wealth management front-end component, it queries the specified type or all wealth management products of the user based on the redemption application information entered by the user, and calls the AI engine to filter out wealth management products that do not meet the conditions. The details of the redeemable wealth management products are then sorted from high to low according to priority. The AI engine calculates the redemption amount allocation based on the redemption application information and the holdings of the redeemable wealth management products, generates business data for each redemption, generates the final redemption transaction flow, and saves it.
2. The intelligent redemption method for financial assets for banking business according to claim 1, characterized in that: The essential information of the product includes TA information, product information under TA, opening period information of each product, redemption rule information, and working day information of TA and product.
3. The intelligent redemption method for financial assets for banking business according to claim 1, characterized in that: The custom configuration rules include any combination of one or more of the following: Configure several financial products and specify the redemption order, or sort them according to rules; Configure several product types, and sort them according to rules; Configure several wealth management subsidiaries and specify the redemption order; As a default rule, redemptions are made from low to high product yields; Configure the daily redemption quota for a specific TA, product type or product.
4. The intelligent redemption method for financial assets for banking business according to claim 1, characterized in that: When training the AI engine, input rules limit the number of redemptions by the same user per unit time to control abnormal redemption behavior.
5. The intelligent redemption method for financial assets for banking business according to claim 1, characterized in that: The redemption application information includes the redeemer's unique identity identifier, redemption shares, and redemption funds repayment bank account information.