Machine learning system for routing payment requests to different customer accounts

Patent metadata

Field	Value
Publication number	US-11829985-B1
Application number	US-202017038554-A
Country	US
Kind code	B1
Filing date	Sep 30, 2020
Priority date	Sep 30, 2020
Publication date	Nov 28, 2023
Grant date	Nov 28, 2023

Abstract

A transaction approval system makes an assessment in real time whether to approve a purchase transaction and to select an account of the customer to fund the purchase transaction. The transaction approval system can use a machine learning model to select the appropriate account of the customer to fund a purchase transaction that satisfies one or more preferences or goals established by the customer or the system. The machine learning model can receive, as inputs, information on one or more of the customer's accounts, user preference information, information regarding the transaction and/or information regarding prior transactions by the customer and use the inputs and machine learning model to make a selection of a customer account that satisfies the preferences or goals established by the customer or system.

First claim

What is claimed is:

1. A system for processing a purchase transaction by a customer, the system comprising:
a memory configured to store a plurality of accounts and, in association with each of the plurality of accounts, (i) a current account balance, and (ii) activity data relating to historical activity, by a customer, on the account; and
at least one processor configured to:
receive, from a third-party system, a request to approve a transaction of a requested value, the transaction having been initiated by the customer associated with the plurality of accounts;
determine whether the plurality of accounts of the customer have funds greater than the requested value;
train a machine learning model by:
identifying account selections for customers who have been successful in obtaining particular goals;
segmenting the customers into groups based on a K-means clustering of transaction histories of the customers;
inferring account selections for a cluster set of the customers from the K-means clustering;
and
feeding, to the machine learning model, training data that is representative of a group of the cluster set of the customers who selected a customer account for achieving a goal the same as at least one customer preference or goal of the customer;

select, using the machine learning model, at least one account of the plurality of accounts to provide the requested value of the transaction based on a plurality of inputs to the machine learning model in response to a determination indicating that the plurality of accounts have greater than the requested value, wherein the plurality of inputs includes the at least one customer preference or goal of the customer and information regarding the plurality of accounts;
detect a change to an account of the plurality of accounts in a customer account database;
based on detecting the change to the account, select, using the machine learning model, a different at least one account of the plurality of accounts to provide the requested value of the transaction based on account balances, the at least one customer preference or goal, and cost information for the purchase transaction; and
transfer the requested value from the selected different at least one account of the plurality of accounts to the third-party system.

2. The system of claim 1 , wherein the plurality of inputs further includes historical data of the customer and aggregated historical data of several customers.

3. The system of claim 1 , wherein the selection of the at least one account of the plurality of accounts to provide the requested value includes selecting two or more accounts of the plurality of accounts and the at least one processor is further configured to determine an amount of the requested value to be provided by each of the two or more accounts.

4. The system of claim 1 , wherein the at least one processor is further configured to request additional funds from a partner financial entity in response to a determination indicating that the plurality of accounts have less than the requested value.

5. The system of claim 1 , wherein the at least one customer preference or goal includes one or more of:
information regarding a number of times an account of the plurality of accounts is used in a predetermined time frame;
information regarding a type of transaction permitted with an account of the plurality of accounts;
information regarding an amount of funds permitted to be withdrawn from an account of the plurality of accounts in a predetermined time frame; or
information regarding a priority of each account of the plurality of accounts.

6. The system of claim 1 , wherein the plurality of inputs further includes information regarding prior transactions by the customer.

7. The system of claim 1 , further comprising a user interface, wherein the user interface is configured to permit the customer to assign a weight or priority to each customer preference or goal of the at least one customer preference or goal.

8. A method comprising:
receiving from a third-party system, a request to approve a transaction of a requested value, the transaction having been initiated by a customer having a plurality of accounts at a financial entity;
determining whether the plurality of accounts of the customer have funds greater than the requested value;
training a machine learning model by:
identifying account selections for customers who have been successful in obtaining particular goals;
segmenting the customers into groups based on a K-means clustering of transaction histories of the customers;
inferring account selections for a cluster set of the customers from the K-means clustering;
and
feeding, to the machine learning model, training data that is representative of a group of the cluster set of the customers who selected a customer account for achieving a goal the same as at least one customer preference or goal of the customer;

selecting, using the machine learning model, at least one account of the plurality of accounts to provide the requested value of the transaction based on a plurality of inputs to the machine learning model in response to the determining indicating that the plurality of accounts have greater than the requested value, wherein the plurality of inputs includes the at least one customer preference or goal of the customer and information regarding the plurality of accounts;
detecting a change to an account of the plurality of accounts in a customer account database;
based on detecting the change to the account, selecting, using the machine learning model, a different at least one account of the plurality of accounts to provide the requested value of the transaction based on account balances, the at least one customer preference or goal, and cost information for the transaction; and
transferring the requested value from the selected different at least one account of the plurality of accounts to the third-party system.

9. The method of claim 8 , wherein the plurality of inputs further includes historical data of the customer and aggregated historical data of several customers.

10. The method of claim 8 , wherein the selecting the at least one account of the plurality of accounts to provide the requested value includes selecting two or more accounts of the plurality of accounts and determining an amount of the requested value to be provided by each of the two or more accounts.

11. The method of claim 8 , further comprising one of: 1) providing a loan to the customer, or 2) permitting the customer to have a negative account balance in response to the determining indicating that the plurality of accounts have less than the requested value.

12. The method of claim 11 , wherein the loan to the customer makes up a difference between a total balance of the plurality of accounts and the requested value.

13. The method of claim 8 , wherein the at least one customer preference or goal includes one or more of:
information regarding a number of times an account of the plurality of accounts is used in a predetermined time frame;
information regarding a type of transaction permitted with an account of the plurality of accounts;
information regarding an amount of funds permitted to be withdrawn from an account of the plurality of accounts in a predetermined time frame; or
information regarding a priority of each account of the plurality of accounts.

14. The method of claim 8 , wherein the plurality of inputs further includes information regarding prior transactions by the customer.

15. The method of claim 8 , further comprising notifying the customer of the transfer of the requested

Assignees

• Chime Financial Inc

Inventors

• Ducker Michael
• Lentz Taylor
• Sloboda Tilo

Classifications

• G06Q20/227 Primary

  characterised in that multiple accounts are available, e.g. to the payer · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06Q20/027

  involving a payment switch or gateway · CPC title

• G06Q20/405

  Establishing or using transaction specific rules · CPC title

• G06Q20/401

  Transaction verification · CPC title