	US 11,989,665 B2
	Methods and systems for customizing recommendations based on user actions
Joseph Goldstein, McLean, VA (US); Thomas Dodson, McLean, VA (US); and Eric Delgado, McLean, VA (US)
Assigned to Capital One Services, LLC, McLean, VA (US)
Filed by Capital One Services, LLC, McLean, VA (US)
Filed on Apr. 13, 2021, as Appl. No. 17/228,800.
Application 17/228,800 is a continuation of application No. 16/654,235, filed on Oct. 16, 2019, granted, now 11,004,004.
Prior Publication US 2021/0232952 A1, Jul. 29, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 40/02 (2023.01); G06N 5/04 (2023.01)

CPC G06N 5/04 (2013.01) [G06Q 40/02 (2013.01)]

20 Claims

1. A system for customizing user records that aggregate data from multiple other records based on user actions and underlying record requirements, comprising:

a user input interface configured to:

receive a first user input requesting to create a record for a user, wherein the record aggregates data from multiple user records established with respective third-party entities; and

receive a second user input selecting a record type for the record from a plurality of record types; and

control circuitry configured to:

query the user for user data corresponding to one or more record requirements;

receive the user data, wherein determining that the user meets the one or more record requirements comprises determining that the user data meets the one or more record requirements;

determine a user type of the user based on the user data;

select, based on the user type, a recommendation type from a plurality of recommendation types;

retrieve the one or more record requirements for the record type;

determine that the user meets the one or more record requirements;

create the record in response to determining that the user meets the one or more record requirements as required by the record type;

select, based on the one or more record requirements, an action type from a plurality of action types for the record after creating the record, wherein actions completed by the user of the action type generate recommendations, and wherein action types within the plurality of action types correspond to different types of detectable electronic transactions;

determine a first record state after creating the record, wherein the first record state is measured by a metric;

monitor record activity of the record;

determine that the user completed an action of the action type;

determine a current record state based on the record activity, wherein the current record state is measured by the metric;

select, based on a rate of change between the first record state and the current record state, a recommendation for the record from a plurality of recommendations of the recommendation type, wherein the recommendation type is selected from the plurality of recommendation types that include a reward recommendation type, a penalty recommendation type, and an incentive recommendation type;

generate for display, on a user output interface, the recommendation for the record based on completion of the action of the action type; and

update the record indicate acceptance of the recommendation in response to a third user input accepting the recommendation.