Determining user types from behavior

Patent metadata

Field	Value
Publication number	US-12488348-B1
Application number	US-202418972104-A
Country	US
Kind code	B1
Filing date	Dec 6, 2024
Priority date	Dec 6, 2024
Publication date	Dec 2, 2025
Grant date	Dec 2, 2025

Abstract

Determining user types from behavior is described. An artificial intelligence (AI) model is trained to classify user accounts of a payment service into different user types using contextual data associated with processed payments between the user accounts. The AI model is used to analyze additional contextual data associated with additional payments between additional user accounts to classify the additional user accounts, and, if a particular user account of the additional user accounts is associated with a user type of the different user types that requires an action to be performed, an instruction is sent to a user device associated with the particular user account to cause a payment application to present a user interface element prompting a user to perform the action, and, based on whether the action was performed, account data indicating whether the particular user account is an authorized account is stored in a datastore.

First claim

What is claimed is:

1 . A computer-implemented method comprising:
processing, by a computing system associated with a payment service, payments between user accounts of the payment service, wherein the payments are associated with contextual data, and wherein the user accounts are associated with different user types; filtering, by the computing system, the user accounts to obtain a filtered set of user accounts by excluding a subset of the user accounts that are associated with:
user profile data indicating that date of birth information has been changed one or more times; or
identity verification (IDV) data indicating inconsistent IDV attempts;
training, by the computing system and using the contextual data associated with the payments between the filtered set of user accounts, a first machine learning model to classify the filtered set of user accounts into the different user types, wherein the first machine learning model uses at least one of notes associated with the payments, network interactions associated with the payments, or utilization patterns associated with the payments to determine which user type of the different user types is associated with a user account; analyzing, by the computing system and using the first machine learning model, additional contextual data associated with additional payments between additional user accounts of the payment service to classify the additional user accounts; determining, by the computing system and based on the analyzing, that a particular user account of the additional user accounts is associated with a user type of the different user types that requires an action to be performed; analyzing, by the computing system and using a second machine learning model, data associated with the particular user account to identify one or more user accounts of the payment service; sending, by the computing system, an instruction to a user device associated with the particular user account and executing a payment application associated with the payment service, the instruction causing the payment application to present a user interface element prompting a user of the user device to request sponsorship from at least one of the one or more user accounts as a sponsor of the user; storing, by the computing system, in a datastore, account data indicating that the particular user account is an authorized account based on an indication of an interaction with the user interface element and the particular user account having been linked to the sponsor subsequent to the interaction with the user interface element; and disabling, by the computing system, a functionality associated with the payment service for the particular user account based at least in part on the particular user account having been linked to the sponsor.

2 . The computer-implemented method of claim 1 , wherein the different user types represent at least one of:
different levels of risk; different levels of authentication; or different levels of access to the payment service.

3 . The computer-implemented method of claim 1 , wherein the first machine learning model further uses at least one of emojis associated with the payments or payment failure metrics associated with the payments to determine which user type of the different user types is associated with the user account.

4 . The computer-implemented method of claim 1 , further comprising analyzing, by the computing system and using the first machine learning model, contact book data associated with the additional user accounts to classify the additional user accounts, the contact book data comprising at least one of:
percentages of contacts that have been classified as a particular user type; names of the contacts; electronic mail (email) address types associated with the contacts; or profile pictures associated with the contacts.

5 . The computer-implemented method of claim 1 , further comprising, in response to the determining that the particular user account is associated with the user type, placing, by the computing system, the particular user account in a restricted state until the action is performed.

6 . The computer-implemented method of claim 1 , wherein the additional contextual data analyzed using the first machine learning model comprises location data indicating locations of user devices associated with the additional user accounts at times at which the additional payments were made.

7 . A system comprising:
one or more processors; and memory storing computer-executable instructions that, when executed by the one or more processors, cause the one or more processors to perform operations comprising:
processing payments between user accounts of a payment service;
filtering the user accounts to obtain a filtered set of user accounts by excluding a subset of the user accounts that are associated with:
user profile data indicating that date of birth information has been changed one or more times; or
identity verification (IDV) data indicating inconsistent IDV attempts;

training a first artificial intelligence (AI) model to classify the filtered set of user accounts into different user types using contextual data associated with the payments between the filtered set of user accounts;
analyzing, using the first AI model, additional contextual data associated with additional payments between additional user accounts of the payment service to classify the additional user accounts;
determining, based at least in part on the analyzing, that a particular user account of the additional user accounts is associated with a user type of the different user types that requires an action to be performed;
analyzing, using a second AI model, data associated with the particular user account to identify one or more user accounts of the payment service;
sending an instruction to a user device associated with the particular user account and executing a payment application associated with the payment service, the instruction causing the payment application to present a user interface element prompting a user of the user device to request sponsorship from at least one of the one or more user accounts as a sponsor of the user;
storing, in a datastore, account data indicating that the particular user account is an authorized account based at least in part on an indication of an interaction with the user interface element and the particular user account having been linked to the sponsor subsequent to the interaction with the user interface element; and
disabling a functionality associated with the payment service for the particular user account based at least in part on the particular user account having been linked to the sponsor.

8 . The system of claim 7 , wherein the different user types represent at least one of:
different levels of risk; different levels of authentication; or different levels of access to the payment service.

9 . The system of claim 7 , wherein, prior to the analyzing of the additional contextual data, at least some of the additional user accounts are selected for classification into the different user types based at least in part on determining that a predefined period of time has lapsed since the first AI model was last used to classify the at least some of the additional user accounts into the different user types.

10 . The system of claim 7 , wherein the additional contextual data analyzed using the first AI model comprises location data indicating locations of user devices associated with the additional user accounts at times at which the additional payments were made.

11 . The system of claim 7 , wherein the determining that the particular user account is associated with the

Assignees

• Block Inc

Inventors

• Joshi aditya
• Davis Alexa

Classifications

• G06Q20/4015

  using location information · CPC title

• G06Q20/3224

  Transactions dependent on location of M-devices · CPC title

• G06Q30/0609

  Qualifying participants for shopping transactions (payment transaction verification G06Q20/401) · CPC title

• G06Q20/12

  specially adapted for electronic shopping systems · CPC title

• G06Q20/227

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