Patent Issued for System and method for fraud detection using event driven architecture (USPTO 11315119): United Services Automobile Association

2022 MAY 12 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- A patent by the inventors Brandt, Randall Martin (Dallas, TX, US), Hendry, Jason Paul (Selma, TX, US), Jimenez, Marco Aldo (San Antonio, TX, US), Kalich, Rodney (Shiner, TX, US), Perry, David (San Antonio, TX, US), filed on May 31, 2019, was published online on April 26, 2022, according to news reporting originating from Alexandria, Virginia, by NewsRx correspondents.

Patent number 11315119 is assigned to United Services Automobile Association (San Antonio, Texas, United States).

The following quote was obtained by the news editors from the background information supplied by the inventors: “Banks and other financial institutions rely on information about financial transactions in order to perform a variety of functions, including providing account summary information to customers, performing fraud detection of financial transactions (or facilitating fraud detection performed by a third party), and managing accounts. Presently, these financial institutions must retrieve information about ongoing transactions (including scheduled and pending transactions) from various other financial transaction processing systems, such as credit card processors and automated clearing house (ACH) payment providers. Because the status of various transactions is constantly changing, the financial institution must make many requests to many different systems in order to obtain real time transaction information. This may lead to higher latency in providing current information, such as current account balances. Moreover, data retrieved directly from external systems may comprise raw data that has not been enriched (for example, cleaned and categorized) and may therefore be less reliable. This data may lead to inaccurate account information (such as an inaccurate account balance), and may not be useful for fraud detection.

“Furthermore, because different financial transaction processing systems and providers may use different data formats and record different data milestones, it may be difficult for a bank or other financial institution to obtain a complete historical record for any given financial transaction. This can be a problem, for example, when the financial provider needs to close an account, since there may not be a record of all ongoing transactions that must be canceled before the account can be closed.

“There is a need in the art for a system and method that addresses the shortcomings discussed above.”

In addition to the background information obtained for this patent, NewsRx journalists also obtained the inventors’ summary information for this patent: “In one aspect, a method of reducing latency in providing an account balance for a financial account includes steps of listening for financial transaction events on an event stream, and detecting a transaction event for a financial transaction, where the transaction event is published by a financial transaction processing system. The method also includes steps of storing transaction data associated with the transaction event in a datastore, receiving an account balance request, retrieving the transaction data from the datastore and calculating the account balance using the transaction data, and providing the account balance.

“In another aspect, a method of sending account alerts to a customer for a financial account includes steps of listening for financial transaction events on an event stream, and detecting a transaction event for a financial transaction, where the transaction event is published by a financial transaction processing system and the transaction event is associated with transaction data. The method also includes steps of sending a request to enrich the transaction data, receiving enriched transaction data and publishing an enriched data event, and sending an account notification to a device associated with the customer, where the account notification includes information from the enriched transaction data.

“In another aspect, a system for providing account information for a financial account includes an event stream and a first subsystem. The first subsystem is configured to receive transaction events in the event stream from a plurality of financial transaction processing systems, communicate with a data enrichment service to create enriched transaction data, and publish enriched transaction events to the event stream. The system also includes a second subsystem system configured to receive enriched transaction events from the event stream and persist the enriched transaction data in long term storage. The system also includes a third subsystem system configured to receive the enriched transaction data and use the enriched transaction data to provide account information to an end user.

“In another aspect, a method of detecting fraudulent activity for a financial account includes listening for financial transaction events on an event stream and detecting a transaction event for a financial transaction. The transaction event is published by a financial transaction processing system and the transaction event is associated with transaction data. The method also includes sending a request to enrich the transaction data, receiving enriched transaction data and publishing an enriched data event, analyzing the enriched transaction data, and detecting fraudulent activity associated with the financial transaction.

“In another aspect, a method of sending fraud alerts to a customer for a financial account includes steps of listening for financial transaction events on an event stream and detecting a transaction event for a financial transaction. The transaction event is published by a financial transaction processing system and the transaction event is associated with transaction data. The method also includes steps of sending a request to enrich the transaction data, receiving enriched transaction data and publishing an enriched data event, analyzing the enriched transaction data for fraudulent activity, and sending a fraud notification to a device associated with the customer. The fraud notification includes information from the enriched transaction data.”

The claims supplied by the inventors are:

“1. A method of detecting fraudulent activity for a financial account, comprising: listening for transaction events published to an event stream; detecting the transaction events for a financial transaction, the transaction events being published by a financial transaction processing system and stored in a log for subsequent use by an event consumer, and the transaction events being associated with transaction data, wherein the transaction events are received from the financial transaction processing system and from an event producer; sending a request to enrich the transaction data, the transaction data originally including raw data, the enriching occurring in real time after detection of the transaction data and being performed by a data enrichment application programming interface (API); receiving enriched transaction data and publishing an enriched data event by the event producer for use by the event consumer, the enriched transaction data including enriched pending transaction data and enriched scheduled transaction data, the enriched transaction data being converted from a raw format into an enriched format by being cleaned by removing simple errors and applying a consistent format and being categorized, wherein the financial transaction processing system includes a datastore configured to temporarily store transaction data and a long term storage configured to persistently store transaction data, and wherein the enriched pending transaction data is stored in the long term storage and wherein the enriched scheduled transaction data is stored in the datastore; reconciling scheduled and pending transactions using a reconciliation API to find completed and expired transactions for use when detecting fraud; aggregating, by a data aggregator, enriched manual and scheduled transaction data from the financial transaction processing system and enriched pending and processed transaction data for use when detecting fraud; analyzing the enriched transaction data; and detecting fraudulent activity associated with the financial transaction using fraud detection algorithms that provide improved fraud detection based on the enriching, reconciling, and aggregating the transaction data, wherein the fraud detection algorithms detect fraud using a trained model, and the trained model is a neural network, wherein the enriching reduces false positives by cleaning the transaction data and aids in identifying differences between groups of fraudulent and valid data by categorizing the transaction data, and wherein the reconciling and the aggregating facilitate access to the transaction data by improving the organization and reliability of the transaction data.

“2. The method according to claim 1, wherein the fraud detection algorithms include clustering and classification techniques.

“3. The method according to claim 1, wherein the fraud detection algorithms include time series analyses.

“4. The method according to claim 1, wherein the method includes a step of posting the financial transaction, and wherein the step of posting the financial transaction occurs after the step of detecting fraudulent activity.

“5. The method according to claim 1, wherein the method includes a step of publishing a fraudulent activity event to the event stream after the fraudulent activity has been detected.

“6. The method according to claim 5, wherein the method includes a step of notifying an owner of the financial account when the fraudulent activity event is published.

“7. A system for detecting fraud for a financial account, comprising: a financial transaction processing system configured to: receive transaction events published to an event stream from a plurality of financial transaction processing systems and stored in a log for subsequent use by an event consumer, the transaction events being associated with transaction data, the transaction data originally being in a raw data format; communicate with a data enrichment service to create enriched transaction data, the enriching occurring in real time after reception of the transaction data and being performed by a data enrichment application programming interface (API); and publish enriched transaction events to the event stream by an event producer for use by the event consumer; a management and storage system configured to: receive enriched transaction events from the event stream, the enriched transaction data including enriched pending transaction data and enriched scheduled transaction data, the enriched transaction data being converted from the raw format into an enriched format by being cleaned by removing simple errors and applying a consistent format and being categorized; persist the enriched transaction data in a long term storage of the management and storage system, wherein the financial transaction processing system includes a datastore configured to temporarily store transaction data and includes a long term storage configured to persistently store transaction data, and wherein the enriched pending transaction data is stored in the long term storage and wherein the enriched scheduled transaction data is stored in the datastore; reconcile scheduled, manual, and pending transactions using a reconciliation API to find completed and expired transactions received for use when detecting fraud; aggregate, by a data aggregator, enriched manual and scheduled transaction data from the financial transaction processing systems and enriched pending and processed transaction data for use when detecting fraud; a fraud detection system configured to: receive the enriched transaction data; and analyze the enriched transaction data to detect fraudulent activity using fraud detection algorithms that provide improved fraud detection based on the enriching, reconciling, and aggregating the transaction data, wherein the fraud detection algorithms detect fraud using a trained model, and the trained model is a neural network, wherein the enriching reduces false positives by cleaning the transaction data and aids in identifying differences between groups of fraudulent and valid data by categorizing the transaction data, and wherein the reconciling and the aggregating facilitate access to the transaction data by improving the organization and reliability of the transaction data.

“8. The system according to claim 7, wherein the fraud detection algorithms include clustering and classification techniques.

“9. The system according to claim 7, wherein the fraud detection algorithms include time series analyses.

“10. The system according to claim 7, wherein the fraud detection algorithms include anomaly detection techniques.

“11. The system according to claim 7, wherein the fraudulent activity is associated with a pending transaction.

“12. The system according to claim 7, wherein the fraudulent activity is associated with a scheduled transaction.

“13. The method according to claim 1, wherein the fraud detection algorithms include anomaly detection techniques.

“14. The system according to claim 7, wherein the management and storage system is configured to publish processed complete transaction events and processed expired transaction events.

“15. The system according to claim 14, wherein the financial transaction processing system listens to the event stream for processed complete transaction events and processed expired transaction events.

“16. The system according to claim 7, wherein each financial transaction is associated with a transaction record in the system.”

URL and more information on this patent, see: Brandt, Randall Martin. System and method for fraud detection using event driven architecture. U.S. Patent Number 11315119, filed May 31, 2019, and published online on April 26, 2022. Patent URL: http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.htm&r=1&f=G&l=50&s1=11315119.PN.&OS=PN/11315119RS=PN/11315119

(Our reports deliver fact-based news of research and discoveries from around the world.)