US 12,169,816 B2
Automated transaction processing based on cognitive learning
Matthew E. Carroll, Charlotte, NC (US)
Assigned to Bank of America Corporation, Charlotte, NC (US)
Filed by Bank of America Corporation, Charlotte, NC (US)
Filed on Aug. 31, 2022, as Appl. No. 17/900,416.
Application 17/900,416 is a continuation of application No. 16/821,117, filed on Mar. 17, 2020, granted, now 11,468,415.
Prior Publication US 2022/0414624 A1, Dec. 29, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 40/00 (2023.01); G06F 9/50 (2006.01); G06N 20/00 (2019.01); G06Q 20/10 (2012.01)
CPC G06Q 20/10 (2013.01) [G06F 9/5083 (2013.01); G06N 20/00 (2019.01)] 21 Claims
OG exemplary drawing
 
1. A method for transferring funds from a payer to a payee based on a transmission of an electronic message from the payer to a financial institution associated with the payer, the method comprising:
receiving, via a communication interface, a first message indicating a first request to process a fund transfer via a first transfer channel of a plurality of available transfer channels, the first message indicating a first request to process a funds transfer via a first transfer channel computing device in electronic communication with a transaction control computing platform, wherein the first transfer channel computing device is configured to transfer the funds based on the first request to a destination account of the payee maintained at a payee financial institution having a destination computing device in electronic communication with the first transfer channel computing device,
wherein the first request comprises transaction details associated with the funds transfer including a transaction type,
wherein the transaction type is associated with the first transfer channel and indicates that the first transfer channel of the plurality of available transfer channels is to be used for the funds transfer,
wherein the first transfer channel comprises a system queue for storing one or more fund transfer requests which are processed in an order specified by the system queue,
wherein the transaction control computing platform comprises a transaction analytics computing unit for monitoring funds transfer attributes associated with the plurality of available transfer channels and for storing determined attributes in a log file and periodically updating the log file based on the determined attributes and the monitoring, and
wherein the funds transfer attributes are associated with at least one of a cost, timing and efficiency of transfer channels;
determining from the first request that the first transfer channel of the plurality of available transfer channels is to be used for processing the first request;
determining respective attributes of the other available transfer channels;
monitoring a real-time status of the attributes of the first transfer channel to detect when the first transfer channel does not meet at least one criteria based on the first message;
selecting, in response to the monitoring and based on a real-time status of the attributes of the other available transfer channels and the first message, a second transfer channel of the other available transfer channels;
switching from the first transfer channel to the second transfer channel to ensure a successful transfer of funds; and
transmitting, via the communication interface and via the second transfer channel, a second message indicating a second request to process the funds transfer, wherein the transaction control computing platform selects the second transfer channel based on a machine learning algorithm to determine one or more payment systems and intermediary accounts for the funds transfer and to determine at least one of wait times, expected usage costs, operation status, country based restrictions, and customer satisfaction scores associated with the second transfer channel,
wherein the machine learning algorithm of the transaction control computing platform comprises an artificial neural network configured to execute the machine learning network to determine a real-time status of the attributes of each of the available transfer channels,
wherein an artificial neural network obtains attribute information about the available transfer channels through a plurality of input nodes comprising logical inputs from different data sources, wherein a plurality of first output signals from the input nodes are processed by a plurality of processing nodes, wherein the processing nodes comprise a plurality of parallel processes executing on multiple computing devices, wherein each of the processing nodes may be connected to one or more other processing nodes, wherein the connections connect an output of one node to an input of another node, wherein the connection is correlated with a weighting value in which the output may be weighted as more important or significant than an output of another processing node, thereby influencing a degree of further processing as an input traverses across the artificial neural network, wherein the connections are modified by data feedback of correct or incorrect decisions such that the artificial neural network learns and is dynamically reconfigured as more inputs are processed, and
wherein one or more second output signals from the plurality of processing nodes are processed by a plurality of output nodes to obtain attribute status information about the available transfer channels, wherein the output comprises at least one determined transfer channel(s), expected wait times, threshold queue lengths, threshold times, predicted expected handshake wait times, threshold customer satisfaction scores, confidence values, operation status of transfer channels, predicted expected usage costs, and classification output.