US 11,810,120 B2
Methods and systems for a virtual assistant
Bradley O. Matthews, Minneapolis, MN (US); and Tory K. Passons, Minneapolis, MN (US)
Assigned to U.S. Bancorp, National Association, Minneapolis, MN (US)
Filed by U.S. Bancorp, National Association, Minneapolis, MN (US)
Filed on Sep. 3, 2021, as Appl. No. 17/466,764.
Application 17/466,764 is a continuation of application No. 16/369,697, filed on Mar. 29, 2019, granted, now 11,113,696.
Prior Publication US 2021/0398122 A1, Dec. 23, 2021
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 20/40 (2012.01); G06Q 10/10 (2023.01); G06Q 20/32 (2012.01); G06Q 10/02 (2012.01)
CPC G06Q 20/405 (2013.01) [G06Q 10/02 (2013.01); G06Q 20/3223 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method comprising:
receiving, by a virtual assistant computing device, an indication that a transaction by a mobile device was completed;
determining, by the virtual assistant computing device, a transaction purpose of the transaction by:
retrieving, from a database in a central server, entries that are within a time frame before or after a time of the transaction or emails that contain words that match words of the transaction;
identifying one or more parameters of the entries or emails;
associating metadata with the identified one or more parameters of the entries or emails;
determining a first potential transaction purpose by feeding the parameters of the entries or emails, with the associated metadata, into a machine learning model;
transmitting the first potential transaction purpose to the mobile device;
receiving, from the mobile device, a first indication that the first potential transaction purpose is incorrect;
responsive to receiving the first indication that the first potential transaction purpose is incorrect, generating a training data set comprising the entries and the first indication that the first potential transaction purpose is incorrect;
training the machine learning model using the training data set;
after training the machine learning model, determining a second potential transaction purpose by using the trained machine learning model;
transmitting the second potential transaction purpose to the mobile device; and
receiving, from the mobile device, a second indication that the second potential transaction purpose is correct;
determining, by the virtual assistant computing device and based on the second potential transaction purpose, whether the transaction is in accordance with policy;
responsive to determining the transaction is in accordance with policy, generating, by the virtual assistant computing device, an expense report associated with the transaction; and
submitting, by the virtual assistant computing device, the generated expense report to an expense report system.