CPC G06Q 20/227 (2013.01) [G06Q 20/102 (2013.01)] | 16 Claims |
1. A computer-implemented method comprising:
training, by the one or more processors during a training phase, a machine-learning model to output scores corresponding to recommendations regarding payment of an invoice, the machine-learning model trained using a training set including a plurality of data entries for invoices from the first data source and account data for a plurality of accounts from the second data source;
deploying, by the one or more processors, the machine-learning model responsive to outputs from the machine-learning model satisfying one or more criteria during the training phase;
receiving, by an application programming interface (API) gateway of a computer system corresponding to the second data source, a first API call for a first API managed by the API gateway, the first API call including information corresponding to authentication of an enterprise application executing on a second server, the enterprise application being authenticated for accessing information via the API gateway based on the information corresponding to the authentication, the enterprise application linked to a common account holder of an account of the second data source, the enterprise application corresponding to the first data source;
establishing, responsive to receiving the first API call and authentication of the information, a connection between the computer system and the enterprise application;
detecting, by the one or more processors, a data entry for an invoice between the common account holder and a third party, the data entry detected via a webhook based on an indication of the data entry for the invoice at the first data source
extracting, by the one or more processors, a trade term from the invoice responsive to detecting the data entry;
applying, by the one or more processors, the trade term from the invoice to the machine-learning model as a first input;
applying, by the one or more processors, data of the account of the second data source to the machine-learning model as a second input;
receiving, by the one or more processors from the machine-learning model, an output comprising a score for early-payment of the data entry according to the trade term and the data of the account; and
transmitting, by the one or more processors, in response to determining that the score exceeds a predetermined threshold, a notification for the invoice to a computing device associated with the common account holder.
|