US patent for Esker ML

Global cloud platform and leader in AI driven process automation solutions provider Esker (www.esker.com) announces the issuance of U.S. Patent No. 11631265 by the U.S. Patent and Trademark Office (USPTO) for Esker’s machine learning technology.

Spanning both the source-to-pay (S2P) and order-to-cash (O2C) cycles, Esker’s AI-powered solutions are built to vastly reduce repetitive and low-value tasks.  The patented technology addresses the need for accurate automated data extraction on incoming documents.  

Esker’s newly patented AI technology uses traditional and machine learning algorithms to identify data fields, make routing decisions and suggest next actions with the goal of reducing the number of touches during processing as much as possible.  The data is read from a document and the AI creates rules that define what data to retrieve and in which format.  Through user input, these rules are adapted, and the accuracy of the extraction process increases over time.  This interplay of AI technology and human supervision reduces mindless and error-prone manual data entry down to almost zero.

“When we started on the AI journey at Esker, we focused our efforts on data extraction in order to decrease manual data entry for incoming invoices and orders,” says Jean-Jacques Bérard, Vice President of Research and Development, Esker.  “This project has expanded to enriching the extracted data with predictive and prescriptive functionalities such as detecting anomalies in orders, predicting the invoice analytical axis, or proposing an answer to a customer request.  This combination of human and artificial intelligence allows for making the work more engaging and efficient.”