Duplicate invoice payments often go undetected by accounts payable. They are most of the time due to bad data entries leading to significant losses if not detected. It all belongs to the past as our algorithm is designed to detect these duplicates. The detection algorithm goes through multiple stages. Firstly preprocessing steps are applied to clean all of the invoice data. Here you can think of replacing an"i" for a "1" within an invoice number to overcome potential scanning failures.
Then fuzzy string matching is applied over all invoices (records) using the available data attributes in the dataset. Potential invoice duplicates will be indentified based on a similarity score presented as percentage where 100% indicates a perfect match.
Our duplicate payment algorithm provides the following benefits to your company:
For this analysis we make use of invoice data only. Using our algorithm some of the value adding attributes for detecting duplicate invoices are vendor name, invoicenumber, amount, currency, invoice date etc.
The resulting matches (incl. similarity scores) will be made available in:
For a fully automised solution preventing duplicate payments get in contact to get more information on our webservice.
In most cases duplicate invoices do not match exactly. Therefore we introduced fuzzy matching to our model in order to identify potentially duplicates based on a similarity score.
Invoice data is cleaned in order to find duplicates effectively.
With the help of the duplicate invoices algorithm you can select the invoices that are completely or largely similar and that have been or will be paid twice. The algorithm is using Fuzzy Matching for this. In this Duplicate Payment demo you can select an invoice identification number and see what the percentage matching is.
Note: all data-attributes and visuals in this webpage are based on fictitious data
Get in touch to receive more information on how we can bring your data to actionable insights
