As a finance executive you want to keep on top of opportunities and threats within your finance department. As your industry and vendors get smarter and improve on a technical aspect, the process even gets more complicated. Therefore, we wanted to become the thought leader in accounts payable (AP), to help our clients in the best possible way. This gives you the opportunity to focus on other subjects and we provide the best in class AP controls and technology.
We share our story starting from manually analysing and recovering invoices within the Accounts Payable department, to developing our own APRA (Accounts Payable Recovery Analyser) software. Where 10 years ago an AP department yearly handled already millions of invoices, that number has even grown multiple times now. You can imagine that with those numbers it’s impossible for people to keep track of duplicates, double payments and unauthorised invoices. And even. now, the best ERP systems can’t catch all the mistakes. But APRA can. Let’s look at how we did that.
Our way of technological development
Within SpendLab both technological- and the manual auditing play an important role. As the complexity of the accounts payable (AP) makes it extremely difficult for only technology to find all flaws and missing liquidity. Our technology forms a synergy with our manually working employees, having a crucial role in the process. Not only to find those missing flaws, but also to support technology. The development of our technology is based on our continuous loop of improvement and validation, and needs the input on often occurring incorrect processed invoices, noticed by our auditing team, to get new points of focus for additional algorithms to increase the flaws found by our technology. As part of the continuous loop, those new algorithms also have to be validated by the auditing team before it is used in practice.
How does this development help you?
As we mentioned, we started out as any other AP Recovery team. A number of years ago we found that with continuously improving technology, we could let software handle the easy to recognise duplicates, while our team took on the harder ones and find new ways of discovering duplicate payments together with APRA. Due to continuous improvement, specifically focused on tracking those payments in the AP department, we managed to create 450 algorithms in our APRA system that recognise almost any type of double payment. Now our people can focus even more on what they excel in, while APRA can even handle robotic process automation (RPA) and other intelligent technologies. The increasing number of found incorrect payments is due to the fact that we don’t only retrieve AP-data, but also extract data out of all ERP-systems regarding the payment process of the invoices.
Advantages of NLP
After retrieving the data, our technology team uses of different methods to analyze the information. One of them is with the use of Natural Language Processing (NLP). NLP creates a vocabulary which recognizes all the words out of the different documents. These words need to be compared to one another and are given a score how closely related they are to one another. Afterwards all the words are clustered and all the different documents are compared based on their score. In this way we can really look into the content of an invoice and not just it’s features.
We are honored to announce that our improvement recently was recognized and is awarded by receiving a spot in the final of the Dutch BI and Data Science Award on the 9th of December.
Do you want to know more about what our technology could mean for your organization? We are happy to answer all questions you have, so feel free to fill in our contact form or contact us directly!