Plataforma de banca moderna FIS
Haga avanzar su banco con una plataforma central moderna.
Leave Manual Cash Application Processes Behind and Achieve 90 Percent Hit Rates
March 29, 2019
It wasn’t so long ago that companies were searching for solutions on how to improve cash collection. Processes were inefficient, and companies were flushing millions of dollars down the drain by not effectively managing their collection teams. The more modern companies sought forward-thinking solution providers to help them with bringing practices under control. The biggest and most attractive benefit has been an improvement in cash flow.
During this renaissance of collections, many companies overlooked the impact of their cash application processes. It was thought to be a manual process that could be solved by migrating customers from checks to electronic payments. The result was quite the opposite. Electronic payments have actually increased the amount of manual work for cash appliers. This increased manual workload has shifted companies’ focus to solving their cash application challenges.
Analysts have been predicting the death of the check in the U.S. for a better part of two decades. At first ACH was supposed to kill it, then credit cards, and so on. At this point, the check is still a major payment method. Across Europe, electronic payment methods have been the norm for a very long time. It does look like the emergence of “real-time payments” is going to have a major impact on checks and some electronic methods, but the verdict is still out on whether it will kill them off.
What does this mean for the cash application process? At one point, checks were pretty much the only payment method between businesses, so companies began shifting the manual work of keying remittance information to their bank (lockbox). The bank would charge a nominal fee per keystroke and then send some form of an electronic file containing all of the keyed remittance information. The sophisticated companies built a process to automatically apply this electronic file against open invoices in their accounting system. It helped some but was also riddled with keying errors and no validation process as banks did not have access to see the open invoice information.
Additionally, ERP/Accounting Systems are inflexible with handling the matching process. If the exact invoice number was not provided on every payment, the ERP/Accounting System could not determine how to apply the payment, making it a manual process again. Flexible rules are required to look for different information, such as PO number, varying invoice combinations, tolerance thresholds, and so on, to make sure that payments can be applied automatically and accurately.
ACH/EFT was a new payment vehicle that was introduced to provide a cheaper and quicker payment option for companies. Again, European countries led the way with adopting this payment method very early on. At first, Accounts Receivable departments were ecstatic about an electronic payment that they could receive in two days rather than waiting for a check to arrive in the mail. However, this excitement was short-lived. It turns out that the electronic payment method organizations were pushing customers to sign up for was creating more manual work. While the payments arrived quicker, the remittance information of how to apply the payments was sent through separate channels or sometimes not at all. This situation also occurred with other electronic payment methods, such as credit cards.
Companies were then faced with how to re-associate these payments and remittance advices. This led to automated processes that involved Optical Character Recognition (OCR). OCR is a process by which the system will automatically retrieve the remittance advice, scan and read the information, and digitize it in a format that can be used for matching. The idea is great; however, OCR alone requires the system to have templates setup so that the system knows how to read and locate the information. Creating these templates is a time-consuming process that has to continually be updated with each new remittance layout received.
Companies that leveraged lockbox keying for checks and OCR plus matching rules for electronic payments, average only about 70-80 percent first-pass hit rates (first-pass hit rates refers to payments being correction applied to open invoices automatically, without any manual intervention).
The good news is, there is a better way to simplify the cash application process. The introduction of artificial intelligence (AI) can work in partnership with other matching methods, like the ones mentioned above. A cash application process that uses Intelligent Document and Data Recognition (IDDR) is able to read and recognize document formats and data without a template being necessary. As the incoming data is recognized the AI engine knows how best to apply the payment against open invoices. The one drawback to an AI-based machine learning system is the amount of time it takes to become accurate. The AI engine requires multiple examples of something to become proficient at executing it on its own. This was, however, until the introduction of Accelerate Machine Learning.
With Accelerated Machine Learning, the AI engine monitors user action to learn how a remittance was applied. Once the user completes the posting action they simply confirm for the AI engine that this method should be used going forward for that customer. The next incoming payment will then be applied automatically. This Accelerated Machine Learning allows companies to achieve 90 percent or greater first-time hit rates right away. No longer do they have to wait for a ramp-up period to achieve the automation they so desperately need.
Companies now have a remedy for their cash application challenges. One of the additional benefits of implementing a cash application solution that uses AI with Accelerate Machine Learning is the fact that it can handle any new payment methods that may come down the road. By modernizing cash application process with the latest automated cash application solutions, manual work has become a thing of the past.