We just released a member Viewpoint that summarizes the recently completed AFP annual event, which like all other industry trade gatherings in 2020 was delivered via a virtual event experience. In that event overview, we discuss some of the key trends in the payments track sessions, of which there were a couple of dozen.
Interestingly enough about 20% of the payments track sessions had something to do with receivables, which has traditionally been the somewhat neglected segment of cash cycle operations.This started changing in the past couple of years and now receivables technology is getting as much if not more focused attention as other parts of financial operations.
This referenced piece is found in Cision PR Newswire and was provided through the receivables automation fintech Billtrust, which is in process of going public via a stock deal with South Mountain Merger Corp. It talks to one of the key developments in the AR space, which is the increasing use of machine learning (generally included under the AI umbrella technologies).
‘Billtrust, the B2B accounts receivable automation and integrated payments leader, has recently upgraded its Cash Application software’s advanced machine learning capability, significantly improving match rates and reducing manual processing while converting payments to cash as fast as possible….Billtrust’s Cash Application quickly adapts to a supplier’s ERP system without being explicitly programmed, delivering a tailored experience based on how accounts receivable teams work with their systems and data. Modeling from remittances and data, match rates improve over time as the model learns usage while adapting to any invoice structure changes. Higher match rates allow users to get through their worklist efficiently with fewer exceptions meaning faster access to cash.’
In AR processes, companies want to optimize match rates between remittance data/payments and the associated invoices. The faster this can be done, the more quickly the cash can be applied to the correct accounts in the GL. Like everything else during the past 9 months, the DSO improvement here can be critical to working capital effectiveness.
This matching process has been further complicated (ironically enough) by the increase in various forms of e-payments, which often arrive disassociated from remittance data. Machine learning (ML) is being used to improve automated matching rates over times as patterns emerge and strengthen the algorithms. This is an area that Billtrust has been adding capabilities.
‘Since the July 2020 upgrade, Billtrust customers have seen a 12.4% increase in overall match rates and an 18% increase for electronic payments. One Billtrust customer, heavy equipment dealer Gregory Poole Equipment Company, transitioned to upgraded machine learning in July 2020 and has reported strong match rate increases and an increase in auto-matched envelopes. “We’re a complex organization, so improving our match rates was a challenge,” said Mary Stumpf, Accounts Receivable Supervisor. “Billtrust more than met the challenge, and the transition to a new machine learning-powered platform was seamless. We continue to see excellent results with strong match rate performance, which is more important than ever with a remote workforce. It’s really incredible how machine learning actually adapts to our systems for continuous improvement.“‘
Overview by Steve Murphy, Director, Commercial and Enterprise Payments Advisory Service at Mercator Advisory Group
