Treasury Today Country Profiles in association with Citi

Now for intelligent accounts receivables

Frustated person in the middle of mountains of paper work

A new solution claims to ease reconciliation pain using AI and machine learning. Are your frustrations over?

Accounts receivables has, for a long time, been one of the most frustrating aspects of a corporate treasurer’s work.

This was largely due to the need to manage swathes of paper-based payment types. Today, the challenge has arguably intensified as corporates must also handle a plethora of digital payment types, all of which offer their own reconciliation challenges. As a result, many treasury teams remain burdened by inefficient, costly and slow reconciliation processes.

Bank of America Merrill Lynch (BofAML) thinks it has solved this issue. Its new Intelligent Receivables solution uses AI, machine learning and optical character recognition to match incoming payments with disparate remittance data. In doing so, it enables payments to be posted in the ERP on a straight-through basis, removing most of the manual work that currently exists in the reconciliation process.

Payment pain

Broadly speaking, the challenges corporates face reconciling incoming payments are no fault of their own. “Businesses are inherently poor at sending remittance information with their payments,” says Rodney Gardner, Head of Global Receivables, Global Transaction Services at BofAML. “In the US, for example, when businesses make ACH payments they often don’t include any remittance information. Instead, this information is provided in a separate email. To add even more complexity, businesses often bundle payments for multiple invoices into one ACH file as well.”

These poor payment practices have created multiple issues for corporates receiving hundreds, if not thousands of payments daily. “There is the cost required to staff a team to manually match incoming ACH payments with the remittance information,” says Gardner. “There is also the negative impact of days sales outstanding (DSO) that the slow posting of payments creates. We must also not overlook the impact it has on the customer’s ability to place further orders if their credit limit has been reached.”

Many banks, including BofAML, have attempted to devise solutions to solve this issue. Yet most, if not all, have fallen short of achieving this, claims Gardner. “There is nothing wrong with the solutions that have been developed,” he comments. “In principle, they can solve the issue. The problem is they require the payer to alter its processes. This is the fundamental issue and why these solutions haven’t had the desired effect.”

Execute, learn, repeat

The fact that payers do not have to change processes for the Intelligent Receivables solution to be effective is what Gardner believes sets it apart from its predecessors. “Payers can continue to send ACH payments and remittance data separately, in any form they like, because the solution is able to read and match these two disparate data sets.”

It does this in four steps. First, it identifies payers, associating their payments to remittances that are received separately. It then scans the remittance email inbox for information that matches the payment and matches payments to open receivables, using the enriched remittance data. “The solution can read free-formatted emails, attachments and even access web portals using its Web Crawler feature,” says Gardner. “Those payments that are matched are posted to the ERP on a straight-through basis; those that can’t are shifted to an exception portal to be manually matched.”

When an individual manually matches the payments and remittance information in the exception portal, the solution utilises machine learning technology to closely follow the steps taken. “Next time a payment is made it will repeat these steps to attempt to post it straight through,” says Gardner. “The aim is to have 80% of transactions flowing on a straight-through basis after a few months.”

Leaving paper behind

There is no debate that when it comes to payments: electronic is better than paper. “It can cost US$8 to US$12 for a corporate to process a cheque payment compared to two cents for an ACH payment,” says Gardner. “Yet because of the reconciliation issues with ACH payments, corporates in the US persevere with cheques.” Solutions such as Intelligent Receivables may be about to change this as more corporates embrace digital payments. BofAML’s offering is currently live in North America and Asia.