How AI can drill into data to prevent late payments

The U.K.’s late payments culture and its resulting impact on small and medium-sized (SME) suppliers has many worrying about whether their clients are the next Carillion, a U.K. construction giant whose collapse left approximately 30,000 SMEs with millions of pounds in unpaid invoices.

This is an endemic problem. A third of the U.K.’s 1.7 million SMEs face delays of at least a month beyond agreed payment terms with large corporate buyers, and nearly a fifth regularly wait more than two months before receiving payment, according to Bacs Payment Schemes.

Invoice on computer
Close-up Of Businessman's Hands Working On Invoice On Laptop At Office
Andrey Popov/Andrey Popov - stock.adobe.com

But soon technology — and specifically artificial intelligence — will be able to help. Tech startup Previse is offering a solution whereby large buyers provide access to their daily electronic payables files as well as historic payments data, enabling Previse to run machine learning algorithms and compute probability scores that the buyer will pay a given invoice within a certain number of days. These analytics are then provided to lenders, typically banks, which pay the supplier immediately upon receiving their invoice.

The crucial aspect of this strategy is that Previse’s analytics are based on large amounts of data, typically three years of invoices. If there’s a high probability that the buyer will both approve the invoice and eventually repay the money, this gives the lender confidence that it is underwriting it at relatively low risk.

“Banks have wanted to do this kind of pre-approval financing in the past, but they haven’t been able to get precise information on the risks entailed,” says Paul Christensen, CEO of Previse. “Because we’ve trained our algorithms on over £100 billion of payments data, we can show the banks exactly what their risks are, so they can price it and get comfortable that they can take that on.”

However, for the supplier, being paid upfront comes at a cost.

In return for being paid instantly, it has to accept a small discount on the agreed payment. This discount is then split between the buyer (who receives a data access fee), Previse (who receives a processing fee) and the lender (who receives compensation for taking the risk). If the buyer goes beyond the agreed payment deadline, it pays interest to the lender.

“As an example of how this works, if the buyer and the small supplier have a 60-day payment cycle, the supplier is given the choice of waiting and chasing for two months, or receiving 98% of the bill today, with the 2% discount being split between the buyer, lender and Previse,” Christensen says.

Market research conducted by Previse on SMEs has shown that 93% are willing to accept this discount in order to guarantee immediate payment, because they need the cash flow and cannot afford the lost time chasing payments. The majority of SMEs spend four hours each week pursuing overdue payments.

Previse’s solution is proving popular as it compares favorably to the existing options for SMEs. These include dynamic discounting whereby a marketplace is created in which the buyer can negotiate to pay a certain number of days earlier, in return for a reduced payment, but this doesn’t give the supplier any certainty.

An alternative is supply chain finance, which allows buyers to extend their payment terms while their suppliers are paid through bank loans during the agreed time frame. However, supply chain finance is only available to the top 1% of suppliers.

“These solutions don’t quite work as they’re either painfully time-consuming for the supplier and involve a lot of paperwork; or in the case of supply chain financing, if the buyer has 10,000 suppliers, it only works for the top 50,” Christensen says. “This is what happened with Carillion. When they went bankrupt, all the big suppliers had already been paid through supply chain financing, but all the little ones hadn’t.”

Previse is currently working with four large corporates that are preparing to offer their new solution to suppliers this year.

However, some SME experts say it is still far from ideal for many small suppliers.

“It may work for some types of firms, but it still means a reduction on income that they should by rights have to pay,” says Robert Blackburn, director of the Small Business Research Centre at Kingston University. “If they have an agreement, then the large firm should pay on time.”

But Christensen points out that even if a supplier receives all its invoices within the agreed payment terms, it is still waiting on average for more than 30 days between sending the invoice and receiving their money. Sometimes it is waiting two or more months.

“That time is not free; it has an opportunity cost and in many cases the SME is covering their cash flow with expensive financing in the meantime,” Christensen says. Choosing to receive payment the day a business issues an invoice "often makes more financial sense than waiting to be paid ‘on time,' " he says.

For reprint and licensing requests for this article, click here.
Payment processing B-to-B payments U.K.
MORE FROM AMERICAN BANKER