The way we all work has undergone a significant shift. We’ve all had to adapt, and criminals have exploited these changing circumstances. As we mentioned in a previous article, COVID-19: The potential for fraud, disruption caused by the pandemic has created an opportunity to commit fraud and leaves companies with a tricky problem: how to spot it when there’s no more ‘business as usual’?
Fraud tends to happen when there is opportunity, incentive, and the ability to rationalise the act. The pandemic has created all three. Our latest Global Economic Crime Survey showed the highest rate of fraud in businesses that we’ve ever recorded. 56% of respondents said that they’d experienced fraud in the previous two years, and we believe that the rate has accelerated during the pandemic – reflecting impacts seen in the general public.
An increase in procurement fraud is one of our biggest concerns; this currently accounts for about a fifth of all reported fraud incidents. The movement to mass home-working, the disruption to supply chains (such as the need to find alternative goods and suppliers quickly), and economic hardship creates more fertile ground for fraud. Fraudsters know that people are working from home and that controls may be weaker than normal, with businesses distracted by more novel or urgent pressures.
Procurement fraud is one of the biggest hidden costs for companies, often manifesting only as a slow drain from the bottom line that can take a painfully long time to spot. We’ve seen cases where millions of pounds have been misappropriated, and no one noticed for years.
Technology is increasingly being leveraged to address the problem, with automated solutions designed to spot unusual activity. However, the pandemic has created new challenges. Transactions that would have been a warning sign under normal circumstances – such as an invoice logged out of hours or from an unusual location, or for goods that an organisation wouldn’t normally order – have become routine. The baseline normal against which transactions can be compared has changed, resulting in a need to recalibrate.
This is where machine learning comes in. Our Procurement Protect solution, for example, uses unsupervised machine learning techniques to spot subtle patterns in procurement data that deviate from new norms. These tools allow for a more proactive approach as monitoring can be continuous and adapt to changing circumstances, and are a key component of an effective fraud risk management framework. Fraud is a sophisticated business and fraudsters quickly learn the techniques that are employed to spot them; machine learning can help us stay one step ahead.
