How is AI & machine learning being used in fraud detection?

Procurement Magazine looks at how artificial intelligence (AI) and machine learning (ML) are being used in fraud detection

In the last two years, 56% of organisations have experienced fraud (the greatest number ever recorded, according to PwC’s Global Economic Crime and Fraud survey). In particular, procurement fraud is of high concern, accounting for a fifth of all reported incidents. 13% of those that experienced fraud reported a loss of US$50mn or more.

The most common global types include customer fraud, cybercrime, and asset misappropriation, of those crimes 40% came from internal and 40% from external perpetrators, the rest came from collusion between the two. 

Despite these alarming figures, many organisations fall short when it comes to responding effectively, only 56% conduct an investigation into their worst incident, with only one-third reporting it to their board. 

Specifically when it comes to procurement, the risk of fraudulent invoices slipping through the cracks is greater than ever,” said Vroozi. “The three most common types to watch for are fake vendors, fake invoices and inaccurate or over-billed invoices. Technology advancements have made spotting invoice and accounts payable fraud much easier today than it was years ago – but organisations still need to invest and act.”

What is artificial intelligence?

Leveraging a combination of computers and machines to replicate problem-solving and decision-making capabilities of humans, modern artificial intelligence (AI) is defined as “the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable," said John McCarthy, Stanford University.

In its simplest form, IBM explains that AI combines computer science and robust datasets, to enable problem-solving. “It also encompasses sub-fields of machine learning and deep learning, which are frequently mentioned in conjunction with artificial intelligence. These disciplines are comprised of AI algorithms [that] seek to create expert systems [that] make predictions or classifications based on input data.”

To find out more about artificial intelligence (AI), click here.

What is machine learning?

A branch of AI and computer science, machine learning (ML) uses data and algorithms to imitate the way humans learn to gradually improve its accuracy. IBM’s Arthur Samuel coined the term ‘machine learning’ in his research surrounding the game of checkers.

“Machine learning is an important component of the growing field of data science. Through the use of statistical methods, algorithms are trained to make classifications or predictions, uncovering key insights within data mining projects. These insights subsequently drive decision making within applications and businesses, ideally impacting key growth metrics,” said IBM.

To find out more about machine learning (ML), click here.

How is artificial intelligence (AI) and machine learning (ML) being used in fraud detection?

By harnessing AI-based accounts payable invoice automation (APIA) in procurement, the function can elevate their AP operations and protect themselves from fraud. This technology provides users not only with automation capabilities but smart, intelligent capabilities that can learn over time.

“Rather than manually spot-checking transactions – which takes time, resources, and opens up the organisation to additional exposure – modern organisations are turning to APIA to automatically watch and audit transactions in real-time,” said Vroozi.

Procurement functions can also benefit from automated three-way matches to easily check invoices against purchase orders (POs) and goods receipts. As a result, organisations can then pay invoices against pre-approved funds, agreed terms, and goods/services received to reduce fraudulent charges.

“Invoice verification is especially important when it comes to the 50% of invoices on average that get processed without an accompanying PO. It’s much more difficult to catch instances of fraud when there’s no PO and you don’t have anything to match against. In these cases, you must chase down people who can verify the expense or rely on trend analysis to detect fraud – a nearly impossible task without intelligent technology,” added Vroozi.

Alongside automated three-way matches, APIA can be helpful in this scenario to ensure that suppliers who have been pre-approved are paid, reducing the payments made to fraudsters impersonating suppliers. APIA’s can also help organisations stay compliant with constantly changing country-specific e-invoicing and tax compliance laws.

“While fraud can never be completely eliminated, manual payment processes increase exposure. Digitisation and AI can help identify instances of fraud automatically before it impacts your bottom line and your brand reputation. The key is making the investment in the right technology that will set you up for success,” concluded Vroozi.


Featured Articles

AWS Marketplace: Streamlining Software Solutions

AWS Marketplace simplifies procurement for public sector organisations, offers compliance support & prevents vendor lock-in, revolutionising cloud services

Keith Hartley, LevaData CEO, on Philosophy and Procurement

Keith Hartley, CEO at LevaData, discusses how the first principles philosophy and design thinking can revolutionise procurement processes

Mars Snacking's Amanda Davies joins P&SC LIVE New York

Amanda Davies, Chief R&D, Procurement and Sustainability Officer at Mars Snacking, is set to speak at Procurement & Supply Chain LIVE New York

Coupa Software’s Mike Schanker joins P&SC LIVE New York

Procurement Strategy

ServiceNow’s Kirsten Loegering joins P&SC LIVE New York

Supply Chain Management

McKinsey's Report Outlines Issues Facing Procurement Leaders

Procurement Strategy