How AI Can Prevent Employee Expense Fraud

In a recent article, SAP Concur noted that less than 1% of employee expense claims are rejected, despite being a major cause of corporate fraud. We spoke to Travis Andrade, Senior Product Marketing Manager at AppZen, the AI company powering SAP Concur Detect (an expense management solution) to understand how AI-based technologies can take over expense management and reduce corporate fraud. With his inputs, in this article we look at:

• Three types of expense fraud
• How AI can prevent expense fraud
• Why using AI for employee expense management is excellent for businesses

Employee expense claims constitute a significant part of operational expenditure (OpEx) for any corporate. In recent years, the emergence of new forms of traveling, accommodation, and commute has made spend management even more complex. Employees now want more freedom on where to spend: from booking their own Uber to choosing their preferred hotels/airlines. In fact, 42% of employeesOpens a new window would rather book directly from a supplier’s website than route it through their organization’s travel unit. Faced with this decline in control and a massive variety of products/services used by employees while traveling, expense fraud becomes a pressing issue.

In recent research, SAP Concur found that less than 1% of employee expense claims are rejectedOpens a new window . We spoke to Travis Andrade, Senior Product Marketing Manager at spend management technology company AppZenOpens a new window . During this enlightening discussion, he revealed how expense fraud is cutting into OpEx margins, necessitating new ways to curb fraudulent practices.

Learn more: How to Tackle the Evolving Nature of Expense FraudOpens a new window

The Top 3 Areas of Expense Fraud

Expense fraud is not always intentional. Employees may perceive a certain amount as rightful payment and claim it despite the expense not being covered under corporate policies. Andrade tells us about the top fraudulent instances identified by AppZen, in addition to other common scenarios:

1. Repeated claims for a single expense

This can be a rather expensive type of fraud. “The most common type of fraud detected by AppZen is a duplicate expense that occurs when the same expense is submitted for reimbursement more than once. While this could be by mistake, our data shows that 27% of the time these instances indicate fraud, as AppZen has identified two different employees attempting to expense the same thing,” Andrade says.

2. Claims for unused resources

Corporates often allocate a specific daily allowance for their on-site employees. However, it is possible that employees spend only the bare minimum of this amount and pocket the rest. Worryingly, this is viewed as “not unethical” by a large number of employees, leading to unintentional expense fraud.

3. Claims for fake expenses

Employees could “pad” bills in many ways, especially if organizations do not require invoices for expenses below a particular threshold. Free meals, kilometers on rented cars, and personal costs (like liquor and entertainment) couched as “business expenditure” are among the most common cases of expenses fraud.

Interestingly, a small share of the claims submitted is actually fraudulent. “Overall, we have found that misuse and fraud account for 10% of submitted expense reports,” Andrade tells us. However, the cost-to-company is extremely high, as expense fraud generally involves large sums of money.

“These high-risk expenses account for 33% of the overall T&E [Travel & Expenditure] spend submitted to AppZen,” reveals Andrade. In other words, only a tenth of expenses are fraudulent but constitute over one-third of the total claims value.

So how do organizations combat this rising menace? With the number of claims rapidly rising and so many expense types to reconcile, how does HR scan every report in detail, not missing a single anomaly? According to Andrade, Artificial Intelligence (AI) might have the answer.

Learn more: Are There Burglars in Your Boardroom?Opens a new window

AI Can be Game-Changing for Expense Fraud Management

Detecting expense fraud and flagging/approving claims has always been an effort-intensive process. Currently, audits and reviews have been able to spot only 15% and 13% Opens a new window of fraudulent instances, respectively. Clearly, a smarter alternative is needed – leveraging data stored on corporate travel portals, pay cards, accounting systems, and more, in combination with AI engines. Not only will this reduce the required effort, but it will also drive accuracy and even predict future risks based on historical patterns.

“Companies using AppZen to audit T&E spend have seen a significant improvement in their ability to detect instances of misuse and fraud. A majority of our customers have been able to detect 10 times more risk than before,” says Andrade.

Furthermore, AI is capable of scanning data in near real-time, so that reimbursement can be immediately checked in case of potential red flags. Andrade adds, “This increase in risk detection has resulted in up to a 5% decrease in spend, as AppZen catches non-compliant expenses before the dollars are reimbursed, saving companies from paying for things they otherwise would have let slip through the cracks.”

AppZen works on a simple premise. Its AI engine looks at 100% of claims (compared to just the usual 20% examined by human auditors, according to AppZen’s founder). This happens in real-time, disbursing rightful claims on the same day and directing doubtful ones to human auditors. As a result, the solution was able to uncover expense fraudOpens a new window , which might have otherwise been overlooked – like a $3,000 Starbucks gift card that one employee claimed as “coffee with a contact.”

Learn more: Creating Smart Travel Spend Management Frameworks with Data AnalyticsOpens a new window

Why AI for Employee Expense Management is a Good Idea

Expense reimbursements account for 14% of all corporate fraudOpens a new window . This is a sizeable number that can be prevented easily by reassigning the auditing task to technology. Andrade explains why: “Most companies today are resource constrained – audit teams are often burdened as the number of expense reports grow at a rate faster than they’re capable of auditing. This limits their ability to audit at scale, as most are only reviewing a small portion of their T&E spend.”

AppZen applies next-gen technology to solve this long-standing issue, while also reducing claims processing costs by 80%. Sift Science is another company using AI and machine learning for fraud detection/management, looking at a wide range of transaction areas, within and outside of HR.

With time, AI is expected to become a staple for employee expense management, bringing an exceptional level of speed, intuition, and accuracy to this field. The automation of this task and the consequent reduction in company costs will significantly improve the bottom line of an organization.

Do you plan to use AI for employee expense management? Share your insights with us on FacebookOpens a new window , LinkedInOpens a new window , or TwitterOpens a new window . We would love to hear about your experience with AI-based technologies.