Voices: 5 common types of back office occupational fraud, and how Artificial Intelligence (AI) can stop them

February 15 2018


Published: Accounting Today

By: Anant Kale

Often with their gray file cabinets and outdated paint jobs, corporate backrooms aren’t the most scintillating destinations, or topics of conversation. Yet for many companies they remain vitally important as the epicenter of occupational fraud detection.


So what are the most common tactics of employees looking to defraud companies on their expense reports, and how is the burgeoning field of artificial intelligence already helping stop them?


1. Mislabeled expenses: This trick is as old as cashless transactions themselves. For most of us, when we purchase, say, pillowcases at Walmart, we rightly expect to see the word “Walmart” spelled out clearly on our bank statements. But of course transparency isn’t as much of a value in the gentleman’s club industry.


To monitor these transgressions for employers, today’s AI systems can automatically cross-check each expense report receipt with data from review sites like Yelp and TripAdvisor. For example, one machine learning program detected employee meals reimbursed at “K-Kel, Inc,” which, it turns out, was actually a receipt for the Spearmint Rhino strip club in Las Vegas.


2. Subtle transfers of wealth: Sometimes it’s fun to take the team for coffee. We’ve all been there; it’s not a crime. Sometimes it’s also OK to spend more than $20 at Starbucks—maybe grab a sandwich for the flight, to go along with the latte and the fruit plate.


But what’s not OK is when employees expense a hefty total at a retailer like Starbucks when they’re actually refilling their gift cards with the funds. This isn’t treating the team or treating your belly; it’s treating yourself to low-level fraud.


3. The illicit upgrade: Those who travel often for work learn something pretty quickly: it’s a lot less glamorous than it seemed when we were younger. Such a realization can lead to frustration, and even a sense of entitlement that can manifest itself in upgrades at odds with corporate policy.


In one interesting example, one company’s AI technology caught an employee from a U.S. auto manufacturer upgrading his rental car—against policy—to a Mercedes.


4. Gateways to larger crimes: Occasionally employees will find themselves involved in larger, fraudulent webs. Whether it’s via a side hustle or a broader scheme at the day job, some employees have eyes on prizes well outside the company’s established parameters.


Along those lines, AI discovered, for a major semiconductor company, that one of the attendees at a business meal was an employee of a company on the U.S. Department of Commerce’s barred list.


5. Using the “Miscellaneous” distinction to hide stuff: The “misc” category is helpful; it can be a major time suck to itemize all the tiny transactions of a common business trip. In other words, no one wants to hear all about that $2.25 spent at the parking meter in front of the business lunch.


Yet where there is great freedom there’s also opportunity to exploit. AI has helped companies catch numerous sorts of expense policy violations loitering in the “misc” column. These include: extra luggage, personal dry cleaning, TSA Pre-checks and other willful missteps.


Expense reporting auditing clearly isn’t the most sensational topic, and perhaps that’s why conversations about it rarely leave the drab walls of the backroom. But what is exciting is saving money and reinvesting it back into the company, and its people. In 2018, “artificial intelligence” doesn’t mean robots making humans cocktails while we lounge by the pool; instead, it is beginning to serve us in daily, practical endeavors like catching occupational fraud and making the workplace more transparent. 



Otras Publicaciones


Últimas Publicaciones