Level upon level of corporate management also failed to detect that the same employee was running a scheme to sell more than $200,000 in company equipment on eBay.
Only a fraction of expense reports are closely examined, so it is no wonder that companies experience more than $7 billion in annual losses from fraud, according to the Association of Certified Fraud Examiners.
By using robots, instead of relying on random spot checks, companies are catching fraud more than twice as fast and fraud losses are halved, said Andi McNeal, director of research for the Association of Certified Fraud Examiners.
That is what happened when the industrial equipment company put in place an artificial-intelligence program from Oversight Systems, which was able to quickly ferret out the culprit.
“It started out as a small infraction that led to an investigation that led to other things,” said Terrence McCrossan, chief executive of Atlanta-based Oversight Systems, which audits about $2 trillion worth of employee spending each year and works with employers like the U.S. Department of Defense, McDonald’s and General Electric.
The expense reporting universe is being overhauled to use artificial intelligence to get a 100 percent overview of employee submissions. In addition to monitoring fraud, companies are streamlining the way employees file expenses.
Soon, employees around the world will stop fussing with paper receipts and crying over hotel bills, then waiting weeks to get reimbursed while their paperwork travels through the corporate labyrinth. Managers will no longer be stuck in the middle of the process, policing spending, and companies will stop losing so much money to waste and fraud.
TEST CASES
Some changes have already occurred, ranging from corporate card charges that automatically attach to electronic expense reports to seamless experiences for business travelers who stay at approved hotels.
One of SAP Concur’s newest offerings is Concur Detect by AppZen, which does a 100 percent audit of incoming expense reports.
AppZen analyzes expenses by looking for risk. Only about 10 percent of expenses that flow through a company have a problem that needs to be addressed, said Anant Kale, CEO of AppZen, based in San Jose, California.
The algorithm can clear expense reports with no issues almost instantly, so that these employee outlays can be reimbursed as quickly as two days.
If a charge has a red flag, it goes to a human auditor. One Concur Detect customer, Portola Pharmaceuticals Inc, said it had reduced the number of expense reports that required review by one-third.
Kale has been surprised by the kind of problems that are popping up since AppZen’s 2016 launch.
“Employees are claiming the same expense multiple times. That happens more often than you can imagine,” Kale said. “It’s not fraud, but an honest mistake.”
AppZen also finds many expenses that are disallowed by corporate policy. Some of these are for strip clubs, in-room movies during business travel or charging gifts at a hotel shop.
Oversight Systems has identified questionable expenses like eyelash extensions, lost sunglasses and an employee who billed for a new shirt after he spilled coffee on himself on the way to a meeting.
There is also true fraud. Oversight Systems, for instance, found an employee who expensed for parking over and over using the same receipt each time. By the time the fraud was discovered, the parking lot no longer even existed.
What makes the difference between catching wrongdoers and companies’ losing money? Better compliance and making audits more efficient, said the Association of Certified Fraud Examiners’ McNeal.
As much as machines can learn and improve their performance, people are more complicated. AppZen, for instance, has yet to run a clean screen on a company where it catches no problems, no matter how much effort a company puts into employee education and catching disallowed expenses before they are filed.
“You’re never going to get all of them to comply - that’s just human nature,” McNeal said. “You’re just trying to let the fewest grains get through the sieve.”