Artificial intelligence used to audit expenses

Register now

Most enterprises that don’t use artificial intelligence only audit up to 10 percent of their spending, while companies that do use AI are able to audit virtually all their invoices, contracts and expenses, according to a new report.

The report, from AppZen, a company that provides artificial intelligence technology for auditing company spending and travel expense reports, not surprisingly finds advantages to AI technology. AI is able to flag 8.7 percent of expenses as high risk, usually because of unauthorized expenses, errors in keyed-in amounts, and duplicate spend. AI also flags 4 percent of invoices as high risk, typically due to price, discount or terms that don’t match the contract, inflated prices compared to market data, and duplicate spending across invoices or with T&E.

“There’s a lot that can go wrong in the enterprise typical organization’s spend audit process,” said AppZen CEO Anant Kale in a statement. “Manually reviewing and auditing vendor invoices and employee expense reports is a time-consuming and error-prone endeavor. Most companies resign themselves to conducting partial audits well after the transactions have occurred, which leaves companies at risk for errors, waste and fraud. AI can improve the audit processes by auditing 100 percent of spend in real time and before payment to uncover expense misuse, double-check invoices, find regulatory violations, discover fraud and duplicate, double payments, and most importantly, audit before you pay.”

The report highlighted some “creative” expenses that employees submitted for reimbursement last quarter, based on data from AppZen’s own customers, which include Novartis, Intuit, Airbus and other Fortune 500 companies. They include lunch with an oligarch, a strip club, an “Airbnb” visit on a friend’s couch, a set of cufflinks, private helicopter rides, and an “extramarital” travel companion. The most creative accounts payable transactions included an invoice for IT equipment that was quintuple the market price, a partner contract that expired a year ago, an invoice from a prohibited supplier that the government won’t work with, an automatically renewed contract for a no-longer-used service, and a suspicious-seeming supplier invoice that happened to have the same remittance details as an internal employee.

The report also looked at how well AI technology is able to catch illegal payments to foreign officials, foreign-owned government entities, sanctioned companies, and entities on export control lists. Based on AI discovery, for every 10,000 expenses, two contain a regulatory violation, and for every 10,000 invoices, one contains a regulatory violation. That can save companies from being exposed to stiff penalties, as the average corporate sanction imposed for a Foreign Corrupt Practices Act violation is $233 million in 2019, according to Stanford Law School.

Despite the risks of rubber-stamping expense reports, auditors approved an average of 79 percent of above-limit expenses, according to the report, a sign that it may be time for companies to revise their expense thresholds. For example, the average limit for alcohol spending was $266, but 86 percent of alcohol purchases expenses over the threshold are approved, with an average above-limit approved amount of approved being $1,044. Similarly, while the expense limit for golf is an average of $97, AppZen found auditors routinely overlooked the average over-the-limit expense, which was $742.

Another example of over-the-limit spending was for gifts. The average limit for gifts was $88, but 78 percent of above-limit gift purchases and expenses are approved, with an average above-limit approved amount of $509.

“We’ve heard anecdotally that some companies may be unwilling to enforce heavily because they acknowledge that some of their employees may not know what those policies are,” said AppZen chief marketing officer Jamie Barnett.

Yet the report also found that 85 percent of companies that use AI for expenses are automatically approving low- and medium-risk transactions. That can help get employees paid more quickly, but it also could expose them to fraudulent claims.

“They’re starting to get behind the notion of streamlining the process significantly,” said Barnett. “A lot of companies start to believe it is beneficial to remove some of the steps along the way, particularly manager approval of the expenses, which helps them to get people paid back faster. Whereas in the past, it may have taken a week or two to reimburse spend, today we’re seeing organizations able to turn it around in one to two days. It’s a pretty significant differential, and that’s one of the reasons they’re putting their low- and medium-risk expenses on automated approval and just looking at those higher-risk spend items.”

The report is available here.

For reprint and licensing requests for this article, click here.
Artificial intelligence Audits Expense management Expense management software Corruption