AppZen’s Insights uses AI to automatically detect expense reporting fraud

For enterprises of a certain size, expense reporting is a significant time sink. Nearly 43 percent of businesses manually check each expense to ensure it aligns with corporate policies, according to Certify, and it’s not just lost, late, and erroneous receipts they have to look out for. Expense report fraud costs companies an estimated $2.8 billion per year.
Andrew Chiang, head of product at AppZen, thinks artificial intelligence (AI) has the potential to ease the burden. “You don’t need a manager to approve and see every expense report,” Chiang said. “There’s a better way.”
To that end, the San Jose, Californa-based startup today announced the launch of Insights, a feature within Expense Audit — its software-as-a-service platform — that eliminates the need for managers and auditors to approve travel receipts individually, line-by-line.
It’s available as an additional product offering to current AppZen Expense Audit customers.
Previously, Expense Audit prompted account administrators to track and report expenses submitted through Workday, Oracle, Coupa, Concur, and other reporting services themselves. Now, AppZen is confident enough in Expense Audit’s track record to relegate those tasks to algorithms.

Above: A dashboard showing expense trends over time.

Image Credit: AppZen

“This is an exciting update for us as it proves how AI can truly disrupt an age-old control and reimbursement process in finance while massively improving compliance and employee satisfaction,” former Fujitsu America vice president and AppZen cofounder Anant Kale Kale said. “This new product allows senior managers to spot leakage, misuse and bad behaviors across the organization. We also understand that approving expense reports is not the best use of these talented peoples’ time. That’s where this new product comes in — as the insight layer to our expense audit product.”
Here’s how it works: Expense Audit assigns employees a score — the AppZen Behavioral Index — informed by each report, receipt, and travel document they submit. Leveraging hundreds of internal data sources and the web, Insights can recognize and authorize reimbursements for items within clients’ approved list (e.g., travel), and highlight possible transgressions (alcoholic beverages) that warrant a closer look.
Within the Expense Audit dashboard, managers can filter expenses between policies, cost centers, and departments, and quickly drill down to top spending areas and high-risk employees.
“It provides managers with insights,” he said. “They can look at how the team’s doing and see what’s going on.”
The idea is to surface information that’ll lead to “proactive” and “impactful” conversations with repeat offenders, and to flag under-the-radar expenses like spa treatments and gentleman’s clubs that might otherwise slip through the cracks. Chiang gave an example: one of AppZen’s clients discovered with Insights that an employee had repeatedly expensed “West Palm Beach Steakhouse,” which turned out to be a strip club.
Chiang acknowledges that the system won’t always get it right, which is why it allows managers to mark false positives. But Ryan Floersch, senior director of product marketing, claims that it reduces audit costs by 80 percent, and that it’s able to detect ten times the number of fraudulent claims a human can.
“The current status quo is, expense report auditors sample two to three percent of reports submitted,” Floersch said. “AppZen Insights brings it up to 100 percent off the bat.”
AppZen isn’t the only startup applying machine learning to expensing, of course. Pleo, which launched in July 2016, offers a prepaid MasterCard product that tracks employee spending and automatically categorizes transactions. And London-based Soldo’s solution lets clients lay out different expense criteria for each employee, contractor, and spending department.
But Chiang is confident that AppZen’s AI-powered backend sets it apart from the competition. “Our goal is to be the most accurate and trusted platform for automated business processes,” Chiang said.
AppZen’s more than 290 customers include Fortune 50 heavy hitters like Amazon, Comcast, Salesforce, DHL, Tesla, Intuit, Hitachi, Sunrun, Airbus, and IBM. To date, the company has raised $17 million in funding to date.
Source: VentureBeat