When you’re sifting through hundreds of receipts and come across a questionable one – how can you confidently confirm it’s fraudulent? What gives it away?
Long gone are the days of picking up blank receipts from a cabbie. With the advent of technology, fabricating receipts has never been easier.
With just a quick Google search, I came across several online tools that generate a plethora of receipts – for Uber, hotels, fuel, meals, coffee, and much more. I created a few receipts using the tool and compared them to legitimate receipts. Even having spent many years in the fraud detection space, it was tough for me to spot the differences at first glance. The more I stared at each receipt, the more real they looked.
Spot the fake receipt
We polled LinkedIn users and asked them to spot the fraudulent receipt amongst the two options provided. More than knowing the receipt they chose, it was interesting to read their reasoning. A majority thought receipt #2 was fraudulent. Here are some of their reasons: it didn’t have the content expected on a fuel receipt, an added grey box seemed suspicious, or it appeared like it was typed. A handful thought receipt #1 was fraudulent because it looked ‘too perfect’ and had content that didn’t seem familiar. Some even thought both were fraudulent! Take a close look yourself at the two receipts – which do you think is fraudulent? Why? (Hint: check the bottom of this post for the answer) To the naked eye, identifying fake receipts amongst the hundreds you’re sifting through is tough. The lines start to blur between what’s legitimate and what’s not. A few key identifiers I use to spot fraudulent receipts are: spelling errors, missing data, overlapping timestamps with other receipts in the employee’s expense reports, and incorrect vendor data. Developing an investigative mindset is key to identify fraudulent expense reimbursements. Think like a fraudster.
As we think like a fraudster, let’s look at some fraud schemes employees could engage in by submitting fraudulent fake receipts:
- Billing for expenses that never occurred or were cancelled
- Billing the company for personal charges
- Submitting duplicate receipts
For a team of auditors sifting through pages of receipts as part of the T&E sampling process, identifying one of these expense reimbursement schemes is tough. The audit team could manually research and confirm each hunch they have – whether it be vendor verification, amount total confirmation, and timestamp verification among others. However, the process of manual verification is inefficient and too time consuming to make it cost effective to research every hunch.
That’s where artificial intelligence comes in. Imagine a tool where all receipts are ‘looked at’ by a machine. All data from the receipt and the employee’s other expense reports are combed to spot patterns. Patterns like this one: an individual submitting multiple reimbursement requests for meals he had with his wife using receipts he printed from an online receipt generator. These additional receipts push him above the normal spending distribution for someone at his level in Sales. An AI tool, like AppZen, would spot this pattern and notify the company’s audit team to look closer for fraudulent transactions.
If you guessed receipt #1 was fraudulent – you’re right! Good use of your investigative skill set 😎