2 Link To White Paper

AI-Powered Fraud Detection Is Here

Today’s article was originally published on the Coupa Blog

A company was charged for an unauthorized $30,000 squirrel hunting trip. You read that right – squirrel hunting.

A logistics company unknowingly paid a supplier that didn’t exist to maintain infrastructure that didn’t exist.

The AP team at a technology company received an invoice for goods never purchased.

An executive bought IT equipment with company funds and sold the hardware privately for personal gain.

Yes, these are real stories, shared by sources who wish to remain anonymous for obvious reasons. It’s something no one wants to talk about, but in the complex world of company spend and supplier relationships, it’s very difficult to avoid: fraud, the dirty little secret of the business world.

As we’ve seen above, perpetrators can get creative, and fraud ends up costing companies up to an estimated 5% of annual revenues each year. 5% may seem like a small percentage of annual revenues, but it can add up to millions of dollars in lost revenue for your business, damaged reputations, and a reduction of trust. And it’s not just the amount of money lost that’s important, it’s how the money was lost and how you can approach fixing that.

What Do We Know About Fraud?
The 2018 Report to the Nations Global Study on Occupational Fraud and Abuse found that weak internal controls were responsible for nearly 50% of all instances of fraud.

That’s not good news, especially for mid-market businesses, which typically have fewer fraud controls and bigger cash flow concerns. But if fraud is slipping through the cracks of our internal controls, how do we know that it is so prevalent? And what can we do about it?

Well, at 40%, tips are by far the most common way that fraud is identified, while internal audits and management reviews only identify 15% and 13% of fraud, respectively. Most tips come from employees (53%), but they also come from customers (21%), vendors (8%), competitors (2%), and other sources (21%).

Clearly, it seems some people are trying to do the right thing by reporting fraud. But, as we can see from the 5% of revenue that’s disappearing, many are not.

Unfortunately, cheating or acting unscrupulously for personal gain is an unavoidable part of human behavior. With this in mind, we can turn to the sage advice from Ray Dalio, American billionaire investor, hedge fund manager, and philanthropist, in Principles: Life and Work for guidance: “Have good controls so that you are not exposed to the dishonesty of others.”

Why Most Methods of Fighting Fraud Fail
With so many avenues for both intentional fraud and accidental error, and with the difficulty of telling one from the other, how do we stop the negative impact to our companies?

Audits are a very common method to find fraud, but they pose an interesting dilemma. On one hand, you can audit every transaction, from purchases to invoices to expenses. This method doesn’t guarantee accuracy, though, and takes a great deal of resources.

On the other hand, you can save time by auditing a percentage of these transactions at random and hope that you find any fraudulent transactions. Either way, audits are very manual, whether your company has an internal audit team or outsources the task.

So, what’s the best strategy to fight fraud? It might be to go beyond what is humanly possible.

Leveraging technology to find fraud is not a new idea. Data analysis has already been used to find significant financial crimes. In fact, data monitoring and analysis and surprise audits have been correlated with the biggest reductions (52% lower losses) and duration (58% faster detection) compared to alternatives. The difference between yesterday’s technology and today’s is that now, we can bring these tools into processes at our companies to identify the everyday fraud that chips away at profitability and shareholder value.

Machine learning can look at a user’s spend across the entire organization to identify suspicious activities. This capability to look for fraud in a comprehensive way is important; if someone finds they can get away with fraud in one area, it suddenly becomes much more tempting to commit fraud in other areas, and soon, money is trickling steadily out of company accounts.

Based on fraud research, we know that 77% of all instances of occupational fraud come from Accounting, Operations, Sales, Executive/Upper Management, Customer Service, Administrative Support, Finance, and Purchasing. Given this information, it’s critical to get all spend for these departments (POs, invoices, contracts, expense reports) in one single platform. Only then can you leverage AI and safeguard company cash by controlling in-flight transactions that are potentially fraudulent across the organization.

Get Ahead of Fraud with AI-Powered Audits
It’s time to outsmart fraudsters when it comes to managing your spend. With today’s technology, you can shift from detection to prevention, comparing your organization’s behavior to the norm at a granular level and identifying potential fraud activity proactively.

Artificial intelligence has advanced to point where it can do the dirty work of aggregating and analyzing billions of transactions from the business community to compile profiles and learn what constitutes “normal” behavior. Then, your systems can compare each transaction to what’s normal and flag anomalies for your finance team to review. This approach can identify and stop losses to fraud more effectively than traditional audit processes while freeing up resources to focus on higher value tasks to drive company strategy.

Fraudsters can be creative, but so can you. Using AI-powered solutions, you can arm yourself with valuable tools to combat fraud and accidental errors and protect every dollar of your revenue from slipping through the cracks.

Scroll to Top

Contact Us