
As we all began downloading computer software and other apps, digital criminals got the idea to sneak all kinds of bad things into our computers, whether it was to steal data, demand ransoms for unlocking data, or just making the computers misbehave. Viruses, malware, and spyware have been regular problems for most of us using computers, tablets and smart phones.
With the rise of AI technologies, it should be no surprise that criminals are leveraging these advances as well. Companies of all sizes need to understand how AI-enabled fraud operates–and how to detect it.
One company, AppZen (AppZen. com) is developing AI tools for large companies that can detect fraud being committed by other AI entities.
AI has significantly transformed the landscape of fraud, enabling bad actors to craft sophisticated schemes that are harder to detect. According to Anant Kale, the CEO of AppZen, AI fraud includes creating fake receipts that look authentic, fabricating entire expense patterns, or manipulating invoices–sometimes by external parties, sometimes by employees.
“It is really left to the crook’s imagination,” Kale said in a recent interview. “From the point of creating fraudulent receipts that look very simple, very authentic rather, to being able to create an entire pattern of spending which may have never happened.”
A company may not be able to detect that kind of scam without AI assistance.
“You also need a robust program which monitors and looks for patterns of spending, and kind of detect anomalies of spending that might be happening, and then figure out a root cause of that,” he said.
For example, the AppZen tool might reveal that an employee is repeatedly ordering expensive dinners or personal items under the guise of business expenses–which would raise red flags for auditors. AI is just better at detecting these patterns faster.
Back before there were AI programs to detect this kind of fraud, I remember a case where a manager was caught submitting meal reimbursements, but only because a sharp-eyed auditor noticed that he was submitting restaurant meal tickets with consecutive serial numbers dated a week or two apart.
The auditor said, “He’s either committing fraud or that restaurant just isn’t very popular.” (The employee had stolen a pad of blank waitress order forms from the diner).
As AI models grow more sophisticated, so do the methods of fraud, ranging from creating convincing phishing emails to mimicking human communication or even intercepting emails to change payment details.
Kale says AppZen’s approach is two-fold: Analyzing documents to determine if they have been falsified in some way using AI, and then recognizing spending patterns that are not normal for the company.
AppZen’s models are continuously updated using a vast dataset covering hundreds of companies and billions of dollars in spend data. By analyzing new patterns and behaviors, they adapt to emerging fraud tactics–like fake websites, phishing emails, or manipulated invoices.
For smaller companies, Anant suggests using AI-enabled tools to review invoices and expenses–particularly if they look different or suspicious.
For example, a small business owner can use basic AI tools like Chat GPT or Claude to verify whether a receipt or invoice is authentic or potentially fabricated.
You can watch an extended video of my interview with Anant Kale at StateBroadcast- News.com
Email steve@compuschmooze.com about how you are using AI. Follow @PodcastSteve on all the socials.

Earlier this month, the Jewish Federation of Southern New Jersey (JFSNJ) hosted the third biennial Love Impact Awards. Named for...

According to Amy Clayman, chief financial officer of the Jewish Federation of Southern New Jersey (JFSNJ), numbers are not just...

April is National Poetry Month, and in observance, we salute American Jewish poet, author, and activist Emma Lazarus. Born on...