News
Insights
Recent Posts
- Planning a Board Retreat that Strengthens Your Nonprofit’s Finances
- Paper Checks May Feel Familiar, But Are They Secure?
- Fraud Awareness Week: Would You Know If Someone Changed the Number?
- How AI Can Help Businesses Fight Fraud
- Get Your Nonprofit Ready for Audit Season Before Year End
- Year-End Tax Planning for Small and Mid-Sized Businesses: Navigating the Post-OBBBA Tax Landscape
- Payroll Tax Compliance: New Social Security Wage Base for 2027
- How Will the 2027 Inflation-Adjustment Numbers Affect Your Year-End Tax Planning?
How AI Can Help Businesses Fight Fraud
By Jake Carruth, CPA
Artificial intelligence (AI) is finding its way into businesses across industries. While much of the conversation has focused on using AI to improve productivity, automate routine work, and analyze data, another area where the technology may offer significant value is fraud prevention and detection.
Fraud remains a serious risk for businesses. Common schemes can involve fraudulent vendors, improper payments, payroll manipulation, expense reimbursement abuse, and unusual accounting entries. As transaction volumes, employees, vendors, and financial processes increase, maintaining effective oversight can become more difficult.
AI won’t eliminate fraud, nor should it replace sound internal controls or human judgment. But when implemented appropriately, it can give businesses another tool to help prevent fraudulent transactions, identify suspicious activity, and respond to risk more quickly.
AI can help teams review more activity, identify exceptions sooner, and focus human attention where it matters most.
1. Identify Unusual Transactions
One potential advantage of AI is its ability to analyze substantial amounts of financial data and identify activity that differs from established patterns.
For organizations processing large numbers of transactions, reviewing every transaction manually may not be practical. Technology can help identify transactions warranting additional attention based on characteristics such as unusual amounts, timing, vendors, or account activity.
- Payments that are unusually large compared with historical activity
- Duplicate or potentially duplicate invoices
- Transactions posted at unusual times
- Unexpected changes in vendor activity
- Journal entries that differ from typical posting patterns
- Payments that fall outside established purchasing patterns
The goal isn’t necessarily for AI to determine whether fraud has occurred. Instead, it can help narrow a large population of transactions into a manageable group for a person to investigate.
2. Strengthen Accounts Payable Monitoring
Accounts payable is an area particularly susceptible to fraud. Fraudulent billing can involve invoices for goods or services a company has never received. Other schemes may involve overbilling, duplicate invoices, fictitious vendors, or employees improperly approving personal purchases.
AI and data analytics can provide another layer of review by comparing information across invoices, vendors, and payment records. A business could look for similar invoice numbers, repeated dollar amounts, unexpected payment changes, or characteristics shared by multiple vendors. When something unusual is identified, the transaction can be routed for further review before or after payment.
Technology shouldn’t replace an effective approval process. Vendor verification, appropriate authorization levels, and segregation of duties remain important controls. Automated monitoring can work alongside these controls to give management additional visibility.
3. Look for Payroll Anomalies
Payroll can be another attractive target for occupational fraud. One classic scheme involves a ‘ghost employee,’ where someone with payroll access directs company funds to a nonexistent employee. Other schemes may involve unauthorized changes to compensation or other payroll information.
Businesses can use analytics to compare payroll information with employee and other business records and look for anomalies. Depending on the data available, management might investigate unusual changes to employee pay, duplicate payment information, unexpected payroll activity, or inconsistencies between payroll and employee records.
Payroll can be another attractive target for occupational fraud. One classic scheme involves a ‘ghost employee,’ where someone with payroll access directs company funds to a nonexistent employee. Other schemes may involve unauthorized changes to compensation or other payroll information.
Businesses can use analytics to compare payroll information with employee and other business records and look for anomalies. Depending on the data available, management might investigate unusual changes to employee pay, duplicate payment information, unexpected payroll activity, or inconsistencies between payroll and employee records.
AI may make this process more efficient by continuously screening information and highlighting exceptions rather than requiring employees to compare multiple reports manually.
4. Monitor More Transactions, More Frequently
Traditional financial reviews are often periodic. Managers might examine information weekly, monthly or when something appears wrong. Technology gives businesses an opportunity to move toward more proactive monitoring.
Instead of selecting only a small number of transactions for review, analytical tools can evaluate a much broader population and identify exceptions based on predefined rules or patterns in the data.
Traditional financial reviews are often periodic. Managers might examine information weekly, monthly, or when something appears wrong. Technology allows businesses to move toward more proactive monitoring.
Instead of selecting only a small number of transactions for review, analytical tools can evaluate a much broader population and identify exceptions based on predefined rules or patterns in the data.
This doesn’t mean every flagged transaction represents fraud. A payment may be unusual for a legitimate reason. But identifying exceptions gives management an opportunity to ask questions earlier.
5. Help Finance Teams Focus Their Attention
Regardless of an organization’s size, finance teams have limited time and resources. Manually reviewing every invoice, journal entry, payroll change, or vendor transaction isn’t always feasible.
Rather than asking accounting personnel, controllers, or business owners to examine every transaction manually, technology can help direct attention toward transactions or trends that appear unusual. Management can then focus resources on higher-risk items while maintaining appropriate human oversight.
This can be particularly valuable when combined with established controls such as approval requirements, segregation of duties, and independent review.
AI Is a Tool, Not an Internal Control Strategy
Despite its potential, AI shouldn’t be viewed as a substitute for the fundamentals.
Businesses still need clearly defined approval authority, segregation of duties, vendor controls, reconciliations, and independent review. They also need appropriate controls surrounding access to financial systems and sensitive information.
AI-generated results require human judgment. An unusual transaction isn’t automatically fraudulent, just as a transaction that isn’t flagged isn’t automatically legitimate.
The strongest approach isn’t AI instead of internal controls. It’s AI working alongside internal controls.
Start With Your Risks
Before investing in a new AI platform or fraud detection tool, consider where fraud could realistically occur within your business.
Which employees can create vendors? Who can change payment information? Who approves invoices? Can the same person initiate and authorize a transaction? How frequently are journal entries, payroll changes, and vendor records reviewed?
A fraud-risk assessment can help businesses identify gaps in their existing controls and determine where additional monitoring may provide value. The scope should reflect the organization’s size, systems, industry, and specific risks.
From there, businesses can evaluate where AI, automation, and data analytics make sense. For some organizations, the best starting point may be monitoring accounts payable. For others, it could be payroll, expense reimbursements, journal entries, or vendor activity.
The objective isn’t to implement AI everywhere. It’s to use technology thoughtfully where it can provide better visibility, complement existing controls, and help your team identify potential problems sooner.
As AI capabilities continue to evolve, businesses have an opportunity to strengthen their fraud defenses without losing sight of what matters most: strong controls, reliable information, and informed human oversight. If you’re looking for support with your organization’s fraud awareness, contact the team at Maner.
The materials provided in the News & Insights section are for general informational purposes only and may not reflect the most current legal, tax, or financial developments. While we strive to ensure accuracy at the time of publication, Maner Costerisan does not guarantee that the information remains up-to-date or free from error. We recommend consulting directly with a Maner Costerisan team member to confirm the applicability and relevance of any information to your specific situation.