Skip Navigation
Let's Talk
Cream colored wall with cutouts in various shapes with rounded corners.

Internal Controls Revisited in the Age of AI: Trust the Technology, But Verify the Results

Artificial intelligence is quickly becoming part of the finance function, much like spreadsheets transformed accounting decades ago. Back then, many professionals viewed Excel as a threat to traditional accounting roles. Instead, it has become one of the profession’s greatest productivity tools. AI is following a similar path. Rather than replacing accountants, it’s helping finance teams work smarter, spot risks faster, and spend more time analyzing information instead of manually reviewing it.

But like any new technology, AI brings new risks alongside new opportunities. Generative AI can now produce convincing fake receipts, invoices, and other supporting documentation in minutes, and industry research shows it’s making fraud schemes more sophisticated even as it gives finance teams more powerful tools to detect unusual activity. AI also isn’t infallible. It can misread data, miss important context, and deliver wrong answers with complete confidence. That fallibility means internal controls that worked well a few years ago need to evolve, both to guard against AI-enabled fraud and to verify the output of the AI tools themselves.

Key Insights for Finance Leaders

  • AI creates both opportunity and risk. It can automate repetitive accounting tasks, accelerate month-end close, and analyze thousands of transactions in seconds. But the same technology can simultaneously be misused to fabricate receipts, alter invoices, or support fraudulent expense reports. When you shift the traditional reconciliation process from the Controller performing a high-level review and signing off on a reconciliation, to the Controller running AI and data analytics over the data as an additional level of review, then the Controller can focus their time also investigating the higher risk exceptions generated and making selections from those transactions for further review.
  • Internal controls still matter, perhaps more than ever. Strong segregation of duties, approval workflows, and management review remain your first line of defense. AI should strengthen these controls, not replace them. Use AI to test the process and identify blind spots in each process by having it “break” the control environment to identify ways which controls may be circumvented.
  • Keep people in the review process. AI is excellent at identifying anomalies, but professional judgment is still essential. A few random spot checks by an experienced reviewer can confirm that automated processes are working as intended. Organizations implementing AI governance consistently emphasize the importance of human oversight.
  • Continuous monitoring beats annual reviews. Rather than waiting for the annual audit, AI allows organizations to perform ongoing monitoring throughout the year, helping identify issues while they are still small and easier to address. This practice results in improved accuracy and reliability of financial reporting throughout the year, enabling management to spend more time focusing on cash flow and other strategic decisions for the business.
  • Use AI as a second reviewer. Many organizations are beginning to use AI after transactions have been processed to identify unusual patterns that deserve further investigation, including:
    • Duplicate payments
    • Vendor mismatches or other inconsistencies with the invoices
    • Journal entries posted outside normal business hours
    • Activity in rarely used general ledger accounts
    • Entries with vague or missing descriptions
    • Approvals completed unusually fast
    • Payments made that do not include all approvals
    • Payments just below approval thresholds
    • Employee and vendor address overlaps
    • Payroll amounts that vary from historical levels
    • Revenue entries that reverse shortly after posting

The Bottom Line

As always, the best internal control environment is still built on a foundation of people and technology working in synchronicity. AI can review every transaction, identify patterns that humans might overlook, and significantly reduce the time spent on routine analysis. But experienced professionals remain in critical roles to ask the right questions, investigate exceptions, and exercise sound judgment.

As AI continues to mature, organizations that embrace these tools while maintaining strong governance and human oversight will be better positioned to improve efficiency, reduce fraud risk, and strengthen confidence in their financial reporting. The goal isn’t to replace internal controls with AI, but to make those controls smarter, faster, and more effective.

Insights

As we approach 80 years, Ellin & Tucker remains firmly in the room, driven by a legacy of excellence in teamwork, leadership, and service. Our strength has always been in our people, and together, we’ll continue to stand with the next generation of difference-makers and leaders, ready to shape the future.
Aileen Eskildsen, Chief Executive Officer

Let's Talk