A better way to detect fraud is coming. Artificial intelligence helps humans overcome traditional investigation barriers and enact proactive monitoring.

The challenges of detecting and preventing fraud

The modern world contains vast networks of interconnected institutions, a myriad of financial tools and an ever-increasing volume of transactions. Criminals hide behind this complexity, taking advantage of oversights and gaps in coverage. A multi-million dollar company likely won’t notice a few hundred dollars taken from different accounts, but the total adds up over time.

With traditional processes, employees lack the capacity to monitor every transaction. Government employees want to spend more time on front-end prevention as part of their job, but nearly half reported that their biggest challenge to achieving this goal is high work volume. Even if workers had more time, manual data gathering and analysis of every transaction is not cost-effective or fool-proof.

Audits are not designed to catch fraud, per se. They generally focus on providing reasonable assurance that the financial statements are free of material misstatement, whether due to error or fraud. These reviews rely on provided information that covers only a portion of transactions, creating sample risk, and are often conducted under tight time and budget constraints, limiting the ability to thoroughly investigate potential red flags or anomalies.

Forensic accounting investigations dig deeper into records but still only cover a fraction of transactions while carrying a hefty price tag. Organizations only turn to these when fraud grows big enough to be detected.

AI empowers proactive monitoring

AI is revolutionizing financial investigations. This technology not only ingests mass amounts of data but also organizes and analyzes the information to expedite review. Forensic accounting teams gain access to the entire library of transactions, eliminating sample risk and providing a comprehensive view of the entire financial history.

Sorting and categorizing data is a laborious manual task. AI reduces that burden. Optical character recognition has been instrumental in digitizing financial documents, but the tool has limitations. Machine learning elevates OCR technology with its ability to understand and interpret a variety of financial document formats, including unconventional or unstructured layouts.

AI automatically cleans and reconciles data, identifying inconsistencies, duplicates, missing information and other errors that might escape human detection — all in a fraction of the time manual processes require. The technology links related transactions across entities and accounts and categorizes them, ultimately generating a central database for streamlined search and analysis.

With current processes, professionals spend 90% of their time on data management tasks, leaving little time for analysis. With AI, the review and reconciliation timeline shrinks from weeks to hours or even minutes, so humans can focus on higher-value tasks such as interpreting results and constructing a narrative of the transactions. The data management capability makes fraud investigations more efficient, effective and affordable.

AI-powered verified financial intelligence (VFI) solutions support human analysis by recognizing subtle signs of fraud that might otherwise go undetected. While technology cannot make conclusions, it can point professionals to transactions that need further investigation. Experts can spend time answering questions instead of identifying them.

The constant reconciliation provided by AI allows organizations to monitor transactions as they occur. VFI platforms can flag any unusual activity for closer inspection, preventing significant losses without an exorbitantly high price tag. An AI system must provide context and evidence to its flagged anomalies, such as the nature of the transaction, the parties involved and the timeline. Early systems lacked this capability, which actually led to more work. The new generation has made significant improvements in this area.

Another AI benefit: storytelling. Analyzing data is one thing; communicating the investigation’s findings is another. Advanced tools turn an overwhelming dataset into a visual narrative so accountants and investigators can help all stakeholders understand the conclusions and how they were reached. The evidentiary chain of custody remains unbroken, resulting in auditable and courtroom-ready evidence.

Humans are indispensable

To be absolutely clear — AI will not replace the need for humans in fraud investigations. While technology can identify anomalies, it can’t understand external context or develop hypotheses. Human critical thinking is imperative to reach — and explain — conclusions. Of course, an AI platform cannot present findings to the board or testify in court. And there is the relationship component. Accountants and investigators must interact with clients, lawyers and other parties during the case, which is beyond technology’s capabilities.

CFOs and other financial leaders bear responsibility for overseeing the monitoring process and making the ultimate decisions on anomalies. They must carefully consider where to apply VFI because no single platform covers an entire financial system.

As technology evolves, the software used by financial investigators and CFOs will become increasingly similar. Technology can put “eyes” on every transaction, empowering organizations of all sizes to catch fraud before it empties accounts. The result will be a more secure financial system. As the volume and complexity of financial data grow, the widespread implementation of AI in fraud detection will become not just an advantage but a necessity.

=====

Tod McDonald, CPA, CIRA, is the co-founder of Valid8 Financial. Early in his career, Tod was an auditor with Ernst & Young and has spent decades navigating complex financial situations, including leading an investigation that unraveled a $200 million real estate investment Ponzi scheme in Washington State. Motivated by this experience, he co-founded Valid8 Financial to build a Verified Financial Intelligence solution to expedite data prep, eliminate sample risk and improve the speed and quality of rendering a professional opinion on complex financial cases.