The grand scheme of things
Read Time: 6 mins
Written By:
Felicia Riney, D.B.A.
A decade ago, examining suspicious transaction activity often meant manually building a narrative, according to Victor R. Cardona, CFE, senior vice president, head of anti-money laundering (AML) and Bank Secrecy Act (BSA) officer for Golden State Bank in Los Angeles, California. He recalls spending significant time moving among disparate systems, spreadsheets, documents and searches to gather information and piece together a customer’s activity. The work required fraud examiners like him to exhaustively find, organize and connect information before they could even begin evaluating what it meant.
Today, Cardona’s work looks starkly different. When reviewing an alert involving a customer with a high volume of transactions, for example, he must quickly understand the customer’s profile, transaction activity, associated counterparties and potential risk indicators. To do that, he relies on transaction monitoring systems, case management platforms, approved research sources and generative artificial intelligence (AI) workflow tools. These technologies organize information, summarize activity, identify inconsistencies and assist with drafting investigative analyses, helping Cardona connect facts across multiple sources and manage the large volumes of information characteristic of AML and fraud reviews.
Cardona describes this change in his investigative work as a transition from manual information gathering toward higher-level analysis. Technology allows fraud examiners to spend more time evaluating evidence, applying professional skepticism and following the investigative trail. Maintaining a “human-in-the-loop” remains critical, he says, particularly when using AI-generated output, which fraud examiners must review, challenge and validate before relying on it.
“The fraud examiner still has to determine what those facts mean, recognize what doesn’t make sense and decide where to dig deeper,” Cardona emphasizes.
Cardona’s experience reflects a broader evolution in financial crime, as technology transforms the skills fraud examiners need, the roles they perform and how they define their work. This article provides a glimpse into the life of a modern investigator and explores the impact of AI and other technologies on investigations, how institutions are balancing innovation and accountability to counter emerging threats, and experts’ predictions for the anti-fraud professional of the future.
Audi6 was a laundering service that promised anonymity and speed. Operating on an industrial scale, the cyber-criminal network received funds from customers and concealed them through a complex chain of transactions built around accounts opened with stolen identities. The operation was dismantled through investigations conducted by the U.S. Secret Service and IRS Criminal Investigation, the Polish Police, EU member states with international law enforcement partners, and Europol, which alleges that Audi6 laundered more than 336 million euros between 2022 and 2025.
This case illustrates a broader challenge for fraud examiners investigating money laundering and other financial crimes: Technology allows criminals to move and obscure illicit funds with unprecedented speed and reach. According to the U.S. Treasury’s 2026 National Money Laundering Risk Assessment, “The rapid growth of emerging technologies, which is improving consumer access and experiences, also creates opportunities that illicit actors can exploit to obscure the origin of illicit funds more quickly and on a global scale.”
With the growth of more complex fraud and financial crime, financial institutions are challenged to keep up with criminal organizations’ accelerated use of deepfakes, synthetic identities, automated scams and AI-enhanced fraud schemes. Research released in June 2026 by Mitek Systems and financial services research and advisory firm Datos Insights warns of the rapid growth of synthetic identity fraud as a major threat to financial institutions. According to their findings, fraud losses in the U.S. tied to synthetic identities amounted to about $2.94 billion in 2025, a significant increase from $1.8 billion in 2020.
Generative AI allows criminals to create realistic deepfakes that impersonate executives and loved ones to deceive victims into sending money. Vera Senel, CFE, an anti-fraud professional at a financial consulting company in Turkey, says the technology also enables identity theft and poses challenges for financial institutions that rely on remote identity verification.
Increased automation provides another challenge for anti-fraud professionals. Victoria Meyer, CFE, FCCA, a director at Swiss International Business Academy, tells Fraud Magazine that automated processes mean that “accounts can be created in bulk, lying dormant or passing innocuous amounts between themselves until they’re ready to be called into action and then quickly abandoned when they’ve fulfilled their purpose.” She adds that autonomous agentic payments and digital assets allow funds to be layered and dispersed in seconds, and they often involve legitimate online businesses as credits are refunded to subsequent addresses.
“The active lifespan of an illicit account has been reduced to days or even hours, turning the identification and blocking of criminal assets into a high-speed game of whack-a-mole as funds are moved onwards,” says Meyer.
According to Meyer, AI can also be used to compromise a bank’s control framework. Money launderers can identify vulnerabilities with AI and modify a bank’s controls to create loopholes, where they can cause significant damage and remain undetected. Mitek’s research report states that 40% of financial institutions reported an increase in attacks linked to AI, posing additional challenges to institutions that are fragmented in their fraud detection approaches.
Anti-fraud professionals investigating money laundering face an additional challenge as technology enables criminals to operate across increasingly decentralized networks. Tony Prior, CFE, who has a background in law enforcement, regulation, consulting and banking in Australia, tells Fraud Magazine that financial crime, which once depended heavily on physical cash, relatively simple structuring or localized fraud networks, can now operate through highly distributed digital ecosystems. Bad actors can use technology to automate scams and rapidly move or disguise illicit proceeds across borders and platforms. They also use large networks of money mules who transfer money for them.
Senel tells Fraud Magazine that financial institutions may underestimate mule networks, in part because some people coerced into moving illicit funds don’t realize they’re facilitating a crime and may not present an identifiable risk during customer onboarding. Criminals can use social engineering to convince victims to receive, deposit or transfer funds on their behalf, often in amounts below traditional alert thresholds. The funds can then be layered through multiple accounts and transactions, further obscuring their source.
“A sophisticated mule network can evade most of the traditional customer risk assessments and transaction monitoring scenarios most institutions have in place to prevent and detect financial crime,” says Senel.
According to Enobong Essien, CFE, a deputy superintendent and senior prosecutor for the Economic and Financial Crimes Commission (EFCC) in Nigeria, and Adebayo Soares, CFE, an investigator with the Norwegian Refugee Council, decentralized finance poses a challenge for AML and anti-fraud processes because digital assets and cryptocurrencies lack regulations and well-resourced regulatory bodies. Information exchanged solely online may lack a physical means of verification and instead rely on details provided by the user.
Digital fraud is a top predicate offense reported by financial institutions in Nigeria. Data breaches result in stolen identities and unauthorized withdrawals from bank accounts. Essien says that vague know your customer (KYC) standards make it difficult for fraud examiners to identify money laundering perpetrators as they rapidly move funds to other bank accounts with the mere click of a button and then withdraw them at ATMs or through other cash channels. Anti-fraud professionals and institutions confront increasing pressure to identify networks and exchange intelligence quickly.
As financial institutions grapple with growing challenges related to cryptocurrency transactions, decentralized finance platforms, cross-border payment networks, beneficial ownership transparency and increasingly complex sanctions evasion techniques, AI is helping anti-fraud and AML professionals take a proactive, predictive risk management approach.
Once considered experimental technology, AI’s use in AML has become a core component of modern financial crime compliance programs. Organizations are also adopting AI to automate portions of customer onboarding, enhanced due diligence, sanctions screening and suspicious activity report (SAR) preparation. According to SEON’s AI Reality Check 2026 report, 98% of the fraud, risk and compliance leaders surveyed have added AI into their institutions’ fraud and AML workflows. Not only can AI-powered monitoring systems identify emerging risks in near real time, but they can also support “perpetual KYC” programs, which continuously monitor customer behavior, detect significant changes in risk profiles, identify changes in ownership structures and flag unusual transaction patterns.
Organizations must contend with the risks and governance requirements that accompany AI’s rapid adoption. Often promoted for its ability to revolutionize compliance programs through automation, faster investigations and improved detection, AI isn’t a magic bullet. Although AI assists organizations in keeping up with threats, it doesn’t eliminate long-standing challenges around data quality, system integration, false positives and human oversight. And implementing AI can introduce challenges of its own, from explaining and validating model decisions to ensuring that automated tools don’t accelerate flawed processes or produce more alerts for investigators to review.
Those limitations make strong governance essential. Senel says institutions need clear responsibility for AI systems throughout their life cycles, including who owns a model, who monitors its performance and when human intervention is required. She says validation should be continuous, particularly when AI output could influence an investigation.
Continuous monitoring extends beyond transaction surveillance. According to Laura Molloy, managing director at Gryphon Strategies, a provider of enhanced due diligence reports, many organizations now view due diligence as an ongoing process rather than a point-in-time assessment. “Historically, due diligence reviews ended when a client received a deliverable. However, now the initial deliverable is the starting point,” she explains. Organizations increasingly monitor customers, portfolio companies, and executives for new records, litigation, regulatory actions or media reports.
Cardona offers a practical framework to help financial institutions deploy AI responsibly without moving too quickly, and it’s rooted in resisting the temptation to adopt AI simply because the technology exists. “Start with a defined business problem, not with the technology,” he tells Fraud Magazine. Institutions should establish what AI is permitted to do, what information it can access and where human judgment must remain, then validate its performance and continuously measure the results. Cardona says that organizations may need to “crawl backwards” after seeing what AI can do. “Go back and make sure the fundamentals — the use case, data, controls, governance, training and measures of success — are right. The goal is to move quickly enough to learn but deliberately enough to govern,” he says.
The experience of developed and emerging markets underscores the importance of tailoring AML and anti-fraud programs to local risks and conditions. Soares says that emerging economies can provide valuable lessons on combating financial crime in rapidly digitizing environments characterized by fintech growth and informal financial systems.
Developed markets, on the other hand, offer established regulatory frameworks, technological expertise and stronger mechanisms for international cooperation. “There’s no one-size-fits-all approach,” Essien tells Fraud Magazine.
The quality of the information investigators feed into or obtain through AI also demands scrutiny. Molloy says that online aggregator sites, which republish information from original sources, may contain inaccuracies or conflated records. “We train analysts to avoid these sources and always consult the original source for the best information,” she tells Fraud Magazine. Senel similarly warns that incomplete or poor-quality information can produce unreliable findings, adding that generative AI can hallucinate or reinforce a user’s assumptions. Because incorrect responses may sound authoritative, fraud examiners must independently verify information before incorporating it into their analyses or conclusions.
Even with reliable data and appropriate oversight, AI encounters another challenge in combating financial crime: Human experts don’t always agree on the correct outcome.
Kimberly Lacey, J.D., an independent consultant and former chief AML officer at KeyBank, recalls comparing hypothetical scenarios with chief AML officers from several regional banks and finding no universal agreement about which situations warranted filing a SAR. Their decisions reflected their institutions, experiences and risk tolerances. “It really provided a great insight into how much subjectivity is, by its very nature, embedded in this risk-based framework,” Lacey says.
That subjectivity complicates efforts to train AI to reproduce AML judgments. Before automating or assisting those decisions, Lacey says, organizations must determine what constitutes a successful outcome. For example, a filed SAR, a request from law enforcement for supporting information and a prosecution represent different measures of success that could steer an AI system in different directions.
Technology has transformed financial crimes compliance from a largely manual, reactive function into a data-driven discipline capable of analyzing enormous volumes of financial activity. AI, machine learning (ML), automation and advanced analytics help institutions identify unusual patterns and relationships, allowing compliance professionals to devote more attention to investigation and decision-making.
Prior has witnessed this technological evolution from multiple sides of the financial crime spectrum. He describes a “fundamental transformation” to a “proactive, network-driven intelligence ecosystem” that has occurred across four dimensions: scale, breadth, speed and sophistication. Anti-fraud professionals can analyze millions of transactions instead of manually sampling them, identify relationships across criminal networks instead of examining suspects in isolation and use automated processes to analyze information faster than traditional searches allowed. Technology has also strengthened connections among financial institutions, regulators, intelligence agencies and law enforcement, Prior says.
“Historically, the relationship between the public and private sectors was highly restricted and transactional,” Prior tells Fraud Magazine. Financial institutions largely submitted individual reports that law enforcement reviewed separately, sometimes long after the underlying activity occurred. Technology can now support a collaborative approach in which intelligence moves across institutional boundaries.
Prior highlights Australia’s Fintel Alliance as an example of that evolution. The public-private partnership, established and administered by AUSTRAC, brings together government agencies, law enforcement and private-sector organizations to share financial intelligence and collaborate against financial crime threats. According to Prior, these types of partnerships move beyond traditional one-way reporting by allowing participants to pool information and develop a robust picture of criminal activity.
“It’s no longer uncommon for a single suspicious matter report (SMR) to AUSTRAC to list hundreds of interconnected entities and individuals, exposing large-scale criminal enterprises that previously went unnoticed or were reported individually,” he says. “Technology has fundamentally challenged the assumption that illicit wealth remains confined to tangible, fiat-based systems, forcing a pivot toward monitoring digital and intangible asset classes designed to evade law enforcement.”
The growing sophistication of financial crimes investigations has also enhanced fraud examiners’ compliance technology toolkit. According to Essien, investigators increasingly use data analytics, investigative platforms, open-source intelligence (OSINT) tools and government databases to analyze financial records, identify relationships and gather intelligence.
Technology has dramatically increased the volume of information available to fraud examiners, but more data doesn’t necessarily mean better intelligence. According to Molloy, modern investigators are often challenged by the declining quality of online information despite improvements in search capabilities. As privacy protections expand globally, investigators increasingly rely on courthouse visits, direct outreach to educational institutions and former employers, and specialized databases to verify facts. Molloy describes the current environment as one of “greater quantity but lower quality” data, requiring anti-fraud professionals to spend more time validating information rather than locating it.
Advanced analytics can help fraud examiners make sense of that growing volume of information by connecting data scattered across disparate systems. Lacey says that graph analytics, once more familiar to federal law enforcement investigators, is becoming common in the private sector. “We went from having no data to drowning in it,” Lacey says. Graph tools allow anti-fraud professionals to focus on analyzing relationships rather than trying to find them. Prior sees similar value in entity resolution, which connects disparate information involving people, accounts and other entities. Together, these capabilities are moving investigations beyond individual suspects and linear connections toward analysis of complex networks.
Compliance professionals may worry that AI will replace them. However, industry research consistently suggests the opposite. More than 85% of the respondents in SEON’s AI Reality Check 2026 report say that “AI agents should support or augment analysts,” not supplant them. As technology takes on repetitive tasks and helps investigators analyze large volumes of information, anti-fraud professionals will need to merge compliance expertise with analytical skills, technological literacy and human judgment.
Molloy witnesses that balance in due diligence and cross-border research. AI can quickly find potentially significant issues in large volumes of information while advances in translation technology allow investigators to review international media and records more efficiently. “Using AI tools, it’s possible to identify major areas of concern quickly, without having to navigate through less relevant results,” Molloy says.
But speed doesn’t guarantee accuracy or context. Molloy says AI systems can mix up people with common names, overlook information unavailable in open records and present incomplete findings with unwarranted confidence. For fraud examiners, that makes verification and expertise increasingly important. “When we’re obligated to know your customer, we can’t outsource the process to a machine,” Molloy stresses. AI may help anti-fraud professionals find and prioritize information, but people must ultimately determine whether that information is complete, accurate and meaningful.
With the promise of AI and other new technologies comes a call from regulators and policymakers for responsible innovation that emphasizes strong governance, cybersecurity protections, explainability and privacy safeguards. The Financial Action Task Force (FATF) urges financial institutions and regulators to support technology adoption through risk-based approaches, collaboration and enhanced supervisory expertise. Additionally, recent Financial Crimes Enforcement Network (FinCEN) modernization proposals and innovation initiatives direct financial institutions to use technology in ways that enhance the identification and management of financial crime risks while improving information sharing and focusing resources on the highest-priority threats.
As organizations face more complex risk environments, they expect measurable results from their technology investments. According to Kroll’s 2025 Financial Crime Report, successful financial crime compliance programs will integrate advanced technology with strong governance, regulatory awareness and skilled professionals. [See “Ten principles of responsible AI use” at the end of this article.]
To counter emerging risks, Cardona recommends that business leaders stay abreast of modern fraud schemes and be willing to adapt. “Fraud and money laundering often intersect, and understanding the underlying fraud can be critical to recognizing and investigating how illicit proceeds are moved,” he says. “This is an area where the CFE perspective and broader fraud knowledge complement traditional AML expertise.” He tells Fraud Magazine that leaders in AML should recognize the opportunities and limitations of emerging technologies. Fundamentals, such as professional skepticism, integrity and an investigative mindset, remain essential to the role of AML professionals, just as they do for fraud examiners.
Regulators are signaling a shift toward measuring financial crimes programs by their effectiveness rather than primarily by compliance processes. In April 2026, FinCEN proposed rules intended to fundamentally reform AML and countering the financing of terrorism (AML/CFT) programs under the BSA. The proposal emphasizes risk-based, reasonably designed programs while giving financial institutions greater discretion to direct resources toward their highest illicit finance risks and distinguish program design deficiencies from failures in implementation. FinCEN’s intention is to make AML/CFT supervision more consistent while reducing unnecessary compliance burdens.
Federal banking regulators are moving in a similar direction. Under a July 2026 proposal, the Board of Governors of the Federal Reserve System would require the banks it supervises to establish and maintain effective AML/CFT programs designed to identify, assess and mitigate their illicit finance risks. Banks would be expected to keep their risk assessments current as their products, services, customers, distribution channels, geographic footprints and other risk factors change, while directing greater attention and resources toward higher-risk activities.
Recent enforcement actions demonstrate the consequences when AML controls fail in practice. In October 2026, U.S. regulators imposed more than $1.2 billion in penalties on Bank of America after finding significant deficiencies in its AML and sanctions compliance programs. Authorities said the bank failed to report over $5 billion in suspicious transactions, including activity linked to national security concerns, and relied too heavily on outside parties rather than effective internal controls. As part of the settlement, the bank acknowledged BSA violations and agreed to enhanced oversight and remediation measures, and will record a non-tax-deductible charge of approximately $1.2 billion.
That focus on effectiveness applies to how institutions deploy technology. The Federal Reserve proposal encourages banks to evaluate whether ML, generative AI, digital identity, blockchain analytics, application programming interfaces and other innovations could help them combat financial crime more effectively, but it doesn’t require institutions to adopt a particular technology. The proposal also clarifies that technology may not be appropriate for every institution and encourages responsible innovation rather than technology adoption for its own sake.
That approach reflects an overall trend in financial crime compliance. LexisNexis Risk Solutions identifies effectiveness as one of the major compliance trends of 2026, noting growing expectations that institutions demonstrate measurable risk reduction, simplify controls and allocate resources strategically. The organization also identifies AI as a force affecting both sides of financial crime, requiring organizations that use it for detection to balance innovation with data governance, transparency and regulatory accountability. Industry analysis suggests that the emerging regulatory consensus isn’t to discourage AI and other technologies but to expect institutions to understand their systems, document their use, maintain effective controls and remain accountable for the results.
For financial institutions, the regulatory shift reinforces the need for financial crimes compliance programs to evolve with their risks. Although FinCEN’s rule hasn’t been finalized, institutions can evaluate whether their risk assessments accurately reflect current operations and emerging threats, whether resources align with their highest risks, and whether governance and independent testing measure how effectively their programs perform in practice.
Avoiding overreliance on technology is equally important. According to Senel, anti-fraud professionals should understand technology well enough to use it effectively without becoming dependent on it during an investigation. “If a fraud examiner isn’t aware of its limitations and trusts the technology too much, they may fail to challenge the results and could ultimately compromise the integrity of an investigation,” she says.
Cardona says that anti-fraud professionals must continue to develop their investigative reasoning skills and know when to ask questions that challenge AI-generated answers and verify information, as well as recognize when technology might be wrong. “Fraud examination reinforces professional skepticism, evidence-based analysis and investigative thinking, all of which become more important as technology produces more information for investigators to evaluate,” he says. Cardona instills this principle in his forensic accounting students at Cal Poly Pomona, an ACFE Anti-Fraud Education Partnership university, and says he believes the next generation of fraud examiners will need to combine traditional investigative principles with emerging technologies.
Senel tells Fraud Magazine that data literacy is key for developing anti-fraud professionals. “So much of what we do involves identifying, investigating and preventing financial crime through data. You need to be comfortable working with large amounts of information, identifying patterns and understanding what information you can get from the data,” she says.
Soares advises fraud examiners to be proactive rather than reactive. He says those investigating money laundering and other financial crimes often chase transactions after they’ve occurred and the money has dissipated. Technology should be used to identify suspicious patterns early and prevent funds from moving through the system. He says money laundering doesn’t always involve large amounts and is often concealed in seemingly ordinary transactions and businesses that look legitimate. This requires an approach that looks beyond transactions to understand people, their relationships and other activities in which they’re involved.
Looking ahead, Cardona says he expects that new anti-fraud roles will center on AI governance, AI validation and retrieval-augmented generation (RAG) management. He predicts hybrid roles may combine traditional financial crimes expertise with technology, data and AI skills, and that subject-matter experts will still be important for translating investigative and regulatory expectations into AI-enabled workflows. He also encourages forensic accounting students to build investigative and technical literacy, rather than viewing them as separate skill sets.
Senel says she encourages anyone interested in preventing financial crime to pursue a career in fraud, AML and forensic investigations. “Not only is the subject matter fascinating, but the constant changes in criminal methods, technology and regulations mean that you’re always learning and adapting. It never gets boring,” she tells Fraud Magazine.
She also offers advice to those new to the field. “Rather than trying to memorize all the different money laundering schemes and methodologies, try to understand the logic and motivation behind the activity. Ask questions like: ‘Which control points could have prevented this from happening?’ and ‘What other weaknesses could be exploited to facilitate financial crime’?”
According to Senel, it’s important to consider criminal methods when developing fraud prevention techniques. “Once you learn to think from the perspective of a criminal, you become much more effective at identifying and preventing financial crime. You start looking beyond what’s already happened and thinking about what could happen next, which I think is an important skill in this field.”
Continuing to develop an investigative mind along with technological expertise is key. “The professionals who can combine investigative expertise with new technology will be best positioned for what comes next,” says Cardona.
Crystal Zuzek, CFE, is associate editor of Fraud Magazine. Contact her at czuzek@acfe.com.
Anna Brahce, CFE, is assistant editor of Fraud Magazine. Contact her at abrahce@acfe.com.
As artificial intelligence (AI) adoption picks up speed, regulators and industry groups are emphasizing governance and accountability. The Financial Action Task Force’s (FATF) Horizon Scan: Artificial Intelligence and Deepfakes report, published in 2025, examines the impacts of AI on money laundering, terrorist financing and proliferation financing. These principles, distilled from the report, aim to reinforce regulators’ expectations that institutions remain responsible for compliance decisions, even when AI supports those decisions.
Unlock full access to Fraud Magazine and explore in-depth articles on the latest trends in fraud prevention and detection.
Read Time: 6 mins
Written By:
Felicia Riney, D.B.A.
Read Time: 7 mins
Written By:
Patricia A. Johnson, MBA, CFE, CPA
Read Time: 12 mins
Written By:
Roger W. Stone, CFE
Read Time: 6 mins
Written By:
Felicia Riney, D.B.A.
Read Time: 7 mins
Written By:
Patricia A. Johnson, MBA, CFE, CPA
Read Time: 12 mins
Written By:
Roger W. Stone, CFE