Some of Asia’s most significant financial crime cases in recent years show how difficult it is to understand illicit activity by looking at individual transactions or institutions in isolation.

Singapore’s landmark S$3 billion ($2.34 billion) money laundering case exposed an extensive network spanning shell companies, nominee arrangements, luxury assets, and cross-border fund flows; enforcement action is still unfolding almost two years on. In Hong Kong, the sentencing of a former finance director and company secretary for laundering billions of Hong Kong dollars through company accounts shows how criminals exploit legitimate corporate structures to move funds illicitly. These cases are unconnected, but they expose the same vulnerability. Financial crime networks are increasingly interconnected, while institutions still work with fragmented information.

For Hong Kong, a major international financial center with a financial ecosystem spanning banking, capital markets, private wealth, cross-border payments, trade and digital assets, a risk exists, while the information needed to understand it sits in various systems, institutions, and jurisdictions.

Hong Kong’s anti-money-laundering and counterterrorism financing framework is well established, and customer due diligence, sanctions screening, transaction monitoring and suspicious transaction reporting remain essential. But the environment around these controls has changed. Customer information, beneficial ownership data, transaction histories, and risk assessments often sit across different systems, so a compliance team can hold vast amounts of information and still lack a holistic view of the customer or the network around them. This pressure extends beyond banking too, with insurers bracing for tighter scrutiny.

A transaction that appears low risk on its own can look very different once the receiving company’s ownership structure, connected accounts, previous activity or relationships with entities abroad are considered. Institutions therefore need technology that connects signals. For a city defined by cross-border activity, the question is whether institutions can connect information quickly enough to understand the risk before the transaction completes.

The same issue applies to fraud and money laundering, traditionally handled by separate teams, systems, and processes. That is gradually changing as the underlying activity increasingly overlaps. A compromised account can become a vehicle for moving funds illicitly, and a mule account can facilitate fraud while forming part of a wider laundering network. An innocuous account may turn out to carry a fraud signal, and an isolated fraud alert may prove part of a much broader network.

When investigated separately, teams see only fragments of the same network. Bringing fraud and anti-money laundering signals together, known as FRAML, gives investigators a bird’s-eye view, reveals connections earlier, and makes it easier to identify important alerts.

A name appearing across a sanctions list, internal records and corporate filings may not be an exact match. A company may sit at the center of several apparently unrelated transactions. Individually, these tell an investigator little; connected, they provide a much clearer picture of risk. This is the value of entity resolution, network analysis and unified intelligence; not more information, but more context for better decisions.

The Hong Kong Monetary Authority made this clear in June by publishing guidance supporting artificial intelligence adoption in fighting financial crime. Following consultation with authorized institutions, it highlighted use cases including continuously updating customer risk ratings, facial-recognition screening and spotting mule bank accounts used to move funds illicitly, and noted that more institutions are pursuing holistic approaches to monitoring suspicious activity. The next question is how AI can improve the quality of financial-crime decisions.

Static rules-based systems remain important, but they are limited by the scenarios and thresholds that define them, while mule networks change and criminals adapt. AI and advanced analytics can identify patterns rules struggle to capture, prioritize higher-risk activity and connect signals across larger datasets. That’s not to say technology should replace human judgment. The strongest applications will help investigators understand why activity is unusual, what is connected to it and where to look next. The goal is not automation for its own sake, but better intelligence at the point of decision.

As AI becomes embedded in daily risk processes, trust will be the measure of success. Regulators, boards, and compliance teams need to be able to interpret and explain what an AI system flags, because a risk score that cannot be justified is difficult to challenge, govern or act upon. Adoption therefore depends on explainability, data quality, model governance, and human oversight. No AI-generated alert should reach a final decision without a compliance officer being able to review and override it.

AI must not be treated as a stand-alone deployment. Even a sophisticated model has limited value if its data is fragmented or investigators can’t incorporate its output into their workflow. Firms should start with a defined financial-crime problem, test whether technology improves detection or investigation, validate results, then expand. This matters in Hong Kong, where commercial activity often involves cross-border transactions, complex corporate structures, and multiple data sources.

The ability to share and connect intelligence is becoming essential, with the aim of strengthening risk efforts without adding customer friction. Hong Kong already has a strong foundation for cooperation between regulators, law enforcement, and the financial sector. The next opportunity is making better use of the data and insights institutions already possess. That doesn’t mean sharing everything indiscriminately but identifying the signals that matter and using intelligence responsibly within appropriate legal and governance frameworks.

Strong financial-crime controls support confidence in the market and influence how institutions manage correspondent relationships, serve international clients, and participate in emerging areas of finance. As Hong Kong continues to develop its fintech and digital-asset ecosystem, that confidence could become a driver of growth. The challenge is no longer how to access information; it’s how to turn it into actionable intelligence.

Institutions best positioned for fighting financial crime will look beyond individual transactions, understand the connections and context behind them, and combine technology, trusted data and human judgment to make better, more-informed risk decisions.

The author is a senior director at LexisNexis Risk Solutions.

The views do not necessarily reflect those of China Daily.