Moderated by Patrick DeVilbiss, Senior Offering Manager at CGI, the panel combined banking and technological perspectives in exploring how red flags in trade finance and supply chains are evolving, and how institutions can better detect and disrupt TBML. 

TBML refers to the process of disguising illicit funds by moving them through seemingly legitimate trade transactions, in order to frame the money as legally earned, as outlined in a 2020 report by the Financial Action Task Force. 

The implications of TBML are extremely grave, with it playing an integral role in financing serious organised crime networks, including drug smuggling, terrorist financing, and human trafficking. 

“If you can stop the flow of money, it is a shot through the heart to these organised crimes,” said Nigel Hook, Founder and CEO of Tradesun, referring to his own experience working with the FBI on combating human trafficking. 

Identifying red flags 

Given the urgency of this type of fraud, it becomes crucial to implement systems that identify the red flags—the early signs that a transaction might be suspicious—within a trade deal. Panellist Mansour Davarian, Head of Transaction Banking Solutions at Lloyds Banking Group, differentiates red flags into two types: complex and non-complex. 

Non-complex red flags might include urgent funding requests from unfamiliar clients, inconsistent documentation, and unusual payment terms. More complex red flags involve over- or under-invoicing anomalies, readjustments to shipping documents, rerouting goods to exploit tariff advantages, or using intermediaries to obscure the true parties involved. 

These expansive mechanisms present a complicated challenge. Institutions need to go beyond spotting one-off anomalies and begin recognising patterns across disjointed systems.

This process is simpler for documentary trade, where payments are made only when trade documents are verified. In open accounts and supply chains, “there is a level of reliance placed on our counterparties,” said Davarian. This reliance complicates the ability to spot red flags, making due diligence and a deep understanding of who the counterparties are essential. 

AI and the human factor 

As TBML schemes grow more sophisticated and begin utilising advanced technologies themselves, financial institutions face mounting pressure to integrate artificial intelligence (AI) into their risk management strategies. 

AI offers the potential for real-time transaction monitoring (to detect anomalies as they occur), advanced pattern recognition across large datasets, and automated risk scoring through predictive analytics. 

“One of the areas where AI will be very useful, probably beyond a human, is looking at fake documents and fake transactions,” said Hook. “There’s a premise that maybe AI can detect them better than a human, if AI has created them.” 

The World Economic Forum (WEF) underscores the importance of human-machine cooperation. It highlights the relationship between humans and AI as a “true partnership,” where AI models can operate with human oversight and testing, getting trained and evaluated to apply insight in a more efficient, thorough manner. 

However, it is crucial to recognise that technological advancements can sometimes be unreliable for recognising complex systems of TBML. Such laundering networks require interpreting context, intent, and behaviour across a large network, making human judgment essential. 

DeVilbiss highlighted the importance of experience in spotting red flags, noting that much of it relies on “your knowledge, your ability to take that gut check of: ‘Well, it just looks a little bit fishy, a little bit off, and I’m going to dig in just a touch further.’” 

Network complexity and data-sharing

TBML risk often lies in relational networks, where intricate dynamics across individuals determine a deal, not just paperwork. Thereby, financial crime detection becomes reliant on the relationships between customers, counterparties, and intermediaries—creating what DeVilbiss referred to as the “network perspective.” 

“You need to know not just your customers, but your customers’ customers. If you can go even further down the line, fantastic,” said DeVilbiss. 

Understanding networks requires contextual judgment. Criminal networks are highly adaptive and use complex structures that obscure ownership and control. Especially in the absence of veteran red flag-spotters, employing the rich transaction data held by banks becomes essential. 

Mariya George, CEO and Co-founder of Cleareye.ai, emphasised that modern TBML detection requires moving beyond traditional transaction-level analysis to embrace network-based approaches and advanced technology integration. To this end, data integration is a game-changer.

“If you combine your historic transaction data with all the external data sources out there – market trends, competitive analysis, geopolitical changes, sanctions – it’s just a matter of time before we can do red flag checks live on every single transaction,” said George. 

Through this combination, institutions can expand their understanding of TBML data in relation to market trends, competitive analysis, and geopolitical shifts—ultimately generating better insights and becoming more vigilant when it comes to recognising red flags. 

In a report produced by the Wolfsberg Group, an association of 12 global banks, it was recommended that governments, financial institutions, trade bodies, and trade logistics providers work together to combat financial crime. The recommendation highlights how wide exchange of data information, carried out through clear protocols, is essential to this collaborative process. 

The training and enforcement gap 

However, despite these calls for coordination, persistent gaps remain in both training and enforcement that undermine the effectiveness of TBML detection efforts.

On the training front, George highlighted a critical disparity in sophistication levels: “Some of the banks are still doing it the conservative way, but the criminals have gone past. They are at a different level of maturity. They’ve figured out how to get things done in a much more nuanced way. The training levels have to change at the banks.”

This training gap, as DeVilbiss identified, is compounded by the ongoing exodus of experienced trade finance professionals who possess the institutional knowledge and intuitive understanding necessary to spot subtle red flags.

Simultaneously, a significant enforcement gap persists, where financial institutions bear the burden of detection while regulatory and prosecutorial bodies often fail to act on their findings.

According to Thompson Reuters, in 2023 a record 4.6 million suspicious activity reports (SARs) were filed to the Financial Crimes Enforcement Network (FinCen). Hook highlighted how of the activity reports, only 4% received any attention from federal law enforcement. Only 0.3% of these reports led to arrests, and a mere 0.01% resulted in convictions. 

This inaction, particularly regarding an issue that enables serious human rights violations, raises important concerns about the efficacy of current anti-money laundering (AML) strategies.

This dual gap – inadequate training to match criminal sophistication and insufficient enforcement follow-through – creates a particularly concerning dynamic. Financial institutions are caught between increasingly sophisticated criminal networks and enforcement agencies that appear overwhelmed or unable to act effectively on the intelligence provided.

—-

The panelists emphasised how red flag detection can be significantly enhanced through combining human experience, AI technology, institutional cooperation, and data-sharing; BAFT is currently working on an updated red flags and TBML guidance document scheduled for release in early autumn of this year.

However, this enforcement gap suggests that even when red flags are accurately detected, criminal activity often goes unpunished. This may be due to a variety of reasons, such as fragmented systems, where there is weak coordination between regulators and law enforcement, or limited prosecutorial focus, with volume overload leading law enforcement to prioritise high-profile cases. 

What is concerning is that this undermines both the deterrent effect of compliance systems and the broader credibility of financial crime enforcement, which ultimately enables more money laundering and the severe criminal activity that it funds.