Jump to: Rising alert volumes | More oversight | Conclusion

Banks and global financial service providers face increasingly complex regulatory and compliance risks. If they fail to meet Know Your Customer (KYC) and identity verification (IDV) obligations, they can face eye- watering fines and exclusion from key markets and banking systems. Consequently, these organizations are under pressure to ensure that their systems and processes can meet the growing demands placed on them.

In this paper, we examine the sanctions landscape in more detail, the impact that sanctions have had on financial services companies, and how artificial intelligence (AI) technology can help these firms achieve efficiency, accuracy and cost savings in this challenging environment.

Unprecedented sanctions

Sanctions imposed after the Russian invasion of Ukraine have continued to bite. The EU has issued its fifteenth series of sanctions against Russia, and the US has continued to expand its restrictions, including sanctioning Gazprombank and more than 50 other Russian banks, securities registrars and finance officials. In general, sanctions have been a mixture of the familiar and the new:

• Trade sanctions, in the form of import and export restrictions for certain goods, notably Russian crude oil and gold.
• Transport sanctions that deny access to airports, ports and other transport terminals for Russian goods and services.
• Financial restrictions, such as measures that block Russian banks from the Society for Worldwide Interbank Financial Telecommunications (Swift) payments network.

Perhaps the most challenging requirement concerns the monitoring of dual-use goods (DUG). These are items that have both civilian and military applications (such as advanced electronics, precision machinery and certain chemicals). In attempting to curtail Russian activity, DUG sanctions have focused specifically on areas such as industry and semiconductors. This is particularly challenging for several reasons, not least the complexity and highly componentized nature of the goods concerned. Almost all modern electronics, for example, contain semiconductors. And while it is one thing to block shipments of semiconductors in bulk, it is quite another to be able to determine which components within a shipment might contain those semiconductors.

This compounds the more traditional challenges of supply chain monitoring (i.e., tracing the origin of goods and ensuring they have not been tampered with) and, of course, the risk of financing of these lucrative but high-risk trade deals, making this particular activity a major challenge for financial institutions and market participants.

For financial institutions, the practical implications of sanctions have included increased compliance obligations and significant fines for any breaches. And with agencies in many jurisdictions increasing their enforcement capabilities, financial services firms can expect to come under greater scrutiny to prove that they know they are not doing business with a listed entity.

However, many are still failing to adequately manage their sanctions risk. In the UK, more than 100 companies have voluntarily disclosed that they have breached sanctions imposed against Russia. While actual fines are yet to be meted out, it can only be a matter of time before these types of breaches begin to cost firms dearly.

Back to top

Rising alert volumes

Financial institutions are under pressure to strengthen their sanctions systems and processes. One challenge many face is a spike in alerts, as more names and businesses are added to lists, as the restrictions themselves expand in type and jurisdiction, and as sanctioned individuals and entities try to circumvent the measures in place. Traditionally, firms have tackled increases in alert volumes either by hiring more analysts to investigate alerts, or by updating the scoring system so that it is less sensitive and alert volumes are reduced. The first option comes with significant costs, while the second reduces the quality of the output, putting firms at increased risk of breaches. 

The best approach to handling the increase in alert volumes is to focus on detection and resolving alerts using improved matching algorithms, better data quality, and auditable AI decisions – automation that reflects the FI’s unique risk views. Advanced analytics tools currently offer considerable value that does not involve significant addition to headcount. Firms can implement data cleanup and screening technology that combines their internal knowledge with the latest developments in AI and machine learning (ML). As a result, these human-led AI solutions can auto-clear false-positive alerts based on analyst expertise, while teams focus on future risks, making them more efficient and saving on costs.

Back to top

More oversight with human-driven AI

When firms adopt AI technology to support their compliance, the heavily regulated atmosphere of the financial services industry, coupled with the high risk of not complying with sanctions, means that they must emphasize transparency, accountability and auditability. AI can play a critical role in managing these issues by, for example:

• Addressing language barriers. Sanctions compliance often involves the screening of data, transactions and communications across different regions and languages. Traditional systems struggle with the complexities of multilingual content, but AI tools can prove more effective. Natural language processing (NLP) models, for example, can analyze and interpret content in multiple languages, including complex scripts such as Cyrillic, Arabic and Mandarin. They can also detect phonetic variations in names and terms across different languages, accounting for transliteration and differences in regional pronunciation. Meanwhile, AI can recognize contextual nuances, such as distinguishing between similar-sounding names or interpreting variations in terminology.
• Harmonizing differences in watchlists and formats. Global sanctions and watchlists can differ in format, content and update frequency. Managing these discrepancies is a significant challenge, but AI systems can enable normalization across factors such as formatting and conventions, providing automated updates and cross-referencing lists to ensure deduplication.
• Adapting to evolving requirements. Because of the dynamic nature of regulations, firms need adaptable compliance systems. By combining flexible rules creation with adaptable ML models, they can refine their understanding of requirements based on new patterns and data.
• Linking AI solutions to strong governance practices. Strong model risk management (MRM) remains essential. Because AI systems address language- and list-related complexities, it is even more critical that firms maintain oversight of model updates, version control and detailed documentation.

For organizations, a vital goal in implementing these systems is to use the flexibility and power of AI not only to detect risks effectively, but also to support human decision-making. With explainable, traceable AI systems, financial institutions can navigate the growing complexity of the challenges they face. 

Back to top