How AI is making legacy anti-money laundering processes more efficient

Effective and accurate data analysis is vital to effective AML/CFT programs – yet AML teams using legacy transaction monitoring software frequently deal with backlogged systems. Their analysts often suffer from fatigue from processing large amounts of alerts with lots of false positives. Without a way to sort through incoming alerts, highly qualified investigators can spend most of their workday on routine tasks like scanning overburdened systems and low-risk alerts.

Not only does this lead to frustration — it wastes company time, financial resources, and energy, overburdens employees, and makes teams more likely to miss illegal activity. It can also lead to unwanted organizational costs and losses. For example, team burnout means high turnover rates and the costs of hiring and training replacements. Poor screening can lead to fraud losses and resulting disputes.

More importantly, if a company is deemed to have inadequate risk management processes, it could face regulatory fines and legal action. In a particularly prominent case, it was one global investment firm Fined more than $1 billion in 2022 For — along with long-term fraud — “failure to implement key risk controls.”

How does AI solve the cost versus risk dilemma?

Despite these increased pressures and risks, many financial institutions fear that reforming the system will cost more. But it is actually possible to keep a company’s platform in place while simultaneously overlaying AI algorithms to enhance its capabilities. In fact, competitor companies have highlighted their reliance on artificial intelligence and machine learning (ML) as key to their success.

“Effectiveness and effectiveness are key factors in expansion. We cannot grow our team every time we grow our customer base,” explains Valentina Butera, Head of AML and AFC Operations at Holvi, a leading digital bank. In a recent interview, Andreas Braun of PricewaterhouseCoopers Luxembourg He highlighted the massive data processing and analysis made possible by artificial intelligence, helping to solve traditional cost and efficiency dilemmas of risk management.

a 2022 Report by Allied Market Research The financial technology and AI market is predicted to reach more than $61 billion by 2030. Once relegated to speculation, AI and machine learning are now practical realities – and judging by Regulatory responses around the worldIts use has become ubiquitous. Key examples include:

In our annual State of Financial Crime report, 99% of companies surveyed expect AI to positively impact financial crime risk detection. Consider the three most specific use cases for AI in transaction monitoring:

• alert priority – 31% of respondents expected AI to help arrange transaction alerts according to risk. This allows transaction monitoring teams to detect and act upon the most risky activity faster.
• Flexible Adjust – 26 percent believe they would use AI to improve their alert system – helping to set boundaries and set alerts in a responsive manner.
• Determine the relationship – 24% expected that AI would reveal new relationships between the entities and individuals being monitored.

Using artificial intelligence to improve transaction monitoring

How can AI overlay work in practice?

Consider a scenario. A senior analyst, Allison, is dealing with bloated and imprecise alert queues due to strict rules and not sorting by priority. Every day, she spends hours painstakingly working through individual alerts with no effective way to know which ones are crucial and worth her time investigating. When it encounters a high-risk alert, it has less time to research it because of time wasted scanning false positives. In fact, if the system is backlogged, alerts related to actual financial crimes may be queuing for days or weeks before they are found. The team has lost several members recently, but Alison doesn’t have time to keep up with her waiting lists and train her new teammates effectively.

Then imagine her company adding a layer of artificial intelligence to its existing system to handle alerts more intelligently. The new AI overlay combines several powerful risk management technologies, allowing it to:

• Sort alerts automatically – Artificial intelligence knows how to sort incoming alerts by risk level, identifying a high risk level for those who show the most suspicious activity. It will also constantly improve based on analyst feedback. Alison immediately begins looking at the queue for high-risk alerts when she comes to work. Meanwhile, lower risk alerts are either resolved in bulk or used to train new analysts. And when directing advanced team members, Alison can use the high stakes queue to demonstrate how to handle risk alerts.
• Enabling more efficient tuning – The AI ​​also allows the team to improve and modify the parameters and limits of the base rules. This enables more risk-responsive alerts, helping to enhance detections and reduce false positives.
• Expose more bad actors Weak evidence relating to only one person may not lead to escalation. But with a new AI overlay, Allison’s team can take advantage of the weak links in their datasets to identify and disrupt clusters of criminal activity.
• Identify the real actors who work behind the scenes Use identity groups to find hidden relationships. The team can now see connections and red flags that were previously invisible to them.
• Get greater insights and explainability About the reasons for creating the alert. Allison is more confident that she and her team can support their decisions should they be audited or receive an inquiry from their senior leadership.

With minimal initial cost, AI-enhanced transaction monitoring increases Allison’s value to the company, enabling effective, risk-based investigation while allowing her to effectively train team members. Meanwhile, layered machine learning models improve the effectiveness of her company’s AML/CFT risk detection process.

main sockets

For years, compliance teams have recognized that legacy AML programs and processes did not meet the financial crime challenges that their organizations faced. The strict rules and checkboxes might catch some egregious behavior, but they’re missing much of the complexity involved in illegal activity. Nor can they see the bigger picture and the broader connections between entities and people necessary to help law enforcement crack down on root and branch criminal behavior. The tools and techniques are now in place for banks to face this moment.

Ian Armstrong, ComplyAdvantage’s Head of Regulatory Affairs Practice, finds that many companies are already seeing success with AI, so it’s important to be agile and avoid falling behind competitors who may soon be able to operate in a more sophisticated way without comparable cost increases. Indeed, the term “artificial intelligence” is no longer just a buzzword – it is an umbrella for the many practical programs companies can implement today. Regulators around the world are aware of this and will likely soon ensure that AML regulations reflect the innovations available to companies in their jurisdictions.

AI overlay can provide a simple and cost-effective option for companies that need the benefits of AI but are not in a position to undertake an overhaul. Using overlay also involves fewer unknowns, since the algorithms do not replace but augment existing processes. With minimal disruption, companies can improve AML/CFT compliance efficiency by prioritizing AI-enhanced alerting, risk detection, and escalation – reducing related risks and costs while supporting employee retention rates and staying competitive in an ever-changing compliance landscape. .

Originally published January 30, 2023, updated January 30, 2023