John Flowers serves as the Global Head of Financial Markets at eClerx. With over 30 years of experience in the financial technology services sector, he has held various executive roles on both the business’s technology and client-facing sides.
Discover top fintech news and events!
Subscribe to FinTech Weekly's newsletter
Read by executives at JP Morgan, Coinbase, Blackrock, Klarna and more
Asymmetric risk poses a constant threat to banks, fintechs, and other heavily regulated businesses. An incomplete due diligence review on a single customer that misses their involvement in money laundering or other crimes can lead to multimillion-dollar fines, reputational damage and regulatory action at the highest levels of leadership. Because even small errors can produce these outsized consequences, eliminating small gaps in know-your-customer (KYC) processes is essential to protecting both institutions and their stakeholders.
Traditionally, effective KYC and anti-money laundering (AML) compliance has required a comprehensive evaluation of customer risk during onboarding, followed by scheduled monitoring for changes in risk profile or behavior, often through exceptionally manual processes that are prone to delay. Now, AI and automation make it possible to strengthen KYC and enhance AML oversight by using real-time data and enabling a more proactive approach to financial crime prevention.
What are AI’s roles in KYC/AML risk reduction?
Operational errors and penalties are happening despite banks’ substantial investment in AML/KYC processes and solutions. Juniper Research put 2024 global KYC spending at $30.8 billion last year. Yet many institutions still rely on manual processing and updating of customer data, which slows down onboarding and delays updates that could flag changes in risk profile.
Automating some of these processes using rules-based robotic process automation (RPA) can speed things up, but may generate high rates of false positives that require more time for manual reviews. Meanwhile, criminals are using advanced technology to avoid getting caught by KYC and AML processes. With AI and stolen or false identity data, they can create documents and histories that look real enough to fool analysts and basic automated systems.
Adding AI-enabled automation and GenAI to RPA can help banks address these challenges in multiple ways.
1. Customer onboarding experience
As part of the KYC process, firms provide new customers with a list of required documents and data they cannot verify independently. When these requirements are not communicated effectively, it can confuse customers and delay approvals. This is especially true when the requested information does not clearly align with the specific regulatory requirements of the jurisdiction(s), creating extra work for analysts who must then resolve the discrepancies.
With an AI natural language processing model embedded in the onboarding process, banks can communicate effectively and request the appropriate information based on specific regulations of the applicable jurisdictions. The result is a faster onboarding process that’s less prone to errors caused by someone checking the wrong box or submitting documents that don’t correspond to local and internal requirements. This can stop data gaps and errors before they enter the system.
2. Detecting identity fraud
AI-powered computer vision and synthetic identity detection models can flag customers whose documents or financial histories appear to be fake or stolen, even if they look legitimate to human analysts. These tools synthesize data from multiple sources over time, and they can see connections among the data which humans would miss, and traditional rules engines cannot decipher. They quickly correlate a customer identity with real-world activity and raise flags when discrepancies appear so analysts can investigate.
3. Real-time KYC and AML monitoring
Maintaining customer data after onboarding is a never-ending process. Monitoring customer activities with the institution, scanning for adverse news about them, and understanding any changes in their business networks is critical to avoid missing signs of a shift in customer’s risk profile. GenAI models can orchestrate this kind of monitoring in real-time by ingesting data from multiple platforms and data sources, setting a baseline risk profile for each customer, and raising alerts when new data indicates a risk profile change.
4. Compliance and reporting
Comprehensive onboarding and monitoring solutions also give banks the data insights they need to assess AML compliance, identify areas for improvement, and generate reports for internal stakeholders and regulators. GenAI reporting solutions are not limited to ingesting massive amounts of data and answering questions. They also can be taught to display the processed information using intuitive graphs and charts, on dashboards, and in reports. This visibility lets bank leadership identify and stop emerging issues before they become major problems.
5. Adapting to technology and regulatory changes
GenAI and AI-enabled automation systems learn from their inputs. That means they can be trained to adapt when banks connect new data sources and technology platforms, without requiring a major replatforming or a lengthy integration process. This allows institutions to derive more value from their AI investments over time.
AI’s learning capacity also makes it easier for banks to update their requirements when regulations change. Training and testing AI KYC models on new guidelines typically takes less time than manually updating non-AI platforms. It’s also faster than training analysts on new guidelines. AI can actually help with this training as well, by answering simple questions or summarizing the changes in easy-to-read formats. Analysts can quickly have the current information they need to consistently follow and enforce new policies.
Reducing asymmetric risk for KYC/AML with AI
AI-powered KYC and AML tools represent the future of financial risk management. They can sharply limit banks’ exposure to asymmetrical risks today and also adapt to evolving technological and regulatory environments to safeguard against future threats. With regulators increasingly scrutinizing the role of financial institutions in international crime, and criminals growing more adept at evading traditional KYC and AML controls, integrating AI into KYC and AML workflows is the most effective way for Institutions to strengthen protection now and into the future.
