Skip to main content

Blog

Financial Crime Risk Management and Anti-Money Laundering Conference

Risk managers and compliance officers reviewing financial data in a modern boardroom at dawn.

The changing financial crime landscape

Financial crime is growing more sophisticated, more coordinated and more global. Global illicit financial activity is estimated at $4.4 trillion in 2025, with approximately $750 billion flowing through Europe. Threat actors are deploying AI-generated deepfakes, synthetic identities and social engineering at scale. Instant payments have reached 80% global adoption, improving speed and liquidity but also exposing vulnerabilities that fraudsters exploit in real time.

Traditional siloed approaches to fraud, AML and operational risk are no longer sufficient. Banks and financial institutions must now integrate financial crime signals across typologies—from fraud and sanctions to money laundering and counter-terrorism financing—to detect patterns and act faster. At the same time, regulators are centralizing oversight and raising expectations for governance, data quality and human accountability.

AMLA and the new European rulebook

The Anti-Money Laundering Regulation (AMLR) will apply directly across all 27 EU Member States from July 10, 2027, replacing the patchwork of national laws. This marks a watershed moment: for the first time, Europe will operate under a single rulebook enforced by a centralized supervisor.

The Risk Management Summit 2027 on 25–27 May in Amsterdam features sessions on AMLA, DORA, NIS2 and the EU AI Act in practice. The summit brings together senior risk, compliance and audit leaders to discuss how supervisory expectations are shifting and what boards must prove, not just promise.

Customer due diligence and lower thresholds

AMLR substantially harmonises customer due diligence (CDD) requirements, removing discretion that Member States previously exercised. The regulation lowers the transaction threshold triggering CDD obligations from EUR 15,000 to EUR 10,000 for occasional transactions and introduces a requirement for limited CDD measures for occasional cash transactions of at least EUR 3,000. Enhanced due diligence now applies to cross-border correspondent relationships involving crypto-assets and relationships involving assets worth at least EUR 5,000,000 for customers with total assets of at least EUR 50,000,000. The definition of politically exposed persons (PEPs) has broadened to include heads of regional and local authorities, with post-office EDD periods of 12 months as a minimum.

Scope expansion and new obliged entities

AMLR brings within the AML/CFT perimeter professional football clubs and agents, crowdfunding service providers, investment-migration operators, luxury-goods traders, dealers in high-value vehicles, and boats and aircraft. This reflects recognition that money-laundering risks extend beyond traditional financial services and that regulatory nets must widen to address evolving criminal typologies.

AMLA's direct supervision and Joint Supervisory Teams

AMLA will begin directly supervising up to 40 selected financial institutions that operate across at least six Member States and present high money-laundering and terrorist-financing risk profiles, with the first selection process taking place in 2027. Joint Supervisory Teams—comprising AMLA staff and Member State competent authority representatives—will serve as the operational units for each selected entity.

Beyond direct supervision, AMLA sets standards and supervises indirectly. For most firms, the supervisory relationship will remain national but increasingly follow AMLA's methodology. AMLA is already issuing technical standards at pace: in the first half of 2026 alone, it consulted on Regulatory Technical Standards covering customer due diligence, business-wide risk assessment, group-wide policies, enforcement and ongoing monitoring guidelines.

Agentic AI: control and oversight

Autonomous AI agents are emerging in banking to handle fraud detection, transaction monitoring, AML screening and investigation at scale. These systems can process vast transaction volumes in real time, spot patterns humans would miss, and make recommendations faster than traditional rules-based engines. But they also pose a governance challenge: how can banks maintain control, human accountability and regulatory transparency when decisions move at machine speed?

The summit's opening keynote, 'The 2027–29 Supervisory Reset: What Banks Must Prove, Not Just Promise', delivered by the Chief Economist of ESMA, focuses on AI-driven market activity and new sources of systemic vulnerability. The CRO keynote, 'Risk Appetite in the Age of Autonomous Banking', addresses how to define board-level risk appetite when AI agents and automated decisions move faster than governance.

A frontier session on 'Agentic AI Before It Touches a Decision' explores mapping agent permissions, tool use, data access and multi-agent chains into a bank control framework. Key themes include kill switches that stop runaway agents, supervision mechanisms that detect drift and confabulation, and explainability standards that allow audit trails to be built and reviewed. Institutions must also consider the EU Artificial Intelligence Act, which imposes governance, transparency and human-oversight obligations on high-risk AI systems used in banking.

Fraud detection in real time

Fraud is evolving faster than detection systems can keep pace. Criminals are deploying generative AI to craft deepfake documents, synthetic identities and convincing social engineering attacks. Instant payments have accelerated the speed at which fraud must be detected: a transaction that clears in seconds leaves no room for manual investigation. Traditional batch-mode transaction monitoring, designed for overnight processing, is becoming obsolete.

Leading institutions are deploying multi-layered fraud detection: fast statistical checks that run on every transaction to flag anomalies, machine-learning models trained on historical fraud patterns, and agentic AI systems that reason over transaction context, merchant risk, device fingerprints and customer behavior. The best systems blend these layers to catch fraud that any single approach would miss. They also require closed-loop feedback: analyst decisions must be captured, labelled and fed back into model retraining so systems improve continuously.

A critical challenge is false positives. Fraud teams are overwhelmed by alerts they cannot investigate. Reducing noise while improving recall requires tuning thresholds to reflect the real cost of a missed fraud versus a false decline, not arbitrary detection metrics. This demands collaboration: banks working with law enforcement, payment networks, telcos and platforms to share intelligence about organized scam networks and mule-account operations across borders.

Beyond AML: sanctions and typologies

Financial crime is not monolithic. Sanctions evasion, fraud, money laundering, counter-terrorism financing, corruption, market abuse and insider threats follow different patterns, use different tools and require different investigative approaches. Yet they often intersect: a sanctions evasion scheme may involve shell companies, trade-based money laundering, corrupt officials and fronted payment instructions.

AMLR embeds sanctions screening formally into customer due diligence, requiring obliged entities to verify whether customers or beneficial owners are subject to targeted financial sanctions. This converges AML and sanctions compliance, but it also expands the scope: institutions must track expanding sanctions regimes across multiple jurisdictions, understand circumvention tactics, and integrate screening into real-time transaction monitoring.

Crypto-asset service providers face particular challenges. They must apply customer due diligence at both sides of a transaction, screen against sanctions lists and beneficial-ownership registers, and monitor for typologies that traditional banking screens miss: mixing services, decentralized finance, and cross-chain bridging. The speed of blockchain transactions leaves little margin for error.

Supervision and governance at the board

AMLA marks a cultural shift: from box-ticking compliance to prudential-style supervision, where financial crime is treated as a first-order risk alongside credit, market and operational risk. AMLA will apply risk-based, data-driven, and intelligence-led supervision, supported by harmonized methodologies and stronger group accountability.

Boards are expected to take an active role. This means governance must embed the AMLA frameworks into strategic direction, strengthen accountability, and ensure institutions can explain both the design of controls and the logic behind how attention, resources and escalation are prioritized across the business. Supervisors will scrutinize the quality and consistency of customer due diligence, the management of cross-border and high-risk exposures, the effectiveness of transaction monitoring, and the integrity of data lineage and reporting.

For institutions likely to fall within AMLA's direct supervisory perimeter, the pressure is acute. They should conduct mock supervisory assessments, test whether reporting, management information and governance documentation are fit for scrutiny, and ensure that senior management can articulate how risk decisions are made and enforced across geographies and business lines.

Preparing your institution

Readiness assessments should begin immediately, even though many AMLA provisions do not take effect until July 2027 or later. Key priorities include:

  • Enterprise-wide risk assessment. Review and redesign risk methodologies to align with AMLA's framework. Assess inherent risk (from customer, product, geography, distribution channel and transaction type), control effectiveness, and residual risk. Document assumptions and benchmarking.
  • Customer due diligence. Audit CDD policies and processes against the harmonized AMLR standards. Test whether thresholds are correctly implemented, whether beneficial-ownership verification is adequate, and whether sanctions screening is embedded in real time.
  • Group-wide policies. Develop and implement group-wide AML/CFT policies, procedures and controls that apply consistently across all branches and subsidiaries, including those in third countries. Where local law falls short, apply EU standards and document mitigating measures.
  • Data governance. Many institutions need significant remediation of AML data architecture, aggregation capabilities and management information. AMLA is developing an EU-wide reporting ecosystem that will require standardized, high-quality and timely data. Begin data integration work now.
  • AI governance. If deploying AI-driven transaction monitoring, fraud detection or AML screening, develop governance frameworks that ensure human oversight, explainability and auditability. Align with the EU Artificial Intelligence Act.
  • Supervisory readiness. For institutions likely to be directly supervised by AMLA, prepare for more intensive, rigorous oversight. Ensure management information systems can support rapid response to supervisory requests.

Institutions that invest early in harmonized governance, integrated data capabilities and mature risk assessment frameworks will be better positioned to comply with the new rules and improve operational efficiency, strengthen risk detection and enhance decision-making.

Frequently asked questions

What happens if my institution falls within AMLA's direct supervision?

AMLA will select up to 40 high-risk, cross-border financial institutions for direct supervision from 2028. If selected, your institution will be supervised by a Joint Supervisory Team comprising AMLA and national authority staff. This means more intensive, frequent oversight, alignment with AMLA's common supervisory methodology, and expectations that are likely to be more rigorous than those of national supervisors alone. Prepare by conducting a mock supervisory assessment and ensuring your governance, data and reporting are audit-ready.

When must my group implement group-wide AML/CFT policies?

AMLR requires group-wide policies, procedures and controls to apply from July 10, 2027. These must cover all branches and subsidiaries, including those established outside the EU. Where third-country law falls short of EU standards, your group must apply EU standards to the extent local law allows. If local law prevents this, you must take additional measures to manage the risk and inform your supervisor. Document all assumptions and escalate any gaps to senior management and the board.

How should we approach AI governance for transaction monitoring and fraud detection?

Start by identifying whether your AI systems fall within the scope of the EU Artificial Intelligence Act. High-risk AI systems must have documented governance frameworks that ensure human oversight, explainability and auditability. Build kill-switch mechanisms to stop runaway agents, implement drift detection to flag model degradation, and ensure that decisions can be traced back to the data, features and reasoning that led to them. Engage with your supervisors early to ensure alignment with their expectations.

Where can I network with peers on AMLA implementation and financial crime trends?

The Risk Management Summit 2027, held 25–27 May in Amsterdam, brings together senior risk, compliance and audit leaders from banks, financial institutions and regulatory bodies. See the tickets page for passes and pricing. The summit features 30 speakers delivering 33 sessions on AMLA, agentic AI, supervision, fraud, credit, capital, operational resilience and emerging risks. Speaker bios and session details are available on the speakers page.

Sources

  1. Risk Management Summit 2027, 25–27 May 2027, Amsterdam
  2. Schedule – Risk Management Summit 2027, Amsterdam
  3. One Rulebook, One Supervisor: How the EU's New AML Regime Will Reshape Compliance for Firms
  4. How EU AMLA is transforming AML supervision in Europe
  5. Risk Ready EMEA Summit – Tackling Fraud & Financial Crime

Share this article

Blog