AMLA, cost pressure, and the next chapter of AML effectiveness

A perspective informed by conversations with senior FinCrime leaders across leading banks in the Netherlands and Belgium

AML is at a turning point. Expectations around impact continue to rise, while the operating environment becomes more complex, marked by sustained cost pressure and a gradual shift from human-led processes to technology-driven capabilities. Across the financial sector, there is a growing recognition that the current approach to AML is reaching its limits.

Drawing on recent conversations with C-level and senior FinCrime executives from leading banks in the Netherlands and Belgium, a consistent narrative emerges. There is clear acknowledgement of the structural pressure the system is under, combined with a pragmatic focus on where meaningful progress is still possible.

A perspective informed by conversations with senior FinCrime leaders across leading banks in the Netherlands and Belgium

The inspiration for these conversations came directly from Discai’s whitepaper Trusted AI in AML: a blueprint for success, which sparked an important dialogue about the challenges, trends and opportunities in financial crime prevention.

One message stands out: the next step in AML effectiveness will not come from incremental optimisation alone. It will require more fundamental choices about what “good” looks like, whether existing system designs and operating models remain fit for purpose, how proportionality is applied in practice, and how technology and governance are organised to deliver demonstrable outcomes.

Four quick takeaways from the discussion

AMLA is seen as a potential catalyst for consistency, but success depends on truly risk based proportionality.
The combination of rising expectations and cost pressure is pushing banks to prioritise signal quality and freeing up specialised capacity where it matters most.
The technology direction is broadly aligned. The harder questions are ownership, hosting, governance, data quality and resilience.
Without explicit trade-offs across impact, privacy, and harmonisation, AML risks are becoming more complex without improving outcomes.

The deeper story

1. AMLA as catalyst: optimism, with open questions

Participants widely recognised AMLA as a potentially transformative force, particularly in driving greater supervisory consistency and improved cross-border cooperation across Europe. At the same time, there was a shared sense that AMLA’s impact will depend less on intent and more on how expectations are translated into practical, proportional supervision.

Key perspectives raised:

Regulatory pressure is intensifying, with expectations expected to rise further as AMLA becomes fully operational.
Several participants pointed to a tendency toward breadth in data expectations (“collect everything”) and stressed the importance of clarity on relevance and context to keep risk-based decisioning workable at scale.
There were signals that supervisors are increasingly interested in overall effectiveness, such as the relevance and usefulness of reported cases, not only procedural completeness.
Participants noted that supervisory outcomes still depend heavily on local supervisory culture, including how proportionality is interpreted and whether institutions feel able to challenge requirements that add cost without clear impact.

Thought leadership implication

This is the window to engage with local regulators and AMLA. If the sector wants proportional, risk-based expectations that scale, it needs to help shape them now, before practices harden into requirements that are costly to reverse.

2. Staying effective amid complexity, cost pressure, and disruption

The ecosystem banks operate in is becoming harder to manage: professionalised criminal networks, accelerating digitalisation, geopolitical uncertainty, and sustained public scrutiny. Against that backdrop, banks are expected to reduce costs while delivering better outcomes. Many participants described this combination as structurally difficult under current operating models.

A recurring observation was the gap between societal expectations (= catching financial criminals) and the legal mandate (= identifying and reporting suspicion) of banks. That disconnect continues to distort priorities and resource allocation.

Despite the weight of the environment, the discussion was pragmatic about where progress is most achievable in the near term:

Increase signal quality and reduce clutter, so analysts spend time on what matters.
Free up specialist capacity by improving triage, workflow and automation where it genuinely reduces friction.
Avoid chasing narratives that sound transformative but are not yet mature enough to deliver value at scale.

Participants also pointed to visible progress in parts of the industry: more mature typology thinking, better collaboration between functions, and a growing focus on measurable outcomes. The challenge is scaling these improvements without defaulting to volume-driven controls.

Thought leadership implication

The strategic shift is from output to outcomes. Banks that continue to optimise for volume will struggle to reconcile cost pressure with rising expectations. Those that optimise for relevance and capacity will be better positioned to demonstrate impact.

3. Technology direction is clear. Governance is the real debate.

Across differences in maturity, participants showed strong alignment on the technology capabilities most likely to drive the next step in AML effectiveness: contextual analytics, network-based insights, explainable machine learning, and automation across the end-to-end lifecycle.

Where views diverged was not on what technology is needed, but on how it should be organised and governed:

Ownership: should future FinCrime intelligence platforms sit solely within financial institutions, within the public domain, or in hybrid models involving AMLA and FIUs?
Hosting and sovereignty: in an increasingly geopolitical world, data sovereignty and quality, cloud concentration risk, cross-border access, and vendor lock-in remain unresolved design questions.
Trust and resilience: participants stressed the need for transparent, well-governed, auditable architectures aligned with resilience principles such as DORA. Not experimental or opaque deployments.

A consistent warning emerged: technology is a necessary enabler, but not a solution in itself. Without high-quality, contextual data and strong governance, AI can amplify complexity rather than reduce it.

Thought leadership implication

The next competitive differentiator is not “AI adoption”. It is credible governance around data, models, operational controls, and resilience. That is what will determine whether advanced analytics becomes an accelerator or an audit risk.

4. The trade-offs that must be made explicit

While not always stated directly, the discussion surfaced tensions that increasingly define the AML domain:

Compliance obligations versus meaningful crime-fighting impact
Privacy protection versus effective information sharing
European harmonisation versus national supervisory realities

Participants recognised that without explicit choices along these axes, AML risks becoming more complex and costly, while societal impact remains hard to evidence.

Thought leadership implication

Effectiveness is a governance question as much as a regulatory or technology question. Boards and senior leadership need to define where proportionality sits, what risk is accepted, and what outcomes matter, then align operating models accordingly.

Conclusion: momentum, but the need to redefine success

The conversation revealed a strong willingness to rethink roles, challenge long-held assumptions and engage more openly with supervisors. There is momentum. But there is also a growing recognition that incremental optimisation will not be sufficient.

Progress will depend on redefining success: shifting from procedural completeness toward demonstrable effectiveness, and from volume-driven output toward meaningful outcomes. For senior leadership, this requires clearer choices around proportionality, accepted risk and long-term investment. Choices that are supported by evidence, not promises.

Ultimately, improving AML effectiveness is a system design and governance challenge that requires alignment across institutions, supervisors, and the wider public domain.