For decades, anti-money laundering (AML) monitoring has been built on rigid, rules-based systems. These solutions were designed to meet regulatory requirements by triggering alerts whenever a transaction crossed a pre-set threshold or matched a predefined scenario. They were easy to configure, easy to explain, and, at least on paper, made institutions “compliant.”
But simplicity has a price. Those static systems have generated mountains of false positives, placed crushing operational demands on compliance teams, and failed to adapt to new products, payment rails, and criminal techniques. Regulators, financial institutions, and technology providers all now recognize the same reality: legacy rules-based monitoring is no longer fit for purpose.
The crossroads: toward risk-based, contextual, dynamic monitoring
The industry is at a crossroads. The future of AML is risk-based, contextual, and dynamic. Yet as attractive as that sounds, it does not mean throwing artificial intelligence or machine learning at the problem and hoping for a miracle. The transition requires a much more profound rethink of how compliance obligations are operationalized — and what role technology should play in balancing regulatory certainty with risk intelligence.
Regulatory rules vs. decision-engine rules
When we discuss moving away from “rules,” we need to be clear about what kind of rules we are referring to. Regulatory rules are the immovable obligations defined by law, including conducting customer due diligence, monitoring transactions, reporting suspicious activity, following the Travel Rule, and verifying beneficial ownership. These are not going away. If anything, regulators are strengthening them, not relaxing them.
Decision-engine rules, by contrast, are the scenarios and thresholds that financial institutions configure in their monitoring systems to comply with those obligations — “trigger an alert for any wire transfer above $10,000” or “flag three cash deposits over $5,000 within seven days.” It is this second category that is being rethought. The shift to risk-based monitoring does not mean ignoring regulatory rules. It means moving beyond the rigidity of legacy Regtech solutions that rely solely on static scenarios.
Tool sprawl and fragmentation in AML stacks
Many institutions, faced with this rigidity, have tried to compensate by layering systems on top of each other — core banking modules, third-party Regtech tools, and sometimes bespoke in-house scripts. The intention is to fill the gaps, but the result is often fragmentation. Data becomes scattered across silos, alerts are generated inconsistently, and investigators must manually reconcile outputs from multiple systems.
Governance becomes harder, not easier, as institutions struggle with duplicated scenarios and inconsistent logic. Costs escalate as more tools are bolted on, but effectiveness does not improve. The very workarounds designed to overcome rigidity end up creating new risks.
Orchestration as the antidote
The antidote to this fragmentation is orchestration. Rather than bolting on more tools, institutions are increasingly recognizing the need to integrate customer risk assessment, transaction monitoring, and screening into coordinated, policy-driven workflows. Industry analysts, including Gartner, highlight orchestration as a crucial capability for RegTech platforms as it enables institutions to replace siloed systems with a governed, unified monitoring layer. Orchestration, in this sense, is not just about efficiency; it is about creating the conditions for accurate risk-based monitoring.
Why rules-only systems fail
The shortcomings of rules-only systems are well known. Static rules generate false positives at an industrial scale — by some estimates, over 90 percent of alerts are false positives, forcing teams to spend vast amounts of time investigating activity that turns out to be legitimate. Rules can only catch what they were designed for, which means they often miss mule networks, rapid account switching, or layering through crypto assets.
They perpetuate silos by separating fraud and compliance monitoring, rather than delivering a unified risk view. And most importantly, they are inefficient in a regulatory environment where supervisors are no longer satisfied with activity for its own sake. Increasingly, the question asked is not “Did you run the rule?” but “Did it help you identify real money laundering risk?”
AI/ML: promise vs. compliance realities
Against this backdrop, AI and machine learning are often marketed as the solution. Vendors promise models that can cut false positives, spot hidden patterns, and outthink criminals. Indeed, AI has significant potential in areas such as anomaly detection, entity resolution, and alert prioritization. But there is an inherent misfit between AI/ML and AML compliance. Monitoring must be auditable, and every alert must be justifiable to regulators. Black-box models that cannot explain their reasoning are simply not acceptable.
Machine learning also relies on large, clean, and labeled datasets, whereas AML data is often messy, incomplete, and imbalanced — with far more false positives than actual cases of laundering. Criminal typologies evolve faster than models can be retrained and validated. The governance burden of managing AI models — including documentation, testing, and gaining supervisory confidence — can outweigh the promised efficiencies.
Regulators’ early warnings on AI in AML
Regulators are already taking note of this. In 2025, the European Banking Authority warned that careless use of AI and other RegTech tools can itself create money laundering and terrorist financing risks. The EBA noted that criminals are leveraging AI to generate synthetic identities, forge documents, and automate laundering, while institutions struggle to keep up. The warning was clear: AI is not a silver bullet. It can be a potent enhancer of monitoring, but only if wrapped in explainable, auditable frameworks that satisfy compliance obligations.
What risk-based monitoring looks like
The direction of travel is toward risk-based monitoring. Instead of relying on binary thresholds, institutions are expected to adopt dynamic, contextual approaches that reflect the relative risk of customers, products, geographies, and behaviors. A risk-based system allows enhanced monitoring to be applied to higher-risk scenarios while streamlining low-risk activity.
Customer risk scores can be continuously updated as new behaviors or data emerge. Fraud and compliance signals can be integrated into a single intelligence layer. And institutions can demonstrate proportionality and effectiveness to supervisors, which is becoming the benchmark.
Regulatory drivers across key jurisdictions
This shift is not optional. It is being driven, and in many cases mandated, by regulators. FATF has long enshrined the risk-based approach as the global standard. In the United States, FinCEN’s modernization proposal would make a documented risk assessment a core program element and require alignment with national AML/CFT priorities. In Europe, the 2024 AML package — AMLR, AMLD6, and the new AML Authority in Frankfurt — embeds the risk-based approach directly into law.
The UK’s JMLSG guidance anchors AML obligations in proportionality, while its Crypto Travel Rule requires risk-informed data sharing. Regulators in Singapore and Hong Kong have explicitly instructed institutions to tailor their monitoring based on assessed risk, extending these expectations to the virtual asset sector. Across jurisdictions, supervisors are shifting from “show me the rule” to “show me the outcome.”
What this means for RegTech vendors
For Regtech vendors, this means flexibility and configurability become defining features. Platforms must enable institutions to quickly configure, update, and orchestrate risk models as threats and priorities evolve. Black-box AI is unlikely to survive regulatory scrutiny; explainability is non-negotiable. Integration across silos is more crucial than ever, as institutions require unified views of fraud and compliance risks. And hype-driven vendors will struggle to endure.
The winners will be those who build sustainable business models on trust, transparency, and measurable efficiency.
What this means for financial institutions
For financial institutions, the implications are equally profound. Monitoring is no longer just a compliance box to tick; it must be strategically aligned with the institution’s own risk appetite. Fraud and compliance teams can no longer work in isolation. They need connected workspaces and real-time intelligence sharing. Governance capacity will have to grow — not only to support dynamic risk models but also to validate, document, and defend them under supervisory scrutiny. And with compliance budgets under pressure, institutions will prioritize solutions that demonstrably reduce false positives, consolidate operations, and deliver return on investment.
Benefits for regulators, institutions, vendors, and customers
The shift toward risk-based monitoring promises clear benefits for all parties. Regulators gain more effective outcomes and more transparent oversight when monitoring is proportional and explainable. Financial institutions reduce wasted effort, focus resources on genuine risks, and protect themselves from reputational damage. Vendors who adapt will find sustainable, long-term growth. And customers experience smoother onboarding, fewer unnecessary frictions, and stronger trust in the financial system.
A hybrid, explainable future
AML monitoring is undergoing a structural transition. Regulatory rules remain fixed; they are the immovable foundation of compliance. What is evolving is the decision logic — from rigid, rules-only scenarios to risk-based, contextual monitoring, enriched but not defined by AI. The future belongs to hybrid approaches: explainable, configurable systems that orchestrate diverse data sources, integrate fraud and compliance intelligence, and provide clear audit trails. Achieving true risk-based monitoring also requires consolidation. Customer risk assessment, transactional AML monitoring, and watchlist screening cannot remain scattered across separate tools and silos. Only by unifying these pillars into a single, coherent framework can institutions gain the holistic view regulators now expect, reduce operational noise, and deploy resources proportionately to actual risk.
Orchestration as a defining requirement
This orchestration isn’t just a best practice — it is becoming a defining requirement of RegTech solutions. Gartner has emphasized that while automation is essential, policy-driven orchestration is the missing layer without which tool sprawl, fragmentation, and inefficiency persist. In the AML context, orchestration refers to the unification of CRA, transaction monitoring, and screening into coordinated, auditable pipelines. It is this consolidation that turns compliance activity into genuine risk management.
Regulators are demanding this integration, institutions need it to manage costs and complexity, and vendors must deliver it. Platforms such as NOTO have been designed for precisely this future: modular, explainable, and built to consolidate CRA, transactional AML, and watchlist screening into one solution that bridges compliance certainty with adaptive risk intelligence. The road away from rules-only monitoring is already being paved. Institutions that take it early will not only meet regulatory expectations but also build smarter, more resilient defenses against financial crime.
Ready to move beyond rules-only monitoring?
NOTO unifies customer risk assessment, transaction monitoring, and watchlist screening into a single, explainable orchestration layer—cutting false positives, eliminating tool sprawl, and aligning controls to your risk appetite. Meet evolving FATF/EU/US expectations with configurable policies, clear audit trails, and measurable ROI. See NOTO in action—book a quick demo.




