What the EU AML Regulation means for customer due diligence, and why deciding what matters is the hard part

Last updated: 16th November 2026
Much of the market conversation about AMLA centres on a move from periodic KYC to “perpetual” or continuous KYC. But the real challenge is not to keep customer records continuously updated - it is to understand when a change matters, what it means for a customer's risk, and what the institution should do about it.
The EU Anti-Money Laundering Regulation (AMLR) makes this challenge more explicit and more consistent across the EU. From 10 July 2027, Regulation (EU) 2024/1624 will apply directly, establishing maximum intervals for updating customer information alongside requirements to respond when relevant circumstances change or new facts come to light.
For customers subject to enhanced due diligence, the maximum interval is one year. For all other customers, it is a maximum of five years. Institutions must also review, and where relevant update, customer information when a customer's circumstances change, a relevant fact becomes known, or certain other legal obligations arise.
In other words, financial institutions must manage both scheduled updates and event-driven change. They must also bring existing customer records into line with the new requirements.
That combination creates a significant operational challenge for many institutions. Adding more reviews, data feeds or triggers will not solve this alone. Institutions need to distinguish meaningful change from noise, connect relevant information to customer risk, and demonstrate why the action taken was appropriate.
The regulatory shift towards continuous CDD
The EU Anti-Money Laundering Regulation (AMLR) introduces a more harmonised set of directly applicable CDD requirements across the EU from 10 July 2027.
The supporting CDD requirements are becoming more specific. On 1 October 2026, AMLA finalized its regulatory technical standards under Article 28(1), covering the information institutions must obtain and verify when conducting CDD; and Article 19(9), addressing how institutions distinguish business relationships from occasional transactions and identify linked transactions. Together, these clarify both what institutions need to establish about customers and when CDD obligations apply.
Alongside these requirements, daft guideline Article 26 establishes the obligation to maintain an up-to-date understanding of customers and monitor transactions to ensure they remain consistent with the institution's knowledge of the customer, their business activities and risk profile. This goes beyond scheduled reviews, customer information must also be reviewed when circumstances change or relevant new information becomes known. The ongoing-monitoring guidelines, following a consultation that has now closed, provide further detail on how institutions should put this into practice.
The double load: Remediation and ongoing monitoring
Existing customer files and future monitoring obligations will need to be addressed together. Institutions will need to uplift the technology and automation components first, and address missing aspects of customers records in a practical manner starting with the highest risk customer segments first, to ensure records remain fit for purpose and in a timely manner acceptable to the regulatory supervisor.
The scale of the challenge is real. In its final report on the draft CDD standards, AMLA recorded that a significant number of financial-sector respondents who provided a substantiated assessment expected one-off implementation costs to be disruptive. Respondents highlighted work on policies, processes, systems, additional customer information, remediation and the integration of monitoring and screening capabilities.
The operational pressures they identify are familiar, more information to maintain, more evidence to verify, and more work to complete with finite compliance resources.
Increasing the number of external data feeds (such as company registries, corporate data providers, credit bureaus) and generating more alerts when data changes, is not the answer. Without a way to assess which changes are relevant and require action, institutions risk replacing a periodic-review backlog with a trigger event backlog rather than improving how they manage customer risk.
An institution may know that something has changed without knowing whether the change is relevant to the customer's risk, whether it requires a review, or what action is proportionate.
The objective should be to ensure that meaningful changes lead to timely, appropriate and evidenced action.
Materiality: The decision between a signal and an action
Article 26 of the AMLR establishes when customer information must be reviewed and, where relevant, updated. AMLA's draft ongoing-monitoring guidelines go further in explaining how institutions should assess whether a trigger represents a relevant and material change in a customer's risk profile that warrants reassessment.
That distinction is important. Materiality is not a standalone statutory test with one universal threshold. It is an operational control question, given the customer's circumstances, the information available and the institution's risk methodology, what does this change mean and what should happen next?
Getting this right requires three disciplines.
1. Define the decision criteria and who owns them
A change in ownership, geography or business activity will not have the same implications for every customer. Institutions need criteria grounded in their business-wide risk assessment and customer risk methodology, with clear ownership of the rules and thresholds used.
These should not be determined by a technology provider's default settings or left implicit in individual workflows. Compliance teams must be able to explain why a particular change matters and how the methodology translates that judgement into action.
2. Assess relevant signals in context
A change in ownership may not, on its own, justify a significant reassessment. Combined with exposure to a higher-risk jurisdiction, adverse information or a material shift in transaction behaviour, it may warrant closer scrutiny.
The context of this matters. The point is not that every combination should automatically increase a customer's risk rating. It is that institutions need to consider relevant signals together, against what is known about the customer, the purpose of the account and the expected nature of the relationship.
3. Avoid unnecessary work without weakening the control – expired documents
AMLA's draft ongoing-monitoring guideline takes a risk-based approach to the re-collection of expired identity documents rather than treating expiry alone as an automatic requirement to collect a replacement in every case. Why? AMLA recognizes that a document's expiry does not automatically mean a customer's verified core identity has changed, or that the money laundering/terrorist financing (ML/TF) risk has increased.
That illustrates a broader principle, an institution should assess whether new information is needed to understand or mitigate the risk, rather than automatically repeating work that may add little value. Any decision not to obtain further information must still be consistent with the applicable requirements and the risks identified.
The result should be a clear, defensible distinction between a signal that can be recorded, one that warrants reassessment, and one that requires an investigation or other action.
One customer, a coherent risk assessment
In many institutions, CDD, Transaction Monitoring, Screening and Fraud have developed as separate capabilities, often with different data, workflows and decision-making processes.
There are sound reasons for specialised detection and investigation. But the resulting boundaries can make it difficult to understand what the institution collectively knows about a customer.
The AMLR requires ongoing monitoring to consider whether transactions and activities are consistent with the institution's knowledge of the customer, their business activities and risk profile. AMLA's draft ongoing-monitoring guideline also envisages monitoring outputs informing customer profiles and risk reassessment where material changes are indicated.
Consider a customer whose transaction monitoring activity indicates a material change in behaviour. The TM team may own detection and initial investigation. But if the outcome changes the institution's understanding of the customer's expected activity or risk, that information should be capable of informing the CDD assessment without waiting unnecessarily for the next periodic review.
The same principle applies to relevant screening findings, changes in ownership and other risk-relevant information. The institution needs to know which signals should be shared, who assesses their significance, which action follows, and how the outcome is recorded.
The important distinction is between connecting systems and connecting decisions. Data integration alone does not guarantee that a material change will influence the right customer risk assessment or reach the right team.
Proportionality has boundaries
A risk-based approach should result in different responses to different risks. An ownership change may require verification of the ownership structure, a source-of-wealth concern may require additional corroboration, a combination of adverse information and unusual transaction behaviour may require a broader investigation.
Applying the same workflow to every event can waste investigator time and create unnecessary requests for information from customers. But proportionality does not mean that institutions can disregard minimum CDD requirements or avoid enhanced measures where higher risk has been identified.
The AMLR establishes the framework for simplified, standard and enhanced due diligence. The final draft CDD standards provide further detail while retaining a risk-based approach.
They also recognise that institutions should consider information already available to them when understanding the purpose and intended nature of a business relationship. In lower-risk situations, the information required may be limited, and some understanding may be derived from the product or service and information already held. Where that information is insufficient to mitigate the identified risks, further information remains necessary.
For institutions, this creates an opportunity to move away from treating every review as a request to start again. They have the ability to understand what information is already reliable and sufficient, what has changed, and what additional targeted due diligence is needed proportionate to the risk that has been identified.
AI should prepare the decision, not obscure it
Much of the effort involved in complex CDD investigations is preparatory, tracing ownership chains, gathering and reconciling evidence, corroborating source-of-wealth and source-of-funds, and assembling the rationale for a decision.
AI can help reduce that effort by finding relevant information, summarising evidence, identifying inconsistencies and preparing an investigator to make a decision. Specialist agents can support bounded tasks, such as ownership and UBO identification or source-of-wealth corroboration, provided their outputs are checked against reliable evidence and their limitations are understood.
The regulatory boundary matters. Article 76 of the AMLR permits automated decision-making, including AI, subject to conditions. Specified decisions about entering into, refusing or maintaining a business relationship, and changing the extent of CDD, must be subject to meaningful human intervention.
AMLA's draft ongoing-monitoring guideline adds that institutions using automated or advanced analytical tools should be able to explain their role, functioning and outputs, and ensure that responsibility for decisions remains clear. The use of a tool is not, by itself, evidence that monitoring is effective.
For product design, the implication is straightforward. AI should reduce the effort required to reach an informed decision, not make the decision harder to challenge. An investigator should be able to see the relevant evidence, understand the reasoning behind a recommendation, identify uncertainty and record their own decision.
The aim is to produce a decision-ready case with evidence and accountability intact.
Govern the methodology, not just the decisions
Risk methodologies cannot remain static. Institutions need to respond to changes in their risk appetite, business model, customer base, emerging threats and regulatory expectations.
If changes to risk factors, thresholds or rules are difficult to implement, keeping the methodology aligned with evolving risks and business requirements can become challenging. Teams may understand what needs to change but need a more efficient and controlled way to make those adjustments.
The answer is governed flexibility, not unrestricted configuration.
This allows compliance teams a clear way to define and adjust risk factors, materiality criteria and actions within approved boundaries. Changes should be tested before deployment, appropriately approved, version-controlled and traceable to the decisions they produce. Institutions should be able to explain what changed, why it changed, who approved it and whether the monitoring framework remains effective.
AMLA's draft ongoing-monitoring guideline Article 26(5) specifically discusses validation when monitoring logic or analytical methods change materially. This reinforces the importance of treating methodology changes as controlled changes to a financial crime control, rather than as routine configuration with no evidence of impact.
A configurable CDD risk engine, connected to relevant signals and investigation workflows, can help make this operationally achievable. The value is not configuration for its own sake. It is the ability to evolve a risk methodology without losing control, consistency or auditability.
Modernize in sequence
Consider the following four priorities:
1. Remediate into the future model. The back-book exercise is a one-off chance to build structured, current customer data. Remediating again through documents and free text recreates the problem.
2. Define materiality before scaling event-driven reviews. Agree which triggers require information updates, which warrant reassessment, which require investigation and which invoke mandatory escalation controls.
3. Connect financial crime outcomes to customer risk decisions. Establish how relevant findings from CDD, Transaction Monitoring, Screening, Fraud and other sources should influence a customer's risk assessment and whether and when further review or action is required.
4. Automate evidence gathering within a governed framework. Prioritise work that consumes significant investigator time, preserve traceability, and maintain meaningful human intervention where required.
The real differentiator
The institutions that respond most effectively to the AMLR will be those that can maintain a current understanding of their customers, recognise meaningful change, take proportionate action and explain why.
That requires more than continuous data collection. It requires a coherent approach to customer risk, clear decision ownership, governed methodologies and evidence that connects what the institution knows to what it does.
The strategic question is therefore not simply how to make CDD continuous. It is how to ensure that an institution's understanding of its customers evolves as quickly as the risks they present, and that every material change can be translated into a proportionate, explainable and defensible decision.
Related Resources
EU AMLA Technology Readiness Hub
What is AMLA? A brief overview for compliance leaders
AMLA's new CDD standards: Seven KYC developments organizations should understand
EU AMLA Technology Impact Matrix
From Regulation to Action: Getting EU AMLA-ready
AMLA compliance: Five technology questions financial institutions should ask
Learn more about Symphony Risk Intelligence
Discover the Symphony Risk Intelligence platform and begin your journey towards Always-on Compliance.
The EU Anti-Money Laundering Regulation (AMLR), Regulation (EU) 2024/1624, applies directly across the EU from July 10, 2027. From that date, financial institutions must meet its customer due diligence requirements, including for existing customer records.
Under the AMLR, customer information must be updated at least once a year for customers subject to enhanced due diligence and at least every five years for all other customers. Institutions must also update it whenever a customer’s circumstances change or a relevant new fact becomes known.
Periodic reviews take place at scheduled intervals, while event-driven reviews are triggered by changes such as a new ownership structure, exposure to a higher-risk jurisdiction, or adverse information. The AMLR requires financial institutions to manage both.
Materiality is the decision on whether a change in a customer’s circumstances should be recorded, trigger a risk reassessment, or require an investigation. Clear materiality criteria help institutions avoid replacing a periodic-review backlog with a trigger-event backlog.
Yes, Article 76 of the AMLR permits automated decision-making, including AI, subject to conditions. Decisions about entering into, refusing or maintaining a business relationship, or changing the extent of customer due diligence, must involve meaningful human intervention.
.png)