Skip to main content

      AMLA’s recent data collection exercises give a flavour of the intensive reporting requirements 2028 will bring. Obliged entities can leverage lessons learned by participants to build integrated, automated data frameworks that deliver speed, accuracy and efficiency – beyond the questionnaires.

      Supervision is changing, with regulators gathering more data and drawing their own conclusions. AMLA’s two 2026 data collection exercises mark the latest step in the shift towards data-driven supervision.

      The first exercise, to determine the inherent and residual risks of obliged entities (“OEs”), asked participants for no fewer than 254 data points covering administrative information, risk metrics, and AML controls. There was little overlap with existing national reporting templates. In a pattern familiar to financial firms under EU prudential supervision, OEs faced major challenges:

      • Interpreting requirements:

        Definitions are not always clear. Which instruments count as loans? Which relationships count as customers? The effort not only lies in sheer volume. Interpretations can vary between different business units and teams, making it hard to achieve consistency not only across different OEs and also within an institution itself. AMLA’s second exercise of 2026, to identify OEs[AB1]  eligible for direct supervision, revealed its complexity on close inspection. Interpreting "freedom to provide services" alone took considerable time!

      • Identifying data sources:

        OEs have many systems and it is rare for one person or unit to have a full overview. Data fields often look very similar, so picking the right one can take real detective work.

      • Introducing new data:

        Some required data do not exist yet. Adding them costs time and resources. Institutions therefore need to decide whether approximations are acceptable and, if so, what degree of approximation is sufficiently accurate.

      • Achieving automation and reproducibility:

        One-off manual reporting does not scale. New data streams are expensive but often necessary for automated delivery and proper audit trails.


      Addressing these challenges required significant manual effort and improvised workarounds. The good news is that OEs of every type can benefit from the lessons learned by participants, including:

      • Begin with a structured data catalogue

        The right foundation ensures no relevant source is missed and makes developing the business logic much easier.

      • Design iteratively, and document decisions

        AML data has many other potential compliance uses. An agile approach with documented design decisions provides a flexible, lasting solution.

      • Act early, even without full clarity

        Building data structures takes time. Institutions that wait for specifications to be finalized risk not being ready when reporting becomes mandatory.

      • Involve Group entities early, and conduct dry runs.

        Group-wide reporting succeeds or fails at the entity level. Testing the full process early reveals gaps while there is still time to close them.


      The years ahead will introduce thousands of OEs to data-driven supervision. To keep time and resource requirements under control, organisations will need to move from ad-hoc, manual approaches towards integrated and automated processes that treat AML reporting as one of many data products. That includes harnessing the potential of AI tools - under robust supervision - to conduct a range of tasks including data matching, outlier detection, and plausibility checks. 

      AMLA’s data driven approach to supervision calls for a new vision of the business layer, featuring the following elements:

      • Automated data pipelines supported by clear data lineage and ownership – ensuring traceability and enabling every figure to be reconstructed.
      • An architectural set up that allows for drill downs - achieving consistency and comparability over time.
      • Regular, management-ready reporting with clickable reporting extensions – allowing for fast, flexible responses to internal analysis, changing regulation, and requests from AMLA.
      • Integrated compliance data management – aligning AML data submissions with the Risk Analysis and other regulatory reporting.

      Data-driven supervision is here to stay. Institutions that invest now in transparency, automation and integration will do more than meet supervisory expectations; they will gain a clearer view of their own risks, leveraging regulatory pressure to create a real data advantage.

      Supervisors will draw conclusions from your data – ensure that you have drawn them first.



      KPMG AMLA Office

      Navigating AMLA Supervision with KPMG’s Dedicated Office

      Our people

      Fernando Reinoso Martinez

      Associate Partner & AMLA Office Network Coordinator

      KPMG in Spain

      andrea-barnaba
      Andrea Barnaba

      Partner Risk Consulting

      KPMG in Italy

      Lena Zinser

      Partner, Financial Services

      KPMG in Germany

      Claudia Gazzola

      Manager

      KPMG in Italy