Anti-money laundering under full evidentiary obligation

Supervisors expect an effective control system against financial crime: transaction monitoring that works, dependable procedures for customer due diligence, traceable audit evidence and documented internal safeguards. The existing setup could no longer produce that evidence to the required depth.

• Risk-based monitoring

  Monitoring by risk content instead of across the board, supported by rule-based scenarios and learning procedures.

• A dependable basis for decisions

  Reliable data and tested procedures as the precondition for a control that holds up in front of the supervisor.

• Complete evidence

  Full documentation of all requirements, thresholds, settings and changes, verifiable at any time.

Against that background, the task was to put transaction monitoring on a new footing and at the same time create a processing basis that meets these requirements permanently and at the required speed.

1. 1

   Clarifying requirements between compliance, the business units and IT

   Building and maintaining a binding reference for business, supervisory and technical requirements. It brought together the customer due diligence requirements, the monitoring scenarios including thresholds, the data quality indicators, the requirements for evidence and the analysis views for the business units.

   • agreed in workshops with compliance, the business units and IT
2. 2

   Business Architecture and processing basis

   Establishing a processing chain that makes more than 70 million transactions with 16 months of history available daily within 6 hours. The picture of the situation was therefore available before the business day began, not after it. It ran on the cloud platform already established in house, and the tools in use, among them Google Cloud Platform, BigQuery and SparkSQL, followed the requirement rather than the other way round.

   • picture of the situation available before the business day begins
   • provable origin for every record
3. 3

   Securing the basis for decisions

   Continuous checks on completeness, freedom from contradiction and consistency of the holdings, with alerts on deviation. In addition, targeted special analyses to bridge infrastructure outages without losing the ability to report.

   • deviations are detected before they become reportable
4. 4

   Operations, support and enablement

   Operation of the processing chain with continuous optimization. Support for the business units in tuning the monitoring rules, to reduce the number of unfounded alerts and direct review capacity at the relevant cases. Enablement of internal staff so that the institution can take over the steering itself.

   • the knowledge stayed in house

• Processing large holdings

  Holdings at petabyte scale had to be available so that analysis could keep pace with the business day. That was carried by the cloud platform of the institution with BigQuery, Cloud Storage, SparkSQL and Hadoop.

• Business Integration

  Bringing holdings together across system boundaries that had grown over time, so that compliance, the business units and management look at the same basis. The existing database landscapes, among them Oracle and MS SQL Server, stayed in operation.

• Regulatory control

  Rule sets for anti-money laundering and customer due diligence, implemented in NICE Actimize SAM and STAR.

• Analysis and steering

  Analysis views for the business units and for management, supported by Looker, SQL Developer and the steering tools of the institution.

• Trigger

  70 million transactions a day, picture of the situation required before the business day begins

  →
  Advisory approach

  processing re-cut and parallelized, history extended to 16 months

• Trigger

  compliance, the business units and IT pursued different objectives

  →
  Advisory approach

  one binding reference, shared prioritization, early analysis views

• Trigger

  review capacity tied up by unfounded alerts

  →
  Advisory approach

  thresholds and scenarios recalibrated, anomalies weighted by risk content

• Trigger

  dependence on external knowledge

  →
  Advisory approach

  enablement of the internal teams, support during ongoing operations

• SIMO GmbH put our transaction monitoring on a new footing. Unfounded alerts fell by 78%, and our compliance department was able to direct its capacity at the relevant cases. What stood out was the migration of 50 million historical transactions without a single loss.

• Strong professionally and in the collaboration. Detection quality is well above that of our previous systems. The ongoing monitoring of 2.3 million transactions a day has run reliably since go-live.

• The supervisory experience at SIMO was decisive for us. The newly established transaction monitoring meets the requirements of BaFin and is designed for the European rules to come. We had not expected the rollout to be complete in six weeks.

• Pattern recognition identifies suspicious transaction flows with an accuracy of 94.7%. The quality of our reports to the FIU has improved by 60%. What counts most for us is that every report today is traceably justified.

Effect for the institution

• picture of the situation with 16 months of history available in 6 hours
• review capacity directed at the cases that actually matter
• a processing basis that grows with the business
• evidence that withstands a supervisory examination
• the steering sits with the institution, not with the advisor

The case shows what Enterprise Business Data Strategy delivers in a regulated environment: a control function that does not merely exist on paper but carries every day. What mattered was not the technology in use but the clarification of accountability, requirements and evidence between compliance, the business units and IT.

SIMO roles in the mandate

• Business Data Management Team Lead
• Data Manager
• Data Scientist
• Data Quality Manager
• Project Manager
• Product Owner

Is your control function under examination pressure?

Let us establish whether your evidence holds when it counts and where it has gaps.