Data and regulatory reporting: the automation of compliance
Read the PDF edition (2 pages)
Compliance is the area in which automation is advancing fastest, because its tasks are numerous, repetitive and documented. Geneva’s institutions are automating the screening of names against sanctions lists and lists of politically exposed persons, transaction monitoring, the collection of identification information, the production of regulatory reports to FINMA and the National Bank, the preparation of returns for the automatic exchange of information and the tracking of training obligations. Tools based on machine learning reduce false alerts, as the worked example above illustrated. Language-processing tools read open sources in several languages, which answers a difficulty that Part J described.
This automation has its limits and its risks. A badly configured screening tool lets a sanctioned name through or blocks thousands of legitimate clients. A monitoring model trained on old data does not detect new patterns. Excessive dependence on the tool weakens the analysts’ judgement. In its communication of December 2024, FINMA insisted on the institution’s responsibility, the explainability of results and independent review. The prosecuting authorities, for their part, judge an institution’s compliance by its decisions, whatever its tools.1 Well-managed automation frees the analysts for complex cases, whereas badly managed automation gives an illusion of control.
Notes
- FINMA, supervisory communication 08/2024, cited above. ↩
For general information only. This article does not constitute legal or tax advice, does not replace advice on an individual case and creates no lawyer-client relationship. It reflects the legal position as of September 2025.