An Exploratory Study of Sociotechnical Issues for Anti-Money Laundering Workers

Paper Title

An Exploratory Study of Sociotechnical Issues for Anti-Money Laundering Workers

Publication Info

• Topic area: Sociotechnical challenges in anti-money laundering (AML) workflows and software systems.
• Keywords: AML systems, financial institutions, workflow fragmentation, human-centered design, machine learning, collaboration, enterprise software, regulatory compliance, socio-technical gap, qualitative methods.

Background and Problem

• Problem / challenge: AML systems often fail to fully support the workflows of financial institutions, leading to fragmented processes, reliance on manual workarounds, and inefficiencies in detecting and investigating suspicious activities.
• Significance: Effective AML systems are critical for compliance with government regulations, preventing financial crimes, and avoiding substantial fines. Improving these systems can enhance security and reduce operational burdens.
• Motivation and related work: Prior research has explored AML policies, financial technologies, and blockchain implications but lacks a human-centered perspective on AML professionals' workflows and software tools. This study aims to fill that gap by investigating user needs, challenges, and collaboration practices.

Solution

• Proposed approach: Conduct a qualitative study using semi-structured interviews with AML professionals to identify socio-technical challenges, workflow pain points, and collaboration practices.
• Novelty:
  1. Establishing an understanding of AML software and the workflow of AML professionals.
  2. Describing the socio-technical challenges and needs of AML professionals and their tools.
  3. Providing design recommendations for AML software and similar enterprise systems.
• Procedure and key techniques:
  • Recruitment of 15 AML professionals from diverse roles and institutions.
  • Semi-structured interviews conducted remotely via Zoom.
  • Reflexive thematic analysis of interview transcripts using qualitative coding software (Atlas.ti) and affinity diagrams (Miro).
  • Identification of themes relevant to HCI and AML system design.

Results

• Concrete findings:
  • AML systems often lack robust integration with banking systems, leading to incomplete data ingestion and reliance on external applications.
  • Fragmented workflows require AML professionals to juggle multiple applications, causing inefficiencies and reliance on manual processes.
  • Investigators value workflow flexibility but face challenges due to prescribed workflows imposed by AML systems.
  • Collaboration among AML professionals is essential but poorly supported by existing software.
  • Repetitive tasks in AML workflows present opportunities for automation and AI, such as automated data extraction, entry, and decision-making.
• Advantage over baselines: The study highlights gaps in existing AML systems and provides actionable insights for improving workflow support, customization, and collaboration features.
• Experiments / evaluation: The study analyzed qualitative data from 15 interviews with AML professionals across various roles and institution types. Findings were synthesized into themes and visualized using diagrams.
• Limitations and future work:
  • Limited global representation; most participants were from the United States.
  • Challenges in recruiting participants due to confidentiality concerns.
  • Future work should explore customization and flexibility in AML workflows, support for creative processes, and the balance between regulatory oversight and user needs.

Summary

This study provides the first in-depth analysis of AML professionals' workflows and software tools, revealing significant socio-technical challenges such as fragmented workflows, incomplete data integration, and limited collaboration support. Findings suggest opportunities for improving AML systems through better design, automation, and support for creative and collaborative processes. The insights are applicable to other regulated industries and enterprise settings, emphasizing the need for human-centered approaches in software development for specialized domains.