Turning suspicious transaction reports into faster financial fraud investigations

Use case: UC1.2
Location: Portugal
Led by: PJ

Use Case 1.2 summary

Enhances Financial Intelligence Unit (FIU) capabilities by integrating and analyzing SAR/STR data through AI-driven predictive analytics. It reduces analysts’ workload, increases the volume of communications assessed, detects complex and emerging AML patterns, and generates actionable alerts and reports, enabling faster, more informed investigations and supporting the identification of financial fraud.

 

The challenge

The Polícia Judiciária (PJ), through its Financial Intelligence Unit (FIU), faces challenges in preventing and investigating financial fraud, including unauthorized access to confidential information, delays in implementing the GoAML platform, disruptions to critical services, and difficulties in efficiently processing and analyzing large volumes of data from SARs, STRs, and criminal cases.

The scenario

The scenario starts when the Financial Intelligence Unit receives a SAR/STR concerning suspicious banking activities involving a company, individual, or group of individuals. The information is processed and analyzed to identify known and emerging risk indicators, suspicious patterns, and potential modus operandi associated with financial fraud. Predictive analytics and machine learning are used to assess the reported activities, generate alerts, and identify previously unknown behavioural patterns. Analysts validate the results and use the generated insights to guide further analysis and investigative activities. Finally, the identified patterns, risk indicators, and relevant findings are consolidated into reports supporting criminal investigators in assessing the case and preparing potential investigation proposals.

GANNDALF in action

GANNDALF in Action will address these challenges by integrating SAR/STR data and applying AI-driven analytics to efficiently screen and interpret large volumes of financial information. The solution will identify risk indicators, detect known and emerging behavioural patterns, and uncover potential modus operandi associated with financial fraud. Automated alerts will highlight suspicious activities, while near-real-time analysis will support timely decision-making. Analysts will validate AI-generated insights, and comprehensive reports will provide investigators with actionable intelligence to guide investigations and support evidence-based investigative proposals.

Expected impact

  • Reduced analysis time: Automates the screening and processing of SAR/STRs, allowing FIU analysts to focus on higher-value investigative activities
  • Increased analytical capacity: Enables a greater volume of communications and financial data to be processed and assessed, including near-real-time analysis
  • Enhanced pattern detection: Identifies complex or previously unknown AML patterns and behavioural indicators that may be difficult to detect through manual analysis of large datasets
  • Improved investigative support: Generates risk-based alerts and actionable reports, helping investigators prioritize cases, identify potential modus operandi, and make more informed decisions