Financial institutions in the Philippines are facing growing expectations to strengthen their ability to identify and respond to suspicious activity. As digital banking, e-wallets and electronic payments expand, regulators are placing greater emphasis on monitoring systems that can detect unusual behaviour quickly and support effective intervention.
For technology providers such as ZIGRAM, which offers solutions spanning AML, transaction compliance and screening, the challenge reflects a wider need for financial institutions to connect fraud detection with their broader financial crime controls. Risk scoring can support this effort, but institutions must ensure that the models behind their decisions are transparent, appropriately governed and regularly tested.
Fraud risk scoring assigns a numerical value to a transaction, account, customer or session based on indicators that may suggest suspicious activity. These can include transaction velocity, changes to devices or account details, recipient history, unusual locations and deviations from established customer behaviour.
The resulting score can help determine whether activity should be approved, subjected to additional verification, held where permitted or escalated for investigation. However, a score is not evidence of wrongdoing. Its purpose is to support risk-based decisions, not replace human judgement or wider investigative processes.
The regulatory environment is adding urgency to these efforts. Bangko Sentral ng Pilipinas (BSP) Circular No. 1213, Series of 2025, calls for automated, real-time fraud monitoring among supervised financial institutions within its scope. The measures identified include velocity checks, account and device changes, geolocation, blacklist screening and behavioural anomaly detection.
Importantly, the circular does not require institutions to adopt one specific numerical fraud scoring model. Instead, it sets expectations around monitoring capabilities, leaving institutions to determine how their systems should assess and respond to risk in line with their activities and exposure.
This makes model governance particularly important. Institutions need to understand which indicators contribute to a score, how those indicators are weighted and what happens when a transaction crosses a defined threshold. They should also retain records of model versions, decisions, interventions and subsequent reviews so that outcomes can be explained and audited.
The choice of model can influence how easily these requirements are met. Rule-based systems can be easier to interpret because decisions are tied to defined conditions, but they may struggle to keep pace with changing fraud patterns. Machine learning models can identify more complex behavioural relationships, although they introduce challenges around data quality, validation, explainability and recalibration.
Hybrid approaches can combine the transparency of rules with the pattern-recognition capabilities of machine learning. Whichever method is selected, thresholds should reflect an institution’s own risk appetite, customer activity and investigation capacity. A model that generates too many false positives can create unnecessary friction for customers and overwhelm compliance teams, while thresholds that are too permissive may allow fraudulent activity to pass undetected.
Institutions must also avoid conflating fraud risk scoring with AML assessments. Fraud indicators can inform wider financial crime monitoring, but a high fraud score does not automatically establish that a transaction is suspicious under the Anti-Money Laundering Act. Similarly, a low fraud score does not remove the need to consider separate AML risks and reporting obligations.
The Anti-Financial Account Scamming Act (AFASA), Republic Act No. 12010, also strengthens protections against financial account scams. Institutions should assess how its requirements apply to their particular licences, services and activities rather than assuming that a single monitoring model will satisfy every obligation.
ZIGRAM’s Complete AML System, Transact Comply and PreScreening.io sit within the broader technology landscape supporting screening and financial crime compliance. But no single platform or numerical score can guarantee regulatory compliance. Financial institutions must combine effective monitoring with model validation, clear accountability, reliable audit trails and appropriate human oversight.
Copyright © 2026 FinTech Global
Copyright © 2026 RegTech Analyst
