Combining Device and Behavioural Signals to Detect Payment Fraud Earlier

Summary
The article recommends combining device intelligence with behavioural analytics to identify device compromise, scams and social engineering before a payment is authorised.
Key points
- Fraudsters may persuade customers to install remote access tools and then prepare or execute payments while obtaining authentication codes.
- Deterministic signals include remote access tool or malware detection, active calls, VPN or location anomalies, device integrity issues and high-risk IP intelligence.
- Probabilistic signals assess behavioural deviations, identity mismatches, signs of manipulation or duress, and similarity to known scam patterns.
- The article recommends correlating both signal types: deterministic indicators can flag technical risk, while probabilistic signals add context about possible manipulation.
- Earlier pre-transaction intelligence can help identify scams, account takeover and device takeover before customers authorize payments.
Article Details
- Defense Focus
- Detect scams, social engineering, account takeover, and device takeover using risk intelligence gathered before payment authorization.
- Detection Methods
- Detect Remote Access Tools and mobile malware on customer devices.
- Monitor active-call signals during banking sessions and payment setup.
- Assess VPN, TOR, location, device-integrity, and high-risk IP signals.
- Score identity mismatches and deviations from known user behavior.
- Identify manipulation, coaching, duress, and pressure indicators.
- Compare customer activity with known scam journeys and fraudster models.
- Correlate deterministic device signals with probabilistic behavioral signals; use weighted risk scores alongside high-confidence standalone indicators.
- Data Sources
- Signals collected during mobile and web banking sessions before payment initiation
- Device security and integrity observations
- Active-call status
- VPN, TOR, location, and high-risk IP intelligence
- Identity and behavioral models
- Known scam patterns and fraudster models
- Platforms
- Mobile banking
- Web banking
- Defensive Actions
- Collect and assess risk intelligence before authorizing a payment.
- Combine device-compromise indicators with behavioral and scam-similarity signals.
- Intervene before payment authorization when correlated signals indicate elevated risk.
