Combining Device and Behavioural Signals to Detect Payment Fraud Earlier

· Original article ↗

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.

Industries