AI in Financial Services Podcast Why Siloed Fraud and AML Systems Are Failing Financial Institutions - with Debjit Saha of MoneyGram
Jan 26, 2026
Debjit Saha, VP of Engineering & Product for Risk & Compliance at MoneyGram, builds data- and AI-driven controls for fraud, compliance, and payments decisioning. He discusses why siloed fraud, AML, and sanctions systems break down. He covers the market forces pushing unification. He outlines measuring success and designing tiered human-in-the-loop controls. He examines how sophisticated, syndicated fraudsters exploit gaps.
AI Snips
Chapters
Transcript
Episode notes
Market Forces Driving Unified Risk
- Rising investigation costs, regulatory pressure, and smarter fraudsters are forcing firms to unify fraud, AML, and sanctions functions.
- Consolidated customer views and shared behavioral data make model-based anomaly detection more effective than fragmented rules.
Standardize Tooling To Break Silos
- Standardize tooling and processes to overcome organizational silos between fraud and compliance teams.
- Use common platforms so shared data and models can improve detection and reduce costly referrals.
Measure Outcomes, Not Just Models
- Measure success by cost reduction, detection outcomes, and overlap between fraud and regulatory reporting.
- Track fraud losses, SAR overlap, and previously unexamined populations to prove unified controls deliver better outcomes.