Network for Financial Crime Prevention

Transaction Monitoring Framework for Financial Institutions

A system isn't a framework — and most institutions don't realize the difference until a regulator points it out.In this episode, guest host Marika Torsdotter sits down with Francisco Mainez, a transaction monitoring leader with deep experience across global financial institutions and fintechs, to unpack what separates a real transaction monitoring framework from the software underneath it — and where institutions most often get it wrong. We cover what good looks like, why data remains the unsolved problem of 2026, how training has quietly become the second elephant in the room, and what the shift from rules-based detection toward more advanced analytics actually demands of governance. A few things that stuck with me from this conversation: → A system is the technology. A framework is everything around it — typology library, governance, data, documentation. Buying one without building the other is the most common failure pattern in the industry. → Data is the toothache nobody fixes. Institutions keep buying new systems instead, because data work is harder, slower, and less visible. But you cannot monitor anything if you don't have the data to do it. → Training is the second elephant. Even the most advanced system will produce false positives in front of an investigator who doesn't know what to do with them. As Francisco puts it: "Whatever you sweat in training, you won't bleed in battle."Whether you work in AML, compliance, fraud, or financial crime more broadly — if your institution runs transaction monitoring, this conversation is for you.

Beyond ChatGPT: The AI Arsenal Compliance Has Been Ignoring

In this episode, guest host Marika Torsdotter sits down with AI and financial crime compliance expert Aneta Klosek to cut through the noise around AI and reveal what most compliance professionals are still missing.While the industry fixates on large language models like ChatGPT, Aneta argues there's a whole ecosystem of AI algorithms quietly powering or failing to power the fight against financial crime. The conversation breaks down three concrete use cases:* Gradient Boosting — the workhorse inside most transaction monitoring systems, learning from historical patterns to prioritize alerts and reduce false positives far beyond what static rule-based systems can achieve.* K-Means Clustering — an unsupervised learning technique that segments customers by behavior, not just category. A $50,000 transfer that looks normal for a corporate treasury cluster is a massive red flag for a retail student cluster. Context is everything.* Word Embeddings — the technology that bridges unstructured text and risk intelligence. From recognizing that "Muhammad Al-Hassan" and "M. Al-Hassan" are the same person, to understanding that "cleared of fraud" and "charged with fraud" carry opposite risk signals this model brings semantic intelligence to name screening and adverse media.Aneta closes with a hard truth: the biggest gap in the industry isn't technology it's the skill gap between technical teams and compliance officers, and the dangerous tendency to treat AI as a pilot project rather than a core business strategy. Meanwhile, criminals face none of those constraints.Bottom line: Stop treating AI as a side project. Start with your data. Build from clarity, not hype.

Muddling Through the Hormuz Crisis

In our latest episode, "Muddling Through the HormuzCrisis," we spoke with David Tannenbaum of Blackstone Compliance about the real-world implications of the 2026 Iran conflict for sanctions programs, maritime intelligence, and compliance frameworks. Some of what we cover might surprise you: → Commercial vessels are no longer using standardinternational fairways — they're routing through Iranian-controlled waters near Qeshm and Lavan Island → General License U has created unexpected openings for dark fleet vessels, but key federal prohibitions still apply → The single biggest blind spot? The sheer volume of Iranianoil currently on the water — much of it cycling through FSOs → Known typologies are shifting: bigger vessels giving wayto chemical tankers, Swiss traders potentially giving way to new UAE or Hong Kong entities This episode is a must-listen for anyone in AML, tradefinance, commodity finance, or maritime compliance. 💡 The core message:Robust controls still matter — they just need to be pointed in the right direction.

Trade Compliance Overview Presentation

Trade compliance isn't just a customs issue — it's a financial crime issue.In our latest NFCP podcast, I sat down with Francisco Montesinos Gutierrez, a seasoned trade compliance professional, to break down what every risk and compliance professional needs to know about global trade controls.We cover the three pillars of trade compliance — custom controls, sanctions & embargoes, and export controls — and why understanding all three is no longer optional if you're working in AML, fraud, or broader risk management.A few things that stuck with me from this conversation:→ Non-compliance with export controls isn't just a fine — it's a criminal offence→ Custom misclassification penalties can reach up to 100% of commercial value→ Sanctions, HS codes, and export licenses are more interconnected than most people realizeWhether you're new to trade compliance or looking to sharpen your cross-domain skills, this episode is a solid place to start.

The FinCrime Friday Drop Q1

Marco Beranzoni (FinCrime Agent) and Louie Vargas (NFCP)opened the series the right way — no slides, no scripts, just an honestconversation about how two FinCrime professionals built their careers from theground up. A few things that came up worth sharing: → AI is coming for entry-level roles. Marco made itclear: if you're thinking about breaking into financial crime prevention, thewindow is now. In 3–5 years, those first-line, high-volume roles will look verydifferent. → Certifications alone won't get you there. Louie putit plainly — the cert without the practical experience is just a lamp without agenie. Both matter.