What the F Happened? Fraud and Financial Crime, Deconstructed

In this month's briefing, Emily and Jeff unpack two major June 2026 compliance deadlines, including SEC Regulation S-P and Nacha's ACH fraud monitoring requirements. They explore how FedNow and instant payments are forcing financial institutions to detect fraud in milliseconds rather than days, why mule accounts remain a critical threat, and how information sharing is becoming essential to stopping scams before funds disappear.

The conversation also examines the growing role of agentic AI in fraud operations, the rise of AI-powered fraud tactics such as synthetic identities and deepfakes, and what Patrick Harr's appointment as CEO means for DataVisor's next phase of growth in the AI-native fraud prevention market.

Links

• 5 Ways Fraud Teams are Using Agentic AI Right Now
• 5 Ways AML Teams are Using Agentic AI Right Now
• BusinessWire: Patrick Harr Appointed CEO, DataVisor
• DEFEND Webinar | AI Tools Fraudsters Use, 2026 Edition
• DEFEND Training
• Transcript

Chapters

00:00 Introduction: Why Real-Time Payments Change Everything

01:11 SEC Regulation S-P Deadline Explained

02:52 Nacha ACH Fraud Monitoring Phase 2 Requirements

04:08 Authorized Push Payment Scams and Behavioral Detection

05:58 From Legacy Payment Rails to FedNow

07:49 Mule Accounts and Financial Crime Networks

09:35 Collective Intelligence and Real-Time Risk Sharing

10:22 Agentic AI and the Future of Fraud Operations

12:47 Deepfakes, Synthetic Identities, and AI-Powered Fraud

14:18 Patrick Harr Named CEO of DataVisor

16:23 Key Takeaways for Fraud and Compliance Teams

18:16 Closing Thoughts and DEFEND Training

Every time you click "Buy Now, Pay Later," an invisible war plays out in milliseconds, and the fraudsters are organized, well-funded, and sophisticated. In this episode of What the F Happened? Fraud and Financial Crime Deconstructed, hosts Jeff and Emily break down the real engineering battle behind BNPL fraud prevention.

You'll learn how unsupervised machine learning (UML) and link analysis graphs are solving a problem that traditional supervised AI and rules engines simply couldn't, including detecting synthetic identity fraud, account takeovers, and abuse rings in milliseconds.

In this episode:

• How synthetic identities are built and why they fool legacy AI models

• The link analysis technique that exposed a 10,000-person fraud ring sharing the same first name

• A 5:1 "hurt ratio" that was killing one major fintech's business

• How unsupervised machine learning, knowledge graph technology, and real-time behavioral analysis detects coordinated fraud rings in milliseconds

• The measurable impact on operations and revenue, including tens of millions in prevented fraud losses, reduced false positives, and dramatically improved review efficiency

Read the following case studies: 

• https://www.datavisor.com/intelligence-center/case-studies/stopping-emerging-bnpl-fraud-rings-in-real-time

• https://www.datavisor.com/intelligence-center/case-studies/5-case-studies-uml-real-time-fraud-prevention

• https://www.datavisor.com/intelligence-center/case-studies/consumer-lender-raises-the-bar-for-customer-experience-and-reduces-fraud-with-machine-learning-wp

Chapters:

00:03 Introduction & BNPL Friction 

01:44 The Business of the Hurt Ratio

04:17 Inside Organized Fraud Rings

05:54 Why Supervised AI Fails 

07:32 Limits of Traditional Clustering 

09:24 Shifting to Unsupervised ML 

10:02 Feature Generation & Data Tech 

11:29 Graph Databases & Live Testing

17:52 Financial Results & Conclusion

In this Analysis, we examine why fraud and AML case management systems are breaking down under the pressure of modern financial crime, and how high-performing institutions are rebuilding their operations around behavioral intelligence, explainable AI, and unified investigations. 

This episode explores:

• The Shift to Entity-Level Investigation: Why traditional transaction monitoring is failing against coordinated criminal networks and how AI-driven graph visualization can achieve up to 100x efficiency gains by surfacing entire fraud rings in a single view.

• The "FRAML" Convergence Strategy: Navigating the regulatory and operational tension between fraud and AML teams by implementing a shared intelligence layer that maintains separate mandates while eliminating dangerous data silos.

• The 2026 Regulatory Reckoning: How new FinCEN standards for "demonstrated effectiveness" necessitate a move away from "black box" algorithms toward explainable AI and immutable audit snapshots to defend decisions to federal examiners.

Read the full analysis and related research: https://www.datavisor.com/blog/why-fraud-case-management-breaks-down-and-the-playbook-to-fix-it

Chapters:

00:00 Intro to Fraud & AML Crisis

02:05 Why Legacy Systems Fail

05:08 Alert Overload Explained

08:12 AI Triage & Entity Reviews

11:18 Breaking Fraud & AML Silos

14:42 The Engineering Bottleneck

18:30 Explainable AI & Audit Risk

22:42 Compliance & Final Takeaways

This month, we deconstruct the massive shift toward Agentic Commerce and the $176 million regulatory wake-up call for financial institutions. We explore how Unsupervised Machine Learning (UML) enables 10-millisecond detection and why biometric injection attacks are rendering traditional authentication obsolete.

In this episode, we cover

• Agentic Commerce: Why AI will handle 25% of e-commerce by 2030.

• Lethal Trifecta: The architectural flaws in autonomous financial systems.

• Synthetic Biometrics: Bypassing hardware with digital signatures.

• FRAML Convergence: Breaking down the silos between Fraud and AML teams.

Find the resources attached here:

• 5 Ways AML Teams Are Using Agentic AI Right Now‍
• 5 Ways Fraud Teams Are Using Agentic AI Right Now‍
• Case Study: Financial Institution Fights Fraud Transactions in Real Time
• DEFEND Training

Chapters

00:00 Intro: The Rise of Agentic Commerce

02:56 The "Lethal Trifecta" of AI Risk

05:08 Death of Point-in-Time Auth

06:26 The $176M Regulatory Warning Shot

07:38 Tangible Off-Ramps & AML Chokepoints

09:00 Merging Fraud and AML Silos

10:43 Tracing Stablecoins & Multi-Chain Tech

13:16 Synthetic Biometric Injection Attacks

15:01 Unsupervised Learning (UML) Defense

19:23 AI Optimization & Human Liability

• 5 Ways AML Teams Are Using Agentic AI Right Now‍
• 5 Ways Fraud Teams Are Using Agentic AI Right Now‍
• Case Study: Financial Institution Fights Fraud Transactions in Real Time
• DEFEND Training

In this episode of "What the F Happened? Fraud and Financial Crime Deconstructed," we dive into the Forrester Financial Crime Management Solutions Landscape (Q1 2026). As digital transactions become the global standard, financial institutions and non-traditional platforms—from dating apps to gig economy services—are weaponizing data to combat increasingly sophisticated criminal networks.

We explore the revolutionary shift toward FRAML (the convergence of Fraud and Anti-Money Laundering), the rise of synthetic identity fraud, and how Large Language Models (LLMs) are rewriting the rules of detection. We also discuss the "ultimate tightrope" businesses must walk: maintaining ironclad security without alienating customers through needless friction.

Key Takeaways:

• The FRAML Shift: Why breaking the silos between bouncers (fraud) and accountants (AML) is now mandatory.

• Digital Identity: How Decentralized Digital Identities (DIDs) are replacing vulnerable data honeypots.

• AI Explainability: The challenge of ensuring AI decisions remain readable for government regulators.

Find the resources attached here:

FORRESTER LANDSCAPE REPORT | The Financial Crime Management Solutions Landscape, Q1 2026

BLOG | Forrester Report: Financial Crime Is Converging And Most Fraud & AML Systems Aren’t Built for It

DEFEND Training

Chapters

(00:00) Introduction: The Invisible Infrastructure

(01:02) The Forrester Report and Rise of FRAML

(03:40) Threat of Synthetic Identity Fraud

(05:08) Under the Hood: Real-Time Data Lakes

(06:20) Security in Dating and Gig Apps

(08:52) The Customer Experience Tightrope

(10:18) AI Revolution: LLMs and Deepfakes

(12:15) The AI Explainability Trap

(14:32) Decentralized Digital Identities

(16:00) The Strategic Vendor Landscape

(20:15) Conclusion: The Invisible War

In this Webinar Recap episode, we highlight key takeaways from the recent session on Agentic AI for Fraud & AML with industry experts from DataVisor, including practical insights, audience questions, and real-world examples discussed during the live conversation.

You’ll hear highlights on:

• Why agentic AI deployments are failing in regulated fraud and AML environments

• The difference between AI chat, AI agents, and conversational AI agents — and why it matters

• How financial institutions can design AI systems with explainability, human oversight, and auditability at the core

Watch the full webinar: https://www.datavisor.com/intelligence-center/webinars/defend-webinar-ai-agents-for-fraud-aml-a-new-era-of-detection-and-strategy-optimization

Chapters

(00:00) AI and the Modern Fraud Crisis(01:58) Industrial Scale Credential Attacks(03:22) AI Chat vs. Conversational AI Agents(05:14) 3 Fatal Traps of AI Deployment(07:55) Conversational AI in Fraud Operations(08:22) UML & Detecting Unknown Unknowns(09:52) Regulatory SAR Narrative Drafting(12:19) The Future of Computational Warfare

In this Analysis, we examine what Claude Mythos reveals about the future of cybercrime - specifically, how a general-purpose AI unintentionally became a highly effective autonomous hacking system, exposing critical weaknesses in global financial infrastructure.

This episode explores:

• How Mythos’ emergent capabilities—like vulnerability discovery and chaining—are redefining what AI can do without being explicitly trained for cyberattacks

• Why financial institutions, built on decades-old infrastructure and complex vendor ecosystems, are uniquely exposed to AI-driven threats

• What this means for the future of fraud, identity verification, and the shift toward continuous, behavior-based security models

Read the full blog: https://www.datavisor.com/blog/mythos-changes-the-threat-model-the-fraud-and-aml-industry-needs-to-catch-up

Chapters:

00:00 Intro: The DEFEND Podcast 

00:27 Claude Mythos: The Accidental Hacker 

04:38 Understanding Vulnerability Chaining 

05:59 Project Glasswing: The AI Shield 

06:51 Financial Crisis & Legacy COBOL Code 

10:01 Machine Mayhem & Agentic AI 

13:41 Behavioral Biometrics vs. Deepfakes

17:33 The Future of Financial Security

In this Monthly Brief, we break down the most important fraud and financial crime developments from April 2026, including emerging threats, notable enforcement actions, and trends financial institutions should be paying attention to.

This episode covers:

• The Mythos shift and the rise of autonomous AI systems capable of identifying vulnerabilities and executing multi-step fraud strategies

• FTC warning letters on “debanking” and the growing regulatory scrutiny around account closures, bias, and consumer protection

• New federal actions targeting global scam centers, including executive orders and a dedicated fraud task force reshaping enforcement priorities

Get the full newsletter by subscribing to DataVisor. Links to all resources are mentioned below:

Finance Firms Brace for Tighter US Regulations: https://www.pymnts.com/news/regulation/2026/embedded-finance-firms-brace-for-tighter-us-regulations/

5 Case Studies of Unsupervised Machine Learning (UML) for Real-time Fraud Prevention: https://www.datavisor.com/intelligence-center/case-studies/5-case-studies-uml-real-time-fraud-prevention

The First Conversational AI Agents for Fraud & AML:⁠ https://www.datavisor.com/products/ai-agents

Financial Indicators of Human Trafficking: What Financial Institutions Can Do: https://www.datavisor.com/blog/financial-indicators-of-human-trafficking-what-financial-institutions-can-do

Account Takeover in Life Insurance: How AI Is Protecting Policyholders: https://www.datavisor.com/blog/account-takeover-in-life-insurance-how-ai-is-protecting-policyholders

5 Fraud Prevention Lessons You Can Learn from Online Marketplaces: https://www.datavisor.com/blog/fraud-prevention-lessons-from-online-marketplaces

Get Ready for Nacha 2026: New ACH Fraud Rules and Strategies for Compliance: https://www.datavisor.com/intelligence-center/webinars/defend-webinar-get-ready-for-nacha-2026-new-ach-fraud-rules-and-strategies-for-compliance

AI Agents for Fraud & AML: A New Era of Detection and Strategy Optimization: https://us06web.zoom.us/webinar/register/WN_hI98x2u0Ry6-jzeYYn3Qpw#/registration

Chapters:

00:00 Introduction
02:30 Fraud landscape shift
06:00 Mythos & AI threat
11:30 FTC debanking crackdown
17:00 Federal fraud task force
22:30 Closing thoughts

In this Analysis, we examine how online marketplaces act as early-warning laboratories for financial crime, drawing on two decades of industry experience to show why the tactics refined there inevitably migrate to banks and fintechs.

This episode explores:

• The Economics of Abuse: Why the most effective defense is simply increasing attacker costs, forcing "lazy and cheap" fraudsters to spend more time and resources than an attack is worth.

• The Power of Delayed Feedback: How "soft blocks" and strategic delays can deceive attackers into thinking their tactics worked, preventing them from quickly iterating on failed attempts.

• Risk-Based Orchestration: Strategies for applying friction only at high-stakes moments, like payouts or unusual login patterns, to balance aggressive growth with essential security.

Read the full analysis and related research: https://www.datavisor.com/blog/fraud-prevention-lessons-from-online-marketplaces

DEFEND Training: https://www.datavisor.com/defend-training

Chapters:

01:10 Marketplaces: The Fraud Laboratory

03:44 Lesson 1: Economics: Make Fraud Expensive

06:24 Lesson 2: Soft Blocks and Delayed Feedback

09:55 Lesson 3: Solving the Trust-Growth Paradox

13:48 Lesson 4: Dynamic Risk Tiering Strategies

17:19 Lesson 5: Strategic Verification Testing

In this Case File, we break down how multiple organizations across payments, travel, and financial services addressed large-scale coordinated fraud, ranging from from brute-force card attacks and synthetic identities to promotional abuse and B2B payment fraud, and, the approach they took using unsupervised machine learning, and the outcomes they achieved.

You’ll learn:

• The hidden fraud patterns traditional systems missed

• How coordinated attacks were exposed in real time

• The measurable impact on fraud losses and operational efficiency

Read the full case study: https://www.datavisor.com/intelligence-center/case-studies/5-case-studies-uml-real-time-fraud-prevention

DEFEND Training: https://www.datavisor.com/defend-training

Chapters:
00:00 Intro: What the F Happened?

01:22 The Power of UML over Rules

03:31 Case 1: Brute Force BIN Attacks

06:34 Case 2: Synthetic Identity & BNPL

09:15 Case 3: Promo & Reward Abuse

12:41 Case 4: Airline Seat Hostaging

15:44 Case 5: B2B Sleeper Cell Rings

20:10 Rigid Bots vs. Chaotic Humans