Oscilar

Heading to Vegas for Fintech Meetup? Us too. Catch Neha Narkhede, Saurabh Bajaj, Anurag Chowdhury, and Team Oscilar in person. Oscilar @ Fintech Meetup 🔥 Mar 30 – Apr 1 → Booth #127 — stop by or book some 1:1 time Tues, Mar 31 • 8:50 AM Panel: Can AI fight fraud faster than fraud fights back? Neha Narkhede (Oscilar) | Chris Dorsey (Capital One) | David Birch (15Mb Ltd) Justin Keene, Ph.D. (Moveris) | Laura Spiekerman (Alloy) Side events (small groups, real conversations): Mon, Mar 30 • 7:30 P.M. - Executive Dinner Tues, Mar 31 • 8:00 A.M. - Risk & Fraud Breakfast (with iDENTIFY & FS Vector) Tues, Mar 31 • 6:30 P.M. - Lending Dinner (with Spinwheel & American Fintech Council) Spots are very limited, act now if you want in. 👇 Learn more, book 1:1 time, or request a seat — link in the comments See you in Vegas. #FinTechMeetup2026

Three years ago, I believed the hardest problem in financial risk was data. Get the signals right, wire them together fast enough, and you could build a system that outpaced fraud networks. I was partially right. Data matters enormously. But I've changed my mind about what the core constraint actually is. It's not data. It's time-to-action. The best fraud signal in the world is worthless if your risk team can't operationalize it in hours rather than quarters. I've seen banks sitting on model insights that take nine months to reach production. By then, the attack vector has evolved several times. What I've learned from watching Oscilar's customers: → The institutions winning at #fraud are the ones who've decoupled their risk logic from their core systems — so they can move at network speed, not vendor speed. → The ones struggling have the same data. They just can't act on it. → AI-native infrastructure isn't an upgrade. It's a different premise: that risk decisions should be made in real time, by systems that learn from every decision. The future of risk isn't just about better data. It's about collapsing the distance between signal and action. That's what keeps me building.

Today! 2 PM ET. ⚡

Last week, IBM finalized its acquisition of Confluent. Confluent is now the backbone of enterprise AI data infrastructure for one of the most storied technology companies in the world. I’ve had a unique front row seat to this journey. First as someone who understood exactly what Neha Narkhede was trying to build, and now as her co-founder at Oscilar. What has always stood out is the consistency of her instinct. Neha co-created Apache Kafka at LinkedIn as an answer to a broken system. A real frustration felt by engineers working at the limits of infrastructure that couldn't keep up with its own data. Fix the backbone, and everything built on top of it gets better. Kafka became the open-source standard for real-time data streaming, now powering over 80% of the Fortune 100. But open source only goes so far. Confluent was the next step: taking what Kafka made possible and turning it into a fully managed, enterprise platform. It gave organizations access to real-time data infrastructure without years of internal engineering investment. Confluent went public in 2021, transforming an internal engineering solution into a category-defining platform that the JPMCs and BMWs of the world, along with 5,000+ other companies, depend on every day. Now IBM is betting $11 billion that Confluent will power the next era of enterprise AI, because getting data to AI agents instantly, wherever it lives, is a foundational problem of this decade. That alone is a remarkable arc. But if you know Neha, you know she’s a builder at heart. The work is never done. After Confluent, the next question kept nagging at Neha: if data is moving in real time, why are decisions still fragmented and slow? You see this most clearly in financial services. With Kafka and Confluent, banks and fintechs now have real-time data streams, but their risk systems are anything but real time, fragmented across onboarding, fraud, credit, and compliance. Signals exist, but they are not connected. Decisions are still slow, siloed, and often made without full context. At the same time, everything has changed. Financial crime has scaled into a $4.4 trillion global economy, with fraud networks that are faster, more coordinated, and powered by AI. The attackers operate in real time. The systems defending against them do not. That gap is what Neha kept coming back to. And that question is why Oscilar exists. If Kafka and Confluent built the infrastructure to move data, Oscilar is focused on what happens next: turning thousands of signals across identity, behavior, transactions, and networks into the right decision in milliseconds, before the money moves. Watching IBM close this deal only reinforces the conviction we built this company on. Real-time data only matters if you can act on it.

Tomorrow: If you're in sponsor banking or risk oversight, this is a conversation you don't want to miss. ⚡ Join leaders from Oscilar, iDENTIFY, TransPecos Banks, SSB, and the Financial Fraud Consortium as they explore how better data ownership is rewriting the rules of risk in sponsor banking: strengthening oversight, improving fraud and AML monitoring, reducing false positives, and powering AI models that actually work. Learn how TransPecos Banks built structure before scale by bringing data, visibility, and control in-house. 🗓️ Wednesday, March 25 (tomorrow!) ⏱️ 2:00 PM ET 🔗 Link to register in comments Lee Easton I John Walsh I Dub Sutherland I Seth S. I Andrea Valentin

Barte adota IA da Oscilar e reduz detecção de fraudes de semanas para minutos A Barte, plataforma de pagamentos e serviços financeiros de São Paulo que atende mais de 5 mil empresas, adotou a plataforma de decisão de risco com IA da Oscilar para transformar suas operações de fraude e compliance. A integração reduziu o tempo de ajuste em políticas antifraude de semanas para minutos, com decisões de risco rodando em menos de 100 milissegundos. A Barte integrou dados de risco transacional da Mastercard Identity diretamente à plataforma da Oscilar. As decisões rodam em paralelo para atender às demandas de baixa latência dos pagamentos, bloqueando fraude sem adicionar fricção no checkout ou no onboarding. “Um dos objetivos da Mastercard é garantir que segurança nunca seja obstáculo para inovação”, disse Marcelo Tangioni, presidente da Mastercard Brasil. A parceria tripartite entre Barte, Oscilar e Mastercard combina dados globais de risco com modelos de IA locais, criando uma camada de proteção que seria difícil de implementar isoladamente. 🟡 Inscreva-se na nossa newsletter e fique por dentro de todas as notícias do mundo das finanças: https://lnkd.in/dgAm3JiY 📌 Matéria completa: https://lnkd.in/d_puc89y

Mule networks are the circulatory system of financial crime. Every scam, every fraud ring, every laundering operation needs a way to move and extract funds. According to Europol, more than 90% of all money mule transactions are directly linked to cybercrime. Disrupt the mule layer and you disrupt the economics of financial crime at scale. The problem: mules don't look like fraudsters at the point of account opening. They pass KYC. They behave normally for weeks or months. Then they activate. Traditional fraud systems weren't built to catch this. They evaluate transactions in isolation, flag individual anomalies, and miss the network entirely. Here's what actually works: → Graph analytics. Mule networks share infrastructure: linked accounts, devices, common IP addresses, overlapping behavioral patterns. Graph analysis reveals connections that transaction monitoring misses entirely. The signal isn't in any single account. It's in the relationships between them. → Velocity correlation. Coordinated cash-out rings create simultaneous velocity spikes across multiple accounts. Detecting these patterns requires seeing the network, not just the individual transaction. → Lifecycle context. Mule accounts show subtle signals over time: dormancy patterns, sudden reactivation, routing changes, first-time recipient activity. Catching them requires correlating signals across the entire customer journey, not just the moment of suspicion. → Real-time detection. Once funds hit a mule account, they're often withdrawn within hours. Batch processing is too slow. You need to flag coordinated activity as it's happening, not after the fact. This is what Oscilar's Network Intelligence was built to do: graph analytics that reveal linked accounts, shared devices, and velocity spikes in real time, exposing coordinated cash-out rings before funds leave the system. In early deployments, banking customers have seen meaningful reductions in both fraud ops costs and investigation time. Mule detection isn't just a feature. It's the leverage point.