DataVisor

DataVisor is a Mountain View, California company that develops an AI-based platform for fraud detection and anti-money-laundering (AML) decisioning, marketed as a combined "FRAML" platform. The system pairs patented unsupervised machine learning with supervised models, a rules engine, a knowledge graph, device and behavioral intelligence, a feature store, and case management, so that institutions can detect both known and previously unseen attack patterns. Documented use cases include ACH and wire fraud, real-time payments fraud, account onboarding and application fraud, account takeover, card fraud, check fraud, promotion and policy abuse, and AML/BSA compliance.

The platform is delivered as cloud-native, multi-tenant infrastructure with data isolation across business units and geographies. The company reports processing more than $30 billion in events annually, sustaining over 15,000 queries per second in production, and computing features and scores in under 100 milliseconds. A newer layer of conversational AI agents, branded Vera, automates alert triage, rule tuning and SAR/CTR narrative generation with human oversight. Forbes describes the technology as flagging fraudulent payments, wire transfers and loan applications, with particular emphasis on linking users to expose coordinated fraud rings, and blocking or delaying suspicious transactions in milliseconds.

Alongside the product, DataVisor runs a customer and industry community program called DEFEND, comprising an annual conference, training sessions, and recurring webinars and forums on emerging fraud and AML threats. It also publishes research and benchmark material such as an annual fraud and AML executive report.

Founding story

Co-founders Yinglian Xie and Fang Yu immigrated from China to the United States, earned computer science PhDs there, and worked together at Microsoft for seven years before founding DataVisor in 2013 to apply machine learning to detecting coordinated fraud campaigns.

Business model

DataVisor sells enterprise software to financial institutions and digital platforms, deployed as a cloud-native, multi-tenant platform that integrates with customers' existing systems and third-party KYC/KYB, sanctions screening and identity verification providers. Go-to-market includes a direct sales organization, demo requests, and partner and system-integrator integrations.

Enterprise software sold to banks, credit unions, fintechs, payment companies, insurers and other digital enterprises; Forbes reports 2024 revenue of $50 million, up 67% year over year, across roughly 50 customers.

Traction

Published case studies report outcomes such as a 40% reduction in manual review alert volume and 85% reduction in manual review time at NASA Federal Credit Union; 20% higher fraud detection with 94% accuracy and 5x faster model development at a global card network; 60x operational efficiency gain and 25x faster response times at TaskRabbit; a 41% hurt-ratio reduction and 320+ fraud rings detected at a BNPL provider; 40% fewer false positives and a 50% higher fraud capture rate at a US card issuer; 92% of fake accounts detected before loss at a crypto exchange; and $3 million per month in fraudulent loans prevented. Forbes reports $50 million in 2024 revenue, 67% growth and 10 customers added that year.

Latest developments

Recent activity includes the launch of conversational AI agents for fraud and AML (Vera) and multi-agent orchestration for investigation and reporting, publication of a 2025 fraud and AML executive report, content referencing Forrester Wave results and OSFI 2026 compliance for Canadian teams, a May 2026 CEO transition from the co-founder to Patrick Harr, and a June 2026 announcement that Tru Cooperative Bank selected DataVisor for real-time fraud prevention.