The AI company quietly deciding whether your payment goes through - in milliseconds.
FEEDZAI, EST. 2011. Three founders, one Portuguese university town, and a bet that machine learning could outrun fraud. Today its models sit behind some of the world's largest banks.
Every second, somewhere in the world, a payment is quietly approved or declined. Behind a growing share of those decisions sits a company most people have never heard of. Feedzai builds the machine-learning systems that banks use to decide, in the moment a card is tapped or a transfer is sent, whether a transaction is trustworthy - or the opening move of fraud.
Founded in 2011 in Coimbra, one of Europe's oldest university towns, Feedzai grew out of a simple observation shared by its three founders: legacy, rule-based fraud systems could not match the speed or complexity of modern digital crime. Rules are brittle. Fraud adapts in hours. The founders bet that machine learning, trained on real transaction data and scoring events in real time, could keep pace where static rules could not.
That bet has aged well. The company describes its product as the world's first "RiskOps" platform - a single system that folds fraud detection, anti-money laundering, transaction monitoring, identity verification and scam prevention into one continuous view of risk across the customer lifecycle. The work is unusual in that success looks like nothing at all: no fraud loss, no false decline, no friction for the legitimate customer. Feedzai's engineers spend their days engineering absence.
"Success depends on AI that can adapt as quickly as fraud evolves."
A unified, AI-based platform for managing fraud and financial-crime risk across onboarding, transactions and the full customer lifecycle.
Real-time machine-learning scoring of payments to detect and block fraud across channels.
Transaction monitoring, risk scoring, watchlist screening and regulatory reporting for compliance teams.
Generative-AI tool that helps customers recognize and avoid scams in real time.
A tabular foundation model purpose-built for financial risk decisioning across fraud, AML and beyond.
Behavioral biometrics and identity checks that secure onboarding and cut account-opening fraud.
Feedzai's founding difference was philosophical: choose adaptive machine learning over static rules, a decade before it became the industry default. That head start shows in how the platform is packaged - not as a point tool for one type of fraud, but as a single risk fabric spanning fraud, AML and identity.
Its newest differentiator is RiskFM, unveiled in 2026 and described as the industry's first tabular foundation model built for financial data. The pitch: match bespoke, hand-tuned models out of the box, and surpass them once trained across multiple institutions and geographies - shrinking the time, cost and manual feature-engineering needed to deploy.
The company also competes on precision. In fraud detection, blocking a real customer is its own kind of failure; lose them once and you may lose them for good. Feedzai frames its edge as catching more crime while turning away fewer good people - growth and trust in the same decision.
Against rivals like Featurespace, NICE Actimize, FICO Falcon, SAS, ComplyAdvantage and Nasdaq Verafin, Feedzai's positioning is breadth plus real-time AI - and, increasingly, a foundation-model approach that few competitors have matched.
Bars are relative to the largest disclosed round ($200M). 2025 round pushed valuation past $2B. Backers over time include Sapphire Ventures, Citi Ventures, Oak HC/FT, KKR, and in 2025 Lince Capital, Iberis Capital and Explorer Investments.
B2B enterprise SaaS. Feedzai licenses its RiskOps platform to banks, processors, acquirers and merchants through multi-year subscriptions, often tied to transaction volumes.
Nuno Sebastião, Pedro Bizarro and Paulo Marques launch Feedzai to apply machine learning to fraud detection.
Feedzai begins selling in the United States and establishes a Silicon Valley presence in California.
Citi Ventures backs Feedzai, beginning a long strategic relationship with the bank.
A $200M Series D pushes Feedzai's valuation above $1 billion.
Feedzai acquires Demyst, raises ~$75M at a $2B valuation, and is selected by the ECB to help safeguard the digital euro.
Feedzai unveils RiskFM, a tabular foundation model built for financial risk decisioning.
Co-founder and CEO. Previously a systems engineer at the European Space Agency before turning to fraud.
Co-founder with an academic background in high-performance computing and data-stream processing.
Co-founder and former CTO, rooted in academic computer science and large-scale systems.
"Feedzai is the world's first RiskOps platform, protecting people and payments with a comprehensive suite of AI-based solutions."
Explore Feedzai's own channel for product walkthroughs, customer stories and talks from its team.
Feedzai builds an AI-based platform that helps banks detect and prevent fraud and financial crime in real time - covering fraud detection, anti-money laundering, transaction monitoring, identity verification and scam prevention.
It was founded in 2011 in Portugal by Nuno Sebastião, Pedro Bizarro and Paulo Marques.
Coimbra, Portugal, with US operations in San Mateo, California, and a UK presence.
More than $2 billion, reached in October 2025 after a roughly $75 million funding round.
Large banks and financial institutions including Citi, Lloyds Banking Group and Standard Chartered, plus payment processors, acquirers and merchants.
• feedzai.com
• LinkedIn
• Twitter / X
• Facebook
• Instagram
• YouTube
• GitHub