Introduction
Feedzai builds AI-based software that banks and payment processors use to detect fraud, screen for money laundering, and manage identity risk across a customer’s entire relationship with a financial institution. Founded in 2011 in Coimbra, Portugal, by Nuno Sebastião, Pedro Bizarro, and Paulo Marques, the company grew out of a research partnership with Carnegie Mellon University before expanding into the U.S. market.
Feedzai’s central pitch is consolidation: rather than a bank running separate tools for card fraud, account takeover, and anti-money-laundering monitoring, its RiskOps platform tries to bring those functions into one system. That positioning is a useful window into how larger financial institutions are rethinking their AI risk infrastructure, not just which individual model performs best.
What the company does
Feedzai’s RiskOps platform applies machine learning and behavioral analytics to transaction and account data in real time, aiming to catch fraud, flag suspicious money movement for anti-money-laundering (AML) review, and assess identity risk during onboarding — all within a single cloud-based system rather than a collection of separate point tools. The company describes its approach as combining specialized AI models with input from a global team of behavioral analysts and data scientists.
Who it serves
Feedzai’s customers are financial institutions rather than individual consumers: banks, card issuers, and payment processors that need to monitor transaction volumes far too large for manual review. The company states on its own website that it processes tens of billions of events and secures several trillion dollars in payments annually on behalf of its customer base, figures that reflect the scale of a fraud-detection vendor operating behind the scenes of everyday card and account activity.
Company background
Feedzai was founded in 2011 in Coimbra, Portugal, by Nuno Sebastião, Pedro Bizarro, and Paulo Marques. According to the Carnegie Mellon Portugal partnership, which supported its founding, Feedzai was the first startup launched under the CMU Portugal program and later became the program’s second company to reach a private unicorn valuation. The company expanded into the United States in 2014, establishing a second headquarters in San Mateo, California, and now operates as a group with entities in Portugal, the U.S., and the U.K. Nuno Sebastião continues to serve as the company’s CEO and chairman.
Product and AI capabilities
Feedzai’s RiskOps platform is built to unify what were traditionally separate systems — fraud detection, identity verification, and AML transaction monitoring — into shared infrastructure, so a bank’s risk team is not maintaining several disconnected tools that each see only part of a customer’s activity. The company has also developed Fairband, a framework it describes as helping financial institutions test whether their AI risk models produce disparate outcomes across customer groups, an issue that regulated fraud and credit models increasingly have to account for. On its own website, Feedzai states that it processes more than 90 billion events and secures roughly $9 trillion in payments annually for its customers, and that it prevented $1 billion in fraud losses in 2025.
Key developments
Feedzai was founded in 2011 in Coimbra, Portugal, growing out of the Carnegie Mellon Portugal research partnership. It expanded to the United States in 2014, opening operations in San Mateo, California. The company went on to reach a private unicorn valuation, recognized by the CMU Portugal program as its second unicorn spinout, and later introduced its Fairband model-fairness framework. More recently, Feedzai consolidated its fraud, identity, and AML products under a single RiskOps platform, a repositioning the company frames as a response to banks wanting fewer, more connected risk tools rather than more point solutions.
Why it matters
Feedzai is a useful example of how AI adoption in bank risk management is moving from isolated fraud-detection models toward integrated platforms that share data and signals across fraud, identity, and AML functions. That consolidation trend matters because financial crime increasingly spans categories — a single bad actor might combine a stolen identity, a synthetic account, and a laundering scheme — and tools that only see one slice of that activity can miss the pattern. Feedzai’s Fairband framework is also a concrete example of a fraud-technology vendor building fairness testing into its own products rather than treating it as a separate compliance exercise.
Sector context
Within Brel’s AI in finance coverage, Feedzai sits alongside other bank-facing fraud and identity vendors such as Socure and Alloy, though its RiskOps positioning places more emphasis on unifying fraud, identity, and AML than on identity verification alone. It is also a useful contrast to payments-network fraud tools such as those built into Stripe, since Feedzai sells its platform to banks and processors directly rather than bundling it with a payments product.
Sources and references
This profile draws on Feedzai’s official RiskOps product page, independent reference sources, and the Carnegie Mellon Portugal program’s account of the company’s founding.
- Feedzai — “RiskOps Platform” official product page (feedzai.com)
- Wikipedia — “Feedzai”
- CMU Portugal — Feedzai unicorn spinout profile