Sardine has unveiled Sardine AI Labs, a dedicated applied research unit, alongside $375,000 in research fellowships aimed at pushing frontier AI into fraud and financial crime prevention.
The new lab will examine how AI models can interpret financial behaviour as a form of language, predict emerging attack methods and support risk teams in reaching sounder decisions.
As part of the launch, Sardine has invited applications for as many as five fellowships, which will go to independent researchers able to build on the lab’s initial findings.
Sardine argues that the next breakthroughs in foundation models will emerge from specialist AI neolabs that hold proprietary, real-world data. Fraud prevention illustrates the point, as even prominent academics frequently struggle to obtain datasets large enough to move the discipline forward.
The company views payment and account access activity as carrying recognisable patterns, with transaction records, devices and IP addresses combining into behavioural profiles that foundation models can examine to spot irregularities and flag possible fraud.
Financial crime prevention continues to weigh heavily on financial institutions, pushing up processing costs across card, ACH and wire payments. Sanctions screening is one persistent pain point: individuals whose names or identities resemble entries on sanctions lists can be wrongly flagged, leaving their onboarding stalled for weeks during manual checks.
Establishing whether two records belong to the same person remains difficult. Spotting money laundering, terrorist financing and drug-related payments among billions of bank transactions is another, with current tools often holding up genuine payments through false positives while missing illicit flows, leaving institutions open to serious regulatory and financial penalties.
The lab’s research agenda covers modelling extremely long sequences of transaction and user events, transfer learning, adversarial robustness and explainability. Drawing on Sardine’s network of device, identity, behavioural and transaction data, the team is building models designed to grasp complex financial behaviour and catch attacks they were never specifically trained on.
Sardine AI Labs has already released early results from a transformer-based foundation model that learns from full cardholder transaction histories. The model was trained on roughly one billion transactions gathered over two years from more than a dozen card issuers. It targets the cold-start problem, where newly launched issuers lack the data to detect fraud dependably yet are frequently singled out by fraud rings probing fresh products.
When tested on issuers left entirely out of its training data, the model lifted fraud detection accuracy by 68% for a consumer card issuer and 41% for a business card issuer versus conventional machine learning methods. These gains held across consumer, business and global cross-border issuers, suggesting the model can generalise across varied card programmes.
Sardine CEO and co-founder Soups Ranjan said, “We are uniquely positioned to train a model purpose-built for risk because we sit on the industry’s fastest growing fraud and fincrime dataset, which spans more than 6.5B devices, 441M consumers, 3.4M businesses, 6.6 billion transactions and $1.8 trillion in payments. Sardine AI Labs will focus on production-grade models that meet real world latency, governance, and explainability requirements.”
Sardine head of data science Niranjan Shetty said, “The most important finding is that foundation models can learn directly from a user’s transaction behavior, allowing us to identify fraud more accurately than approaches that reduce that behavior to a set of tabular features,” “Because these patterns transfer across financial institutions, the model is not limited to a single card program. The next step is to make this intelligence fast, explainable, and reliable enough to support real-world risk decisions.”
Ranjan added, “The frontier AI labs have shown the immense benefits of having multi-modal AI models which allow you to extract intelligence across different modalities of audio, video and text.
“We think that the next unlock in fraud and fincrime prevention would occur from creating a multi-modal AI that goes across device, identity, behavioral, and transaction data. This is what excites me the most about the potential ahead with our AI Lab and I can’t wait to see what our research fellows build with us,”
Copyright © 2026 FinTech Global
