-
Revolut Research is the institutional engine behind PRAGMA, Revolut’s proprietary foundation model powering real-time risk assessment, platform operations, and tailored product recommendations.
-
Building alongside leading academic and technology institutions, the dedicated research unit provides a central framework for Revolut’s proprietary AI deployments and machine learning initiatives
Dubai, United Arab Emirates, 25 August 2026: Revolut, the global financial leader with over 80m retail customers, has today announced the launch of Revolut Research. Revolut Research is a dedicated division within the broader AI Department, established to pioneer next-generation machine learning architecture across global financial services. Set to build alongside leading academic and technology institutions, the unit provides a unified foundation for Revolut’s proprietary AI deployments and machine learning initiatives.
The global banking industry is hitting a technological wall as legacy institutions attempt to retrofit decades-old infrastructure with off-the-shelf software wrappers. This approach creates both soaring operational drag for organisations and leaves consumers with clunky applications and interfaces, and basic search-bar chatbots. Revolut is building differently, embedding native intelligence, developed in-house, directly into its core financial engine, turning local market data into a compounding global advantage.
Revolut Research serves as the engine behind PRAGMA, Revolut’s proprietary foundation model developed in collaboration with NVIDIA. Engineered specifically to decode complex financial behaviours and power real-time risk assessment, platform operations, and tailored product recommendations, Revolut is rebuilding its AI stack around a single foundation with one shared behavioural intelligence layer.
Head of Revolut Research, Anton Repushko, has already presented key findings at multiple high profile international conferences this year, including ICML in Seoul. Early deployments of PRAGMA on historical data have indicated major performance gains over legacy baselines:
-
Superior Credit Risk Accuracy: 2.3x higher accuracy at identifying credit default risk.
-
Smarter Fraud Prevention: 65% more fraud cases caught, with 17% greater precision in alerts.
-
Hyper-Personalisation: 41% more relevant product recommendations across retail and business accounts.
At the heart of this technical advantage is an unmatched global dataset. Serving 80 million customers across more than 40 markets, Revolut processes billions of cross-border transactions and diverse financial behaviours in real time. This data feeds into Revolut Research’s models, creating a self-reinforcing loop and establishing a new language of financial behaviours. As the dataset grows, the models become exponentially smarter at detecting fraud, evaluating risk, and predicting user needs – creating a proprietary intelligence advantage that traditionally structured banks struggle to match.
“To lead the future of intelligent banking, you cannot rely on third-party blueprints,” said Pavel Nesterov, Head of AI at Revolut. “We have launched Revolut Research to institutionalise our ‘build, don’t bolt on’ philosophy. By training native foundation models on our global operational data, we are giving our engineering teams an unprecedented engine to deploy smarter features faster, eliminate systemic friction, and give our customers a safer, radically better financial experience.”
“Revolut Research has been established to responsibly build financial intelligence at its deepest layer, rather than patching together narrow, specialised models,” said Anton Repushko, Head of Revolut Research. “In PRAGMA, we are developing a single, unified foundation model capable of understanding the true nuance of financial behaviour in real time. Technology is in Revolut’s DNA, and by collaborating with global tech leaders, this division is engineering proprietary capabilities that set us apart from traditional banks.”
Beyond PRAGMA, Revolut’s advanced security models review nearly one billion transactions every month to stop fraudulent activity proactively, while connected engines power front-end capabilities such as AIR, Revolut’s in-app assistant that executes complex financial tasks in a single step (currently available in the UK only).
Looking ahead, Revolut Research plans to regularly publish its scientific findings and open-source technical frameworks. The Revolut Research team will also engage with the broader scientific community through international conferences, including NVIDIA GTC Berlin in October, ICAIF in November and will host quarterly meet-ups for the scientific community at Revolut’s offices.
To read the detailed technical benchmark results, explore the case study at https://www.nvidia.com/en-us/case-studies/revolut/, or view open research roles, visit revolut.com/careers.
