Revolut launches in-house AI research unit
Revolut has created a new AI research division, Revolut Research, to develop proprietary models such as PRAGMA, which the company claims delivers significant improvements in credit risk, fraud detection and product recommendations.

Revolut has announced the creation of a dedicated AI research unit, Revolut Research, as part of its broader AI department. The new division will work with academic and technology partners to pioneer machine-learning technology for financial services.
At the core of the unit’s work is PRAGMA, a proprietary foundation model built in partnership with Nvidia. Revolut says PRAGMA is engineered to understand complex financial behaviour and to power functions such as fraud detection, credit risk assessment, platform operations and personalised product recommendations.
Revolut reported that early deployments of PRAGMA achieved the following improvements:
| Metric | Improvement |
|---|---|
| Credit risk accuracy | 2.3x better |
| Fraud detection | 65% more cases detected |
| Product recommendations | 41% more relevant |
The announcement follows UK digital bank Starling’s launch of smart tools last week, a set of crowdsourced features that will roll out weekly until the end of the year. Starling’s rollout was partly inspired by the continuous model updates of frontier large-language models.
“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’s data from 80 million customers feeds into PRAGMA, creating a self-reinforcing loop. As the dataset grows, the models are set to become increasingly sophisticated. The unit plans to publish its scientific findings and open-source technical frameworks, while engaging with the broader scientific community through international conferences and hosting meetups at Revolut’s offices. The research team will share its progress on its stats page and invite collaboration at its squad hub.
The AI arms race among Europe’s neobanks has intensified over the past year. Starling kicked off the competition in June with a spending-intelligence tool that let customers ask AI about their spending in plain English. Dutch rival Bunq responded in December by upgrading its AI customer-support assistant, and by March Starling launched what it billed as the UK’s first “agentic” AI assistant. Revolut followed with its own assistant in April.
Revolut’s AI research unit is positioned to accelerate the deployment of smarter features, reduce systemic friction and enhance the safety of its customers’ financial experience.





