Revolut has launched a research division to develop AI for fraud detection, credit risk and product recommendations.
Revolut Research sits within the company’s AI department and will oversee work on machine learning models for financial services.
It also plans to collaborate with academic and technology institutions.
The unit is developing PRAGMA, a family of foundation models built using NVIDIA technology.
The models provide a shared intelligence layer for functions including risk assessment and platform operations.
According to an NVIDIA case study, PRAGMA was trained on about 26 million user records from 111 countries.
The dataset covered between 24 billion and 40 billion events over roughly 28 months.
Tests using historical data indicated that PRAGMA was 2.3 times better at identifying credit default risk than existing benchmarks.
Revolut reported that it also detected 65% more fraud cases with 17% greater precision and delivered 41% more relevant product recommendations.
“Revolut Research has been established to responsibly build financial intelligence at its deepest layer, rather than patching together narrow, specialised models. 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.”
said Anton Repushko, Head of Revolut Research.
The division plans to regularly publish its findings and release technical frameworks as open-source software.
Featured image: Edited by Fintech News Singapore, based on image by Revolut
