With 8 billion euros in deposits, Bunq, sometimes referred to as "the bank of the free," provides over 12 million users with a fully online banking experience.

The collaboration attempts to meet the growing need for sophisticated fraud detection techniques. The automated transaction-monitoring system from Bunq greatly improves the speed and precision of identifying fraudulent activity and money laundering thanks to NVIDIA's accelerated processing technology.

NVIDIA is an expert in GPU-accelerated computing, which improves AI model training and data processing. Using the RAPIDS package, which is part of their AI Enterprise software platform, bunq was able to train its fraud detection model about 100 times faster than it could have with other techniques. Because of this partnership, bunq can handle more datasets more quickly, increasing model accuracy and lowering false positives.

After more than thirty years, Bunq became the first new bank to be granted a banking license in Europe when it was created in 2012. It does away with the necessity for in-person bank visits by providing a variety of online banking services via its app. Bunq hopes to expand its fraud detection system and boost overall operational efficiency by using NVIDIA's AI technology.

Bunq is also investigating AI applications for all aspects of its business. Head of data and AI at bunq Ali el Hassouni explained how AI is used to handle user issues, identify phony IDs during onboarding, and automate marketing campaigns. Utilizing bunq's unique broad language model, the personal AI assistant Finn responds to user inquiries and improves the clientele's experience.

Additionally, Bunq is testing a set of generative AI microservices called NVIDIA NeMo Retriever in an effort to boost customer support and increase Finn's accuracy.

According to Ali el Hassouni, "this is a great example of how human intelligence can be coupled with AI. AI has enormous potential to help humanity in so many ways."

"We chose to use larger datasets and speed the training of new models — sometimes by an order of magnitude — resulting in improved model accuracy and reduced false positives," he continued. "NVIDIA's sophisticated, GPU-optimized software is what we used."

El Hassouni said, "We're constantly looking for new ways to apply AI for the benefit of our users," in reference to bunq's larger AI ambitions.