New LIST-BIL Initiative Aims to Set Standard for Evaluating Agentic AI in Finance

On Wednesday 23 July 2025, the Luxembourg Institute of Science and Technology (LIST) and Banque Internationale à Luxembourg (BIL) announced a partnership on an innovation project to assess AI solutions for customer service.

According to the project partners, the collaboration aims to develop advanced tools and methods for evaluating agentic AI systems in banking, with a focus on strengthening user trust and ensuring greater accountability within the industry.

The joint initiative between LIST and BIL aims to develop a rigorous, science-based methodology and software platform for evaluating agentic AI systems. According to the partners, robust assessment is crucial to building user and customer trust, particularly in high-stakes sectors such as banking. The project will address core pillars such as robustness, fairness (i.e. lack of social bias) and human oversight - described as “increasingly vital” as AI becomes more embedded in critical domains. The initiative also seeks to ensure that AI solutions are not compromised by common limitations of generative AI, such as hallucinations and bias.

“Generative AI is already transforming the banking industry,” commented Olivier Gorin, Head of Digital Banking at BIL. “Thanks to our collaboration with the experts from the LIST, we will get invaluable insights on Gen AI technology. It will help us adapt future services to clients’ needs and improve personalisation in a secure and accurate way. This project is also a prime example of our commitment to a thriving Luxembourg innovation and financial ecosystem.”

The project will leverage the LIST AI Sandbox to further advance its development and Technology Readiness Level, with the aim of enabling its application across a broader range of real-world use cases.

“This collaboration is a significant step forward in the evolution of the LIST AI Sandbox,” said Francesco Ferrero, Head of the Flagship Initiative on Artificial Intelligence at LIST. “It already serves as a platform for assessing the ethical, linguistic and technical performance of AI systems. By integrating BIL’s real-world use cases, we’re enhancing the AI Sandbox’s maturity and applicability in the financial sector - moving closer to a scalable and market-ready framework for trustworthy AI deployment in mission-critical applications and regulated industries.”

According to BIL and LIST, this project represents the first milestone in a broader collaboration aimed at promoting the responsible use of AI within BIL and the Luxembourg banking sector. The partners described it as a “replicable model” for assessing and deploying trustworthy AI. The initiative, they added, reflects both institutions’ shared goal of applying AI in ways that ensure operational efficiency, regulatory compliance and protection of customer interests.