AI in Luxembourg finance: real use cases delivering ROI

Luxembourg's financial industry has moved beyond abstract discussions about artificial intelligence. At Nexus Luxembourg’s panel 'From Hype to ROI: Real AI Use Cases in Financial Applications' on 10 June, moderated by Dr Amal Tawakuli, Data & AI Technology Strategist at NTT Data, experts confirmed the sector is now in an execution phase — focused on tangible results and overcoming strategic hurdles. 

Prioritising AI use cases in financial services

The primary challenge is no longer a lack of ideas but how to act on them. Patrycja Garscia, Business Relationship Advisor for Data & AI Finance at Luxinnovation, explained that financial institutions have "an extensive portfolio of use cases" ready to go, including intelligent document processing, client onboarding, know your customer (KYC) and anti-money laundering (AML) transaction monitoring. The real difficulty, she noted, lies in "how to prioritise them and to have a good execution". 

Drawing on lessons from previous technology waves like digital payments and e-wallets, Daniel Steinhauf, Managing Director of Financial Navigator by A352, cautioned against hype. He stressed the importance of finding the "sweet spot for your business" by carefully assessing "what is your business currently doing, what is the new technology doing and how you can bring both together". He warned that AI is like an "academic student coming from the university" who "doesn't really know what you're doing". To achieve real value, businesses must "bring the proper context to AI" and teach it the specifics of their operations. 

AI delivering measurable ROI in treasury and insurance

This practical approach is already yielding results. Mr. Steinhauf detailed how his firm’s treasury management platform, Financial Navigator, uses AI to deliver tangible benefits. The platform aggregates all of a company's financial information into a single dashboard. AI enhances this by providing "very precise forecasts" on cash positions, learning from past data and even factoring in news that could impact the business. Furthermore, AI helps with risk monitoring, identifying potential fraud or suspicious transactions that rule-based systems might miss. 

At insurance leader Foyer Group, AI is enabling long-stalled projects to move forward. CIO Remy Els revealed that his team is now "developing prototypes for business ideas that were under the work for several years". He also highlighted a clear focus on "refining our software delivery process" through the use of AI-assisted coding, which promises to fundamentally change how software is developed and deployed. 

Data sovereignty: a blueprint for secure AI in finance

A key concern for the heavily regulated sector is data sovereignty. A landmark proof-of-concept between Foyer Group and LuxProvide offers a powerful local solution. Filipe Pais, CCSO of LuxProvide, explained the project's goal was to allow Foyer to "benchmark some models for a specific use case" on MeluXina, a "sovereign infrastructure that you can put trust on. The most important is keeping your data where you control it, keeping your models where you have them and understanding who is handling that." 

Mr. Els added that the proof of concept was a crucial opportunity to "evaluate the combination of two partners" in a secure environment. This collaboration serves as a blueprint for innovating responsibly while keeping sensitive data and intellectual property under control within European borders. 

Luxembourg AI Factory: supporting AI adoption at scale

The shift towards practical AI application is actively supported by the Luxembourg AI Factory, a national initiative that offers a suite of programmes designed to demystify AI adoption for businesses. As Mrs. Garscia explained, the factory currently provides "66 services developed in our service catalogue" to help companies de-risk their projects, upskill their teams and develop market-ready AI solutions. 

Mr. Pais concluded "We will invest in a new machine fully dedicated to AI workloads and later in a quantum computer — the idea is to have these three platforms interconnected, helping companies innovate using any type of technology."