By Jim Kent

Some financial actors are already all in. Certain banks now employ up to 250 AI programmers within the EU alone, integrating machine learning into everything from trading models to compliance. Others are still testing the water, identifying which internal processes are ripe for automation. Global consultancies may call for a complete business overhaul, but the truth is: progress will come step by step, not in one great leap.

RegTech Rises in Luxembourg

Nowhere is this more visible than in Luxembourg’s RegTech scene, which has expanded rapidly — you can see it in the growing number of exhibitors at LPEA tech events. Measuring the exact size of the sector, or identifying future winners, remains difficult. But the players tend to fall into two camps:

• established IT providers, some with PSF status, and
• agile startups fighting to scale up.

Driving much of this growth is the wave of regulation from the European Commission and the CSSF. Compliance and client onboarding — both highly document-intensive and data-heavy — have become ideal entry points for automation. These are natural candidates for transformation through GenAI and the emerging Agentic AI models.

For fund houses without in-house expertise, this will become a growing area of spend. Meanwhile, IT providers tackling these challenges internally gain valuable experience they can bring directly to clients — and they all agree on one point: skilled AI programmers are in short supply.

Internal LLMs — A Natural First Step

Many organisations are starting with internal AI. The most common first project? Building an LLM trained on company data, using Retrieval-Augmented Generation (RAG) to extract insights from repositories of PDFs and Word documents.

The goal is clear: turn scattered information into actionable intelligence. The results can be impressive, though they demand careful data labelling, fine-tuning, and internal Q&A guidelines to prevent model drift. When done right, GenAI excels at absorbing regulation and mapping it against internal policies. In theory, any firm could download the GDPR or EU AI Act, cross-reference it with its own rules, and instantly highlight compliance gaps.

Beyond Rules: AI in AML

The next logical step for many IT service providers is anti-money-laundering (AML). AI enables the shift from simple rule-based monitoring to intelligent detection. Machine learning

models now analyse entire datasets, spotting anomalies, flagging suspicious behaviour, and even predicting emerging risks before they become breaches.

Trusted Platforms or Tailor-Made Solutions?

When selecting external partners, the old dilemma still applies: Do you buy into a trusted platform that solves specific needs quickly, or commission a bespoke solution that fits you perfectly — at a price?

Ian Atkinson, CEO of Governance.com commented “The optimal approach to AI in finance is a trusted, mature platform that can be configured to specific workflows, offering the best of both worlds: industrial-grade reliability with tailored application. This is our exact philosophy at Governance.com and for the LHoFT’s Mutualisation Catapult programme, where we are extending our existing, market-proven governance platform to create a shared utility. While our sovereign AI engine will ingest, transform and structure data, the final decision, and the accountability that comes with it, must always remain with a human expert.”

Not all agree and opt for complete customisation. Custom builds offer flexibility but come with higher costs and long service contracts. Platforms deliver faster results but can feel limited and demand a steep learning curve for staff. Just ask a bank that had its core banking system built in the 1980s and is still using it today.

The Road Ahead

AI will not replace RegTech — it will redefine it. The firms that combine domain expertise, technological agility, and ethical AI governance will not just survive this shift; they’ll become indispensable partners in the next generation of compliant finance.