Beyond the buzzwords: Key takeaways from Money20/20 Europe

My colleague Djoko Bodiroga and I are fresh from Money20/20 Europe in Amsterdam, and amid the usual noise and spectacle, a few themes cut through – not just as passing trends, but as directional signals for the future of financial services. Two, in particular, stood out: the accelerating push toward real-time financial infrastructure, and the rise of agentic AI systems – autonomous models not just assisting, but acting.

Open finance & real-time payments: from regulation to rails

The dialogue around PSD3 (Payment Services Directive 3) and instant payments felt more grounded this year – less “vision,” more imminent infrastructure shift.

Unlike the existing PSD2 framework, which suffers from inconsistent implementation, poor UX for consent flows, and delays caused by bank non-compliance, PSD3 aims to fix these shortcomings.

The draft legislation introduces:

• more flexible exemptions and improved UX;
• standardisation and removal of fallback mechanisms;
• mandatory fraud reporting hubs and enhanced collaboration;
• a shift towards Open Finance – giving access to broader financial data beyond just payments.

Crucially, PSD3 aligns with the EU’s instant payments regulation, which requires all banks and PSPs in the Eurozone to support 24/7/365 instant credit transfers.

This is not just regulatory plumbing – it’s a catalyst.

The implications? Reduced transaction costs, better fraud detection, and new embedded models, especially around pay-by-bank and wallet orchestration. Banks that treat this as a compliance box to tick will miss the bigger picture – this is about user experience, competitive positioning, and long-term revenue play.

At Vega IT, we are seeing this firsthand in our work with clients modernising real-time platforms. The interest is no longer hypothetical - it is implementation driven.

Agentic AI: when the system acts

The second major shift was more subtle, but equally profound: the rise of agentic AI – systems that not only learn, but act independently across financial workflows.

This is not just a chatbot trend. We saw demos (Klarna stood out) of LLM-powered assistants resolving entire support journeys, with escalation to humans happening only when edge-case thresholds were hit. No hand-holding, just execution.

In fraud detection, agentic models are already identifying anomalies, freezing accounts, triggering compliance workflows, and escalating to legal or audit – all autonomously.

This is no longer experimentation. It is embedded. And it raises interesting questions – not just about performance, but about trust, oversight, and the evolving role of humans in financial operations.

At Vega IT, we’ve been helping clients explore where AI adds real operational leverage. In some cases, it is about efficiency. In others, it is about enabling new business models. But in all cases, the systems need to be intelligent and accountable.

Agentic AI is no longer experimentation. It is embedded.

Let’s continue the conversation

Money20/20 is always a good temperature check. And this year, the temperature feels higher – urgency is replacing inertia in many areas of financial services. These changes are already a reality, not the future, and they impose important demands for transparency and accountability.

If you are thinking through the implications of instant payments, open finance, or autonomous AI systems, and how to build for them, let’s talk.