In June, PwC Luxembourg released its 2025 (Gen)AI and Data Use Survey, the fourth edition of a study examining how organisations in Luxembourg manage, operate, and increasingly leverage data and AI. This year’s edition broadened its scope and partnerships, notably through collaboration with the Luxembourg Bankers’ Association (ABBL) and the Luxembourg Insurance and Reinsurance Association (ACA). Together, they shed light on how various industries — particularly the financial sector — are approaching both the opportunities and the challenges of GenAI (Generative AI).

Broadening the Lens

Since its launch in 2019, the PwC survey has traced the gradual shift from data management to AI adoption. But with the arrival of GenAI, the conversation has accelerated.

“Ever since ChatGPT has become, let’s say, the norm, or very much embedded in everyday life, everyone has seen what additional capabilities GenAI can offer. So, we decided it would be great to put a strong focus on AI in this year’s edition,” explains Andreas Braun, Managing Director at PwC Luxembourg. “We also wanted deeper insights into how organisations are using today’s most advanced technologies, and what they are actually using them for.”

For ABBL, participating in the study was a natural step. “It was a great opportunity for ABBL to being part of this study, as it aligns with our ongoing work to understand the market appetite for GenAI” notes Ananda Kautz, Head of Innovation, Digital Banking and Payments at ABBL. Data has clearly emerged as both the main issue and the greatest challenge for banks today, making this focus especially timely and relevant. “Gaining this cross-sector perspective was very valuable, allowing us to benchmark against other industries and deepen our understanding of the market and its challenges around data. Our survey found that overall, across sectors, only 25 % of organisations are using most of the data they collect, a striking gap between potential and practice.”

From the insurers’ side, the timing was equally relevant. “Luxembourg insurers know how to manage data — it’s at the very core of their business — but historically this use has been closely tied to regulatory requirements and actuarial work,” adds Sarah Hartmann, Legal Adviser at ACA Luxembourg. “GenAI opens the door to much broader applications, provided the industry can attract and retain the right talent.”

The Data Gap between Governance and Strategy

One of the survey’s most striking findings is the contrast between governance maturity and limited data usage: while many organisations report strong capabilities in data privacy and compliance, only 25% fully leverage the data they collect.

According to Andreas Braun, the issue stems from a lack of strategic alignment. “Organisations, particularly in the financial sector, report very high maturity when it comes to data governance and ensuring privacy. But do they really know what all their collected data is being used for? That’s where we observe a gap. While data governance is well established, the same cannot always be said for a data strategy.”

For Ananda Kautz, the problem ultimately comes down to trust. “There’s no magic. It really comes down to data quality, which determines both the results AI can deliver and the trust we can have in these tools,” she says. “The objective for banks is not to achieve 100% data quality everywhere, but to ensure 100% quality on the data points that are truly strategic. That requires a genuine AI strategy at the highest level.”

Sarah Hartmann points out another challenge: “To move towards optimised data use through GenAI, the main obstacle is the shortage of profiles able to turn this data into tangible value. Companies are increasingly investing in upskilling their workforce, but the financial centre’s ability to attract and retain these skills will be decisive.”

GenAI in Financial Services

Compared with other industries, Luxembourg’s financial sector is further along the adoption curve. The survey signals this momentum: nearly two-thirds of organisations already use GenAI tools, and over 80 % have strong governance frameworks in place.

“Staying competitive now depends on turning AI from experimentation into real impact,” stresses Ananda Kautz. “A lot of initiatives focus on efficiency and automation. We’ve seen this trend for a while now — starting with internal use cases before moving toward enhanced customer insights and innovative client-facing solutions.”

Yet challenges remain. “Interestingly, AI is not very consistently used for customer insights, which is somewhat surprising, since it’s a fairly common use case,” notes Andreas Braun. “All financial sector entities reported lower maturity in this area than the public sector. Fraud detection and cybersecurity also show relatively low maturity, with only 40% of organisations using AI for these purposes.”

For insurers, the foundations are solid but underexploited. “Insurers have always been data-driven, but mainly in an actuarial and regulatory context, relying on their own models and frameworks,” explains Sarah Hartmann. “That’s a strength, because the data is already reliable and well structured, but it doesn’t capture the full potential of GenAI. The sector is clearly more cautious than others, yet the foundations are there to deploy GenAI responsibly and, when the time is right, at scale.”

From Pilots to Execution: Next Steps

Moving from experimentation to execution is the challenge most organisations face today. Many banks have introduced productivity tools, but scaling these initiatives requires both strategy and investment.

“What we see at the moment is significant progress in internal efficiency, now embedded in most financial-sector entities,” says Ananda Kautz. “The next step is to consolidate these early gains and move toward tangible implementations, particularly in areas such as fraud detection, cybersecurity, and compliance. At ABBL, we’ve launched a new Tech and Innovation Cluster to unite C-levels, set a strategic roadmap, and drive collaboration where it matters most.”

For Andreas Braun, execution also means cultivating an AI mindset. “The bottom-up approach remains a very effective way to identify the best AI use cases within an organisation. Using tools like Copilot in day-to-day business often generates the most valuable ideas,” he explains. “That’s why investing in literacy and upskilling is so important. It’s also crucial to keep an eye on what’s happening in the AI market. Even if a project didn’t work a year ago, with GenAI and reasoning systems something new may already have emerged. Organisations need to invest more than the 1% of revenue currently dedicated to innovation if they want to ensure long-term sustainability.”

Sarah Hartmann echoes this call for pragmatism. “Many insurers are still at the pilot stage. To move to the next level, it’s essential to focus on a few concrete and pragmatic use cases — for example, augmented customer service, fraud detection, or the automation of regulatory reports. Success will also depend on cross-functional teams and, above all, trust. Data quality, model explainability, governance, and ethics are indispensable for sustainable deployment.”

Looking Ahead

PwC Luxembourg’s 2025 survey reveals an economy at a crossroads. Governance is robust, but strategy, talent, and investment will determine whether organisations can truly harness the potential of GenAI. Banks are driving efficiency, insurers are laying cautious yet solid foundations, and both sectors recognise the importance of collaboration and trust.

As Andreas Braun puts it: “Governance is not the same as strategy.” To unlock the full value of data and AI, Luxembourg’s financial institutions will need to move from pilots to scale — with a clear vision, stronger budgets, and a shared commitment to innovation.