Over the last decade, financial institutions have invested hundreds of billions in digital transformation. They are racing to deliver intelligent automation and real-time customer experiences, all with the aim of providing faster, more personalised services.
Yet customers continue to experience disruptions that prevent them from accessing their accounts, making payments, or using their cards. While the cause may vary, when these incidents stack up, banks need to ask themselves whether they are focusing too heavily on the symptoms rather than the root cause.
The question isn’t whether banks are investing enough in technology. It’s whether they’re investing in the right foundations to make the technology work for themselves and their customers.
The real problem isn’t the outage
Bank outages aren’t a good look. They frustrate customers, drain internal resources, and immediately shift attention to the system that failed. But focusing on the immediate cause risks overlooking the root cause that created the disruption in the first place.
Research among senior leaders at the UK’s largest banks suggests those conditions are already well understood. Almost half (48%) believe fragmented or siloed data is the single biggest barrier to operational reliability. When critical data is spread across legacy systems, business units and functional teams, ownership becomes unclear, information is duplicated, and dependencies are poorly understood.
Banks are grappling with complex technology estates, where new platforms are layered onto ageing infrastructure and hundreds of systems that depend on one another. That means that even a minor change in one system can result in unintended consequences elsewhere. Without visibility across those connections, a relatively small issue can quickly ripple across the wider organisation and recovery can take longer than expected.
That’s why pre-empting and solving outages aren’t simply about replacing legacy technology. It starts with understanding who owns critical data, how systems interact, and whether governance is strong enough to support increasingly complex operations.
AI is revealing the same problem
The same structural weaknesses are beginning to limit banks’ AI ambitions. Many financial institutions have demonstrated that they can build and launch pilot programmes. But the real problem arises when those successes need to be scaled across the organisation. While 77% of AI initiatives progress beyond the pilot stage, only 13% successfully scale into real-world operations in the UK.
The gap is rarely caused by a lack of investment. Instead, it reflects the challenge of deploying AI across environments where data is fragmented, inconsistent, or difficult to access.
In many cases, the challenge grows as organisations collect more data. Multiple technology stacks, inconsistent governance standards, and disconnected data sources make it increasingly difficult to generate reliable insights at scale.
This is where many organisations are celebrating the wrong wins. Deploying a handful of AI use cases is progress, but the real opportunity lies in redesigning customer experiences end to end, not bolting AI onto isolated processes. Without strong data foundations, AI risks becoming another layer of complexity rather than a catalyst for transformation.
This is where the opportunity really lies for banks. Banks with solid data foundations will be able to move from experimentation to execution far more quickly, giving them a competitive edge. Whether it’s opening an account, applying for a mortgage or resolving a fraud issue, AI should help redesign the entire journey rather than improving isolated tasks within it. That simply isn’t possible without trusted, connected data.
Operating models are the heart of modern banking
Fixing fragmented data isn’t just a technology challenge; it’s an organisational one.
The way many banks are structured makes it difficult to respond quickly, even when the problem is clear. Again, our research found that 56% of significant technology decisions still require C-suite approval, while only 5% are delegated closer to delivery teams.
In a highly regulated industry, oversight is non-negotiable but excessive bureaucracy slows progress and limits responsiveness. Equally, only 7% of organisations are adopting product-led, cross-functional teams that bring together technology, data and business expertise around clear outcomes. Those operating models give teams ownership of both the customer experience and the data that supports it, allowing issues to be identified and resolved before they become wider operational problems.
Banks that hold onto working in silos will find it increasingly difficult to deliver the speed, resilience and customer experience that the market expects.
The road to agentic AI
These challenges will only become more pronounced as agentic AI moves from concept to reality. As agentic AI systems become capable of making decisions in real time, adapting to changing circumstances and completing tasks with minimal human input, banks will need to double down on governance and adapt the controls they have in place.
Autonomous systems are only as reliable as the data and controls that underpin them. If banks are already struggling with fragmented data and inconsistent ownership, introducing increasingly autonomous AI will only amplify those weaknesses. Building stronger governance today isn’t simply about supporting today’s AI capabilities; it’s essential for the next generation of intelligent systems.
The foundations of future growth
Bank outages will likely continue to make headlines, but they raise a much bigger question: whether an organisation’s operating model is equipped to support the next generation of financial services.
Banks with trusted, well-governed data will be better positioned to scale AI, improve operational efficiency, and create more seamless and connected customer experiences. Rather than treating governance as solely a risk-management exercise, banks should see it as an enabler of growth, innovation, and customer value.
As investment in AI accelerates and the technology evolves, success won’t be determined by how many use cases a bank can launch. It will depend on whether it has the operating model, governance and data foundations needed to turn those ambitions into lasting business value.
