From legacy to intelligence: the new data operating model for banks

The challenge

Data has always been central to banking. Over decades, institutions have invested heavily in collecting, storing, and analysing it. In recent years, that investment has accelerated. Core platforms have been modernised. Cloud adoption has increased. Reporting has improved.

Through all of this, however, a simple question still causes hesitation.

Do you trust your data?

For many organisations, the answer is unclear. Not because data is missing, but because confidence in it is inconsistent. Analysis of the figures may seem broadly correct but struggles to explain where the conclusions came from? How were they produced? How resilient would they be under pressure.

In 2026, this uncertainty is no longer abstract.

Recent incidents in UK banking have shown that data failures can occur during routine system changes. A poor outcome does not just happen during cyber-attacks. In at least one case, a software update led to customer data being exposed through a mobile banking app. No money was stolen but compensation was paid. Parliamentary scrutiny followed.

The issue was not malicious activity. It was a lack of control over how data was surfaced and reused.

At the same time, regulators are making it clear that higher quality, clearer and more reusable data is now expected. The direction of travel is away from one-off reporting and towards demonstrable, ongoing capability.

Together, these signals point to the same conclusion. The challenge facing banks is no longer collecting data. It is being able to trust it.

Why current approaches fall short

Most institutions are running modern data platforms on legacy ways of working. This creates a growing gap between technical capability and operational confidence.

Common issues include:

• unclear ownership of critical datasets
• inconsistent definitions across business areas
• limited visibility of lineage from source to output
• governance that sits outside delivery rather than within it
• increasing exposure to cyber and operational risk at the data layer
• greater confidence in decision making
• reduced regulatory and operational risk
• stronger foundations for AI and advanced analytics

As data volumes increase and platforms become more interconnected, complexity rises. Without structure, this complexity does not produce insight. It produces hesitation. It promotes a lack of trust.

In practice, many organisations only discover these weaknesses when something goes wrong. A report is challenged. An incident occurs. A regulator asks for evidence that cannot easily be produced.

What leading organisations are doing differently

We are seeing a shift among organisations that are addressing this problem successfully. Rather than focusing solely on platforms, they are adopting a more deliberate data operating model built around four principles.

Data as a product

Critical datasets are treated as long-lived assets rather than just project outputs. Ownership is clearly defined. Accountability sits with named individuals who are responsible for quality, availability and fitness for purpose.

Governance embedded into delivery

Controls are built into data pipelines rather than applied retrospectively. Lineage is tracked as standard. Auditability is designed in from the start, not added later under pressure.

Resilience built in early

Data is classified consistently. Access is controlled and reviewed. Recovery capability for critical datasets is understood and tested. The focus shifts from preventing every issue to limiting impact and accelerating recovery.

An aligned operating model

Business, data and security teams work to shared priorities. Accountability for outcomes is collective rather than fragmented. Improvement is continuous rather than driven by isolated programmes.

This is not a radical reinvention. It is a structural correction.

The outcome

Organisations that adopt this approach begin to see clear benefits:

Most importantly, leaders stop second guessing their own information. Data becomes something they can rely on under normal conditions and under stress. In other words, data becomes a reliable platform on which decisions can be confidently made.

Where to start

For most organisations, the first step is clarity.

Understanding which datasets matter most, who owns them, how they are governed and where risk sits provides a practical foundation for change. Without that baseline, even the best technology investments struggle to deliver lasting value.

The shift is simple but significant: a shift from managing data to trusting it.