Modernising legacy banking systems: a challenge for today’s CIOs

On top of this comes organisational and cultural change. What are the hurdles to navigate? And what's the role of application modernisation powered by AI? At Zühlke, we've been exploring these questions and considering:

• Why outdated legacy systems urgently need to be replaced
• How AI-powered application modernisation helps break down large-scale projects and gradually build a new IT architecture
• The role of professional change management
• What the bank of the future looks like

Across industries, established companies recognise the imperative for legacy system transformation. According to the Lünendonk Study 2025, IT modernisation and cloud transformation are some of the highest-demand areas for IT service providers. This is hardly surprising given that IT architectures have evolved over decades.

For banks, legacy modernisation is particularly challenging. As part of critical infrastructure, they are subject to strict regulatory requirements. Moreover, core banking systems have often evolved organically, with new functions and services layered onto existing structures over time. As a result, many systems have remained largely untouched by earlier transformation initiatives – until recently, they were stable (enough) to support operational resilience.

Today's core banking systems have become a ticking time bomb, however, threatening the competitiveness and security of German banks as technology evolves at pace. Legacy systems were not designed for the speed, complexity and agility required today. Traditional banks are falling behind, especially compared with fast and agile fintechs.

Within banks, legacy application modernisation is one of the CIO's core responsibilities. As the leader of this transformation, the CIO's own success within the organisation can hinge on progress and outcomes. Given that such transformations can take a decade or longer, many hurdles must be navigated along the way. Costs can spiral out of control, and delays, IT outages or security incidents can severely impact the bank’s competitiveness.

At the same time, the importance of IT – and therefore of the CIO – is growing. IT is increasingly at the centre of the bank’s future. It determines how competitive a bank is and how it is perceived by customers. In other words, whereas a bank used to be a financial institution with an IT department, it is increasingly becoming a software company with a banking licence.

To better understand existing core banking systems, banks are increasingly turning to AI-powered tools for enterprise application modernisation. They enable implicit knowledge, previously held only in developers’ minds, to be made visible through a digital twin.

This digital twin is a knowledge database created through reverse documentation and accessible through natural language queries. It enables targeted changes and, in a second step, the analysis of individual modules, which can then be modernised using AI. In this way, the Herculean task of technological transformation can be divided into smaller, manageable steps.

This approach is already delivering measurable results. For example, a leading Swiss private bank began to machine-read the knowledge layers of its core systems, model dependencies, and make logic transparent. The result: shorter release cycles, lower error rates, faster responses to regulatory changes and, for the first time, an architecture that explains what it does.

For transformation to succeed, adopting the right technologies is not enough. Mastering change management is equally essential, as technological transformation must be accompanied by changes in processes, ways of working, roles and responsibilities.

The CIO must bring employees, and often executive management, on board with the transformation. It affects not only IT, but the entire organisation.

So how should a CIO begin such a comprehensive transformation? Start by mapping the current IT landscape and defining a target vision for how technology and the organisation should evolve.

A pragmatic first step is to select a single module – for example, payments, digital account opening, or customer onboarding – and use AI to fully analyse and understand it, including all dependencies and risks, before migrating it into a modern system.

This creates an initial, verifiable learning model and quickly demonstrates the value of AI-assisted modernisation. From there, modules can be prioritised based on clear criteria – typically starting with those under the greatest pressure, such as regulatory requirements or urgent customer demands.

When restructuring the organisation, pilot projects and teams should also be used to move towards the target operating model. Thinking in learning cycles is essential: observe, understand, adapt, review.

In this way, IT architecture, organisation and culture evolve together, and incremental modernisation moves the bank toward the desired future state.