AI & cost of legacy systems in UK banking
AI tools are revealing the deep-seated inefficiencies in banking IT, not just fixing them.
A new analysis reveals UK banks are spending approximately £3.3 billion every year—about a quarter of their total IT budgets—simply to keep aging core banking systems running. Nearly half of these institutions admit their legacy platforms struggle with modern demands. For decades, this massive cost was accepted as the price of stability for systems that process millions of daily transactions and ensure compliance. However, the rapid advancement of artificial intelligence is fundamentally challenging this status quo, exposing that the real problem runs deeper than old software.
AI development tools that speed up coding and testing are creating pressure for faster software delivery cycles. Yet, introducing these tools into banks' fragmented environments often backfires. The issue is that business logic and requirements are inconsistently defined across siloed teams, with critical knowledge held by a shrinking group of experts. When AI is fed ambiguous or contradictory inputs, it reproduces those flaws at scale, leading to brittle and high-risk releases. The £3.3 billion maintenance bill is thus a symptom of decades of accumulated technical debt—a dense web of integrations and workarounds—that AI alone cannot solve. The technology has broken the illusion that traditional delivery models were ever robust, revealing a hidden fragility that must be addressed at a foundational level.
- UK banks spend £3.3 billion annually, roughly 25% of IT budgets, just to maintain legacy core systems.
- AI tools can amplify existing problems in fragmented bank IT environments, reproducing inconsistent business logic at scale.
- The core issue is decades of accumulated complexity and technical debt, not merely outdated technology.
Why It Matters
For tech leaders, this highlights that deploying AI without modernizing underlying data and process foundations is a high-risk investment.