This month we brought together senior financial services leaders in London spanning global transaction networks, investment banks, central banking, digital challengers, and established high street names.
The conversation cut straight to a question the industry keeps circling: is financial services genuinely re-architecting for an AI-driven future, or is it papering over structural cracks with a layer of intelligence?
The honest answer, for many in the room, is somewhere in between. But what struck us most was not the diagnosis, it was the scale of the opportunity waiting for those willing to act on it.
The productivity paradox and what it points to
The financial services sector has invested heavily in digital transformation over the past three decades. The results have been real: transaction costs slashed, new channels opened, customer journeys reimagined. Yet the very infrastructure built to drive efficiency has created its own drag.
Automation that cuts transaction processing costs by 60 - 70% simultaneously introduces a 12 -15% rise in back-office costs from data management complexity. Custom API and middleware investments consume 15 - 20% of IT budgets. Siloed data contributes to a 30 - 35% spend on marketing campaigns that miss their mark. Inconsistent data has driven a 15 - 20% rise in regulatory inquiries from audit reconciliation errors.
These aren't edge cases. They're structural features of a sector that digitised at speed without always asking whether the data foundations would hold. And crucially, they point directly to where the next wave of value is waiting to be unlocked.
The AI opportunity requires a foundation
Generative and agentic AI don't solve the data problem, but they make the prize of solving it dramatically larger. Real-time personalisation, intelligent automation, predictive credit decisioning, proactive fraud detection, all are entirely dependent on data that is unified, trustworthy, and accessible. An AI agent cannot act intelligently without a clean, governed foundation beneath it.
The institutions that move with urgency on their data foundations now will capture a disproportionate share of AI's value. Those attempting to scale it onto fragmented infrastructure will find themselves investing heavily for marginal returns.
The opportunity is not theoretical. Organisations that have rebuilt on unified data foundations are compressing credit model development from two to three years down to six to eight months, reducing processing time by 80%, and growing digital customer acquisition threefold while cutting cross-channel marketing spend by over a quarter. These are structural step-changes in competitive capability - explored in depth in 'Unlocking Banking's Data Advantage', the full research report published by Cognizant and Snowflake.
From data burden to data assets
The shift required is conceptually clear, even if operationally demanding: from treating data as a byproduct of business activity to treating it as a managed, reusable strategic asset.
The organisations making the most meaningful progress are moving beyond centralisation to creating data products, curated, governed, discoverable assets that any team can access and apply without starting from scratch. The distinction matters. A data warehouse is infrastructure. A data product has an owner, clear lineage, defined access controls, and embedded quality standards. One bank we work with published client relationship value scores as reusable data products through a layered architecture, enabling commercial banking teams to identify cross-sell opportunities with precision, monitor client relationships in real time, and align performance metrics with genuine value creation.
Data that had previously been locked in systems became a shared engine for growth.
The UK governance advantage
There is a dimension to this that is particularly compelling in the UK context. British financial services operates under one of the world's most demanding regulatory frameworks ,which many institutions have experienced as a brake on innovation. We'd argue it is increasingly a source of competitive advantage, for those who architect correctly.
Institutions that embed governance into their data architecture from the outset ,making access controls, lineage, and audit trails features of the platform rather than processes bolted on top ,find that regulatory rigour and innovation velocity become mutually reinforcing. When AI capabilities operate within the same centralised governance plane as the underlying data, institutions can move faster precisely because the controls are already in place. As the FCA and PRA watch AI deployment in credit, risk, and customer communication with increasing scrutiny, that architecture is not just good practice ,it is a competitive differentiator.
The defining question
The sector is not short of AI ambition. What that conversation ,and many like it ,consistently surfaces is that ambition without structural readiness is expensive optimism. The institutions that will define the next era of UK financial services are those connecting their AI roadmaps to the data foundations required to deliver them.
The data opportunity is not a future aspiration. For organisations willing to build properly, it is available right now.
Read the full research we did with Snowflake: Unlocking Banking's Data Advantage, From Isolated Islands to Connected, Reusable Assets, published by Cognizant and Snowflake.
