The Foundation First – ING’s Global CIO on Why AI Pilots Fail to Scale

In a recent interview, Global CIO of Wholesale Banking, Bouke Hoving shared that ING’s AI transformation focus has shifted from pilots to the hard problem – moving AI into production at scale across regulatory fragmentation, data localisation constraints, and talent scarcity that defines Southeast Asia.
 
Hoving reveals the mechanics behind ING's success. First, governance by the highest standard: ING uses forward-looking regulators like Singapore's MAS and Germany's BaFin as the baseline, applying those controls across regional markets regardless of local requirements. Second, data strategy: input and output filters ensure AI models comply with localisation rules without sacrificing cross-border insight. Third, and most critical, scalable foundations. A decade of core banking modernisation across ING's footprint created standardised products and harmonised data infrastructure, the real enabler of AI scaling.
 
On the innovation side, Hoving walks through three concrete wins. The HyPe Engine delivers personalised messaging to 4.6 million customers with double-digit uplift in campaign success rates by moving from rules-based to GenAI-driven micro-segmentation. KYC automation reduced manual onboarding tasks by 80% by shifting verification work into the background. Agentic AI in lending is moving from copilots that augment analysts to agents that execute tasks end-to-end, with humans staying involved for exceptions and high-risk decisions.
 
The biggest takeaway for me is AI transformation failure in banking is not a technology problem, but a scaling problem. Banks that invested in modernised platforms and common data foundations can move pilots into production in weeks. Those without that foundation are still stuck.

Originally published in CIO: Opens in a new tabThe Foundation First – ING’s Global CIO on Why AI Pilots Fail to Scale | CIO