FICO: Banks risking US$170B as AI rewires retail banking

FICO's newest executive brief makes a blunt case: lenders that treat artificial intelligence as just another technology rollout, rather than rebuilding the machinery behind their decisions, will be outpaced by rivals that do. For institutions that get the architecture right, the upside is material. The analytics firm reckons the strongest performers can add up to four percentage points to their return on tangible equity, while laggards face something worse than a missed opportunity.

The brief, Winning in New Era of AI-Powered Intelligent Banking, gathers FICO's read on how retail banking is being reshaped across Asia Pacific. It frames the moment as a structural reset touching growth, risk management and customer relationships alike, and argues that banks which keep AI as a bolt-on are the ones most exposed to losing ground.

Five pressures anchor the analysis: AI-driven search and generative tools changing how consumers discover and choose products; loyalty that frays as a result; scam networks that now pair automation with social engineering to talk victims into authorising payments themselves; watchdogs insisting automated decisions be explainable, governed and monitored; and the looming prospect that agentic assistants divert value away from banks toward new platforms and intermediaries. FICO puts the erosion risk at as much as US$170 billion in worldwide banking profits over the coming decade. The common thread, it argues, is decisioning too fragmented to keep pace with fast-moving behaviour, threats, rules and competitors.

The proposed remedy is a connected decisioning layer, where data, models and policies are reused across fraud screening, credit risk, lending origination and customer management instead of being rebuilt for every new use case. That is an operating-model and governance shift as much as a technology one, and it drags AI oversight onto the board's agenda rather than leaving it with data teams. Banks still running isolated pilots will find each new AI initiative starts from scratch, which is a poor fit for a market where customer attention, fraud tactics and regulatory expectations all move quickly.

Why it matters for Singapore: MAS is among the regulators FICO names as tightening expectations around how automated decisions are explained, governed and audited, which gives local lenders a head start but also puts them under a brighter spotlight. Banks here will be judged on whether AI-driven calls can be defended to customers and supervisors, not merely on how fast they deploy models. As Singapore positions itself as a trusted hub for enterprise AI, the decisioning overhaul the brief describes looks less like an option than a competitive necessity for its financial sector.