DBS Group Holdings, Southeast Asia's largest bank, generated SGD 1 billion (approximately USD 750 million) in verified economic value from artificial intelligence and machine learning in FY2025 — reaching in three years a target publicly committed for five. CEO Tan Su Shan disclosed the result in the bank's Annual Report released on 8 March 2026. The figure is measured against control groups, making this one of the few enterprise AI disclosures in global banking with an explicit before-after methodology.
A public commitment made in 2022
In its 2022 Annual Report, DBS disclosed that AI initiatives had already delivered SGD 180 million in economic value — SGD 150 million in revenue uplift and SGD 30 million in productivity gains and cost avoidance. Then-CIO Jimmy Ng attached a five-year target to that baseline: SGD 1 billion in AI economic value by 2027. The commitment appeared in a document filed with the Singapore Exchange, visible to every investor and analyst following DBS Group Holdings. That public visibility created a governance structure rare among AI programs: the target would be verified or disproven in print.
Three years of compounding deployment
DBS treated AI as a cumulative, cross-business discipline rather than a single transformation initiative. By end of FY2023, AI initiatives had delivered SGD 370 million in economic value — more than double the FY2022 baseline in one year. By 2024 the figure reached SGD 750 million. By FY2025, the bank crossed the billion-dollar threshold, operating more than 2,000 AI and ML models across 430 distinct use cases spanning credit risk, wealth management, customer engagement, and internal operations.
DBS-GPT, the bank's internal large language model, gives employees role-based access to more than 4 million DBS policies and documents. In the technology division, generative AI now automates test-case generation and user-story documentation — timelines that once ran months close in weeks. In July 2025, DBS launched Joy, a generative AI chatbot for corporate and SME banking customers that reached 20,000 active users and delivered a 23% increase in customer satisfaction scores against the pre-launch baseline.
A measurement methodology that sets the standard
The SGD 1 billion figure carries weight because of how DBS calculates it. The bank compares outcomes from AI-powered interventions against control groups — customers or processes that received standard service with zero AI-driven treatment — across three value dimensions: revenue growth through enhanced interest income and fees; cost reduction through operational efficiency gains; and risk avoidance through reductions in risk-weighted losses.
The control group methodology produces precise evidence. DBS's hyper-personalized AI nudge engine engaged 8.6 million consumer customers in 2023; Singapore customers in the exposed group showed 83% higher savings rates, four times the investment activity, and double the insurance uptake versus the control group. DBS's credit risk alert system identified more than 95% of at-risk SME loans three months in advance, with over 80% of flagged borrowers receiving intervention that prevented default. Forrester's analysis characterizes this approach as producing “tangible, measurable benefits rather than projections common in banking industry claims.” Global Finance named DBS the World's Best AI Bank.
What other organizations can learn
The DBS case rests on four replicable conditions. First: public accountability — committing to a number in a publicly filed document that investors and regulators can track. Second: control group measurement built before deployment at scale, generating evidence rather than narrative. Third: unified infrastructure that lets 430 use cases compound on shared learning rather than run as disconnected experiments. Fourth: AI treated as process redesign — DBS completed nine operating model transformation initiatives in FY2025, each re-engineering how work happens rather than layering tools onto existing workflows.
CEO Tan Su Shan's FY2025 report signals the bank's next phase: agentic AI systems capable of autonomous action across complex banking workflows. The SGD 1 billion result establishes that AI generates enterprise value at scale when measured with rigor. The agentic phase raises the follow-on question: at what organizational velocity can that value compound further?
Article by SAGA — Success Stories & Real Cases
SAGA covers enterprise AI implementations with verified outcomes. Every metric is sourced. Every company is named.