DBS

DBS, one of Asia's largest banks operating across 19 markets, recognised over a decade ago that competing in an increasingly digital financial landscape required more than incremental technology upgrades. Its legacy architecture relied on fragmented, siloed data systems that lacked the consistency and governance needed to power AI at scale. Without a unified data foundation, the bank could not deliver the personalised, real-time customer experiences that digital-native competitors were beginning to offer, nor could it drive operational efficiency across its global franchise. The absence of clean, centralised data represented a strategic ceiling — one that would only widen as customer expectations and competitive pressure accelerated.

DBS made a deliberate, multi-year investment in a centralised data lake that became the foundation for all subsequent AI and machine learning initiatives. By consolidating previously siloed data sources into a single governed platform, the bank created the infrastructure required to deploy predictive analytics and personalisation models across its entire customer franchise — enabling hyperpersonalised contextual offers and real-time recommendations at scale. Internally, the bank developed DBS GPT, a secure generative AI assistant that democratises institutional knowledge, giving employees on-demand access to expertise across functions. Alongside technology deployment, DBS launched an extensive role-based training programme to upskill over 12,000 employees in AI and data capabilities. The sequencing was deliberate: data infrastructure and governance first, predictive models second, and workforce capability building running in parallel throughout.

DBS sequentially doubled the economic value generated by its data analytics and AI programme since 2021, reaching S$750m (US$561m) in 2024. The bank is on track to surpass S$1bn (US$764m) in AI-driven economic value in 2025 — a milestone CEO Tan Su Shan attributes to systematically rewiring the organisation around data and AI rather than simply deploying tools.

• Customer service efficiency: An AI assistant deployed to over 1,000 customer service officers reduced average call handling time by 20%.
• Workforce transformation: More than 12,000 employees actively upskilled in AI and data skills through role-based training.
• Knowledge access: DBS GPT expanded institutional knowledge democratically across the entire staff base, improving productivity across functions.

• Data infrastructure is the non-negotiable prerequisite: DBS built and governed its data lake before scaling AI — clean, centralised data determines the ceiling of what models can achieve.
• Workforce investment multiplies technology ROI: pairing AI deployment with structured, role-based training accelerates adoption and extends value well beyond initial use cases.
• Responsible AI requires explicit governance from day one: protocols covering data integrity, accessibility, and security must be in place before scaling, not retrofitted after incidents.
• Sequentially doubling economic value requires organisational rewiring, not just tooling — leadership commitment, culture change, and talent pipelines matter as much as the models themselves.
• Tracking AI value in dollar terms creates accountability and sustains executive investment through the long build-out phase.