DBS Bank’s AI programme has moved from pilots to platform. By integrating Google Cloud — notably Vertex AI — with its internal ADA platform and layering Glean’s Work AI, DBS has turned AI into everyday leverage for tens of thousands of employees. The impact is real: independent coverage and industry analyses attribute roughly SGD 750 million (≈USD 563m) in economic value to AI in 2024, with further growth signposted in 2025.

Why it matters now

Banking productivity is increasingly determined by how quickly teams can find knowledge, summarise complex information, and trigger routine actions safely. DBS’s approach shows how to combine a governed data backbone (ADA), a robust model platform (Vertex AI), and a Work AI layer (Glean) to deliver time savings and better decisions — without compromising compliance.

What’s new?

• Platform integration: Google Cloud’s Vertex AI is integrated into DBS’s self-service data platform (ADA), supporting rapid scale-out of use cases and automated infrastructure management as volumes grow.

• Measurable value: External spotlights report ~SGD 750m in value from AI in 2024, reflecting broad-based productivity and revenue use cases.

• Recognition: DBS was named World’s Best AI Bank in 2025, reflecting execution depth across operations and customer experiences.

• Work AI at scale: Glean and Google Cloud are cited as enabling 40,000 employees to work faster and unlock new AI agents and bots; posts suggest 5–10% time savings in day-to-day work. (Claim per partner communications.)

How the stack works

1. Data foundation (ADA): Clean, governed data products feed AI safely. Access controls and lineage keep auditors happy.

2. Model platform (Vertex AI): Teams use managed tooling to build, evaluate and deploy use cases — from summarisation and classification to agentic workflows with Gemini-class models.

3. Work AI layer (Glean): Employees search across systems, generate briefs, and trigger actions via agents that sit in daily tools. This shrinks the “find-understand-act” loop across operations, risk, and customer teams. (Per vendor + public materials.)

4. Guardrails: Policies and monitoring enforce responsible AI — vital in regulated banking. DBS publishes its responsible AI perspective, emphasising governance and trust.

Practical examples

• Operations: AI reduces manual processing time across key back-office workflows; analysts cite significant reductions alongside higher throughput.

• Frontline knowledge: Staff can retrieve cross-system knowledge and draft summaries in seconds, supporting quicker, more consistent customer responses. (Per partner communications.)

• Risk & compliance: Generative AI accelerates adverse-news screening and documentation, helping teams focus on judgements rather than sifting.

A repeatable rollout blueprint

1. Start with a governed data layer. Define high-value data products and access policies first; AI without clean, accessible data stalls.

2. Adopt a managed model platform. Standardise on a platform like Vertex AI for evaluation, deployment, and monitoring to avoid tool sprawl.

3. Activate Work AI for employees. Deploy an enterprise-search and assistant layer (e.g., Glean) to turn AI into daily productivity — not just a lab project.

4. Measure, then scale. Track time saved, throughput, quality and risk outcomes; reinvest where value is proven (DBS’s 2024 figures illustrate the compounding effect).

5. Embed responsible AI. Implement model governance, human-in-the-loop, and audit trails from day one.

Results to watch

• Time saved per employee: Partner comms suggest 5–10% time savings at DBS from Work AI — a useful benchmark for business cases. (Directional; validate locally.)

• Value creation: Track economic impact similar to DBS’s reported ~SGD 750m (2024) as programmes mature.

• Recognition & resilience metrics: Awards and case studies signal maturity; operational KPIs (cycle times, error rates) prove durability.

Summary & next steps

DBS shows that productivity gains come from system design, not isolated pilots: governed data (ADA), a strong AI platform (Vertex AI), and a Work AI layer (Glean) that meets employees where they work. If you’re building your roadmap, start with one or two cross-functional use cases, wire in governance, and measure relentlessly.

Talk to Generation Digital to plan a pilot that pairs Google Cloud and Glean for fast, safe productivity wins — then scale.

FAQ

How does AI improve productivity at DBS Bank?
By compressing the “find-understand-act” cycle: enterprise search, summarisation and agents reduce manual effort and speed decision-making across teams.

What role does Google Cloud play?
Google Cloud’s Vertex AI underpins model development, evaluation and deployment, integrated with DBS’s ADA platform to scale use cases securely.

Where does Glean fit?
Glean provides the Work AI layer — search, assistant and agent workflows across daily tools — enabling broad employee productivity. (Per vendor materials.)

Is this approach recognised by the industry?
Yes. DBS was named World’s Best AI Bank in 2025, highlighting execution at scale.