DBS AI Agents Take On Over 70 Credit Tasks for 1,500 Employees Worldwide

Agentic AI has crossed from the experimental edges of Singapore banking into the core of institutional lending. DBS is the latest home-grown lender to put the technology to work on a genuinely consequential workflow — the credit assessment that decides whether a corporate client gets financing in the first place.

The bank has deployed an agentic AI tool to help about 1,500 employees worldwide assess corporate clients seeking financing. Specialised AI agents handle more than 70 separate tasks, pulling information from annual reports, industry research and internal records to produce a first draft of a credit memo that relationship managers and credit risk managers then review. The stated goal: cut the time spent preparing credit memos and related work by at least 30% — tasks that today can absorb up to 40% of a relationship manager's time. Employees can instruct the agents to carry out further research and revise the draft before it's signed off.

The rollout is part of DBS's wider bet on agentic AI across customer service and internal operations. In July the bank announced agentic AI upgrades to its DBS Joy and digibot virtual assistants, which serve roughly 10 million customers across Singapore, Hong Kong and Taiwan. Institutional banking is a natural next frontier — Han Kwee Juan, DBS Group Head of Institutional Banking, says agentic AI has let the bank encode the knowledge of its best relationship and credit risk managers into a system that raises the quality of credit analysis at scale.

What makes this notable is the human accountability built into it. The AI drafts and revises, but relationship managers and credit risk managers remain responsible for the final memo, applying their own judgement on the client, industry and wider business environment. DBS expects the time freed up to flow into strategic client discussions and, for credit risk managers, deeper portfolio strategy and risk calibration. That division of labour — machines doing heavy lifting, humans wielding judgement — is emerging as the template for regulated, high-stakes AI use in finance.

Why it matters for Singapore: DBS is the flagship of Singapore's financial AI push, so how it deploys agentic AI in a regulated lending workflow is a bellwether for the rest of the ecosystem. As the MAS encourages responsible AI adoption across the financial sector, a home-grown bank demonstrating that agents can accelerate credit work under tight human oversight gives other lenders here a working blueprint — and reinforces Singapore's position as the region's testbed for production-grade financial AI.