From cutting onboarding times to minutes to reducing data-entry workloads by 50%, DBS Bank India shows how AI can quietly transform core banking.
When most people talk about AI transformation, it usually begins with lofty ambition statements. At DBS Bank India, it begins with something far more simple: documents.
Loan files, KYC forms and financial statements – the manual handling of these slows lending cycles and consumes thousands of hours of frontline time. Over the past 18 months, the bank has quietly pushed an AI-powered automated document processing system from pilot to production. It extracts and validates data, auto-populates workflow systems, and now performs the grunt work that once stretched across days. “It acts as a ‘digital maker’, significantly reducing manual effort,” Ramesh Mallya, CTO, DBS Bank India tells ET Enterprise AI. That reduction, he highlights, has meant two things: faster loans and time freed up for employees to do work humans are actually good at such as relationship-building and problem-solving. AI in the customer journey Across the bank, AI is embedded into onboarding, fraud detection, and personalisation, often in ways the customer never directly sees. One early gain has been onboarding.
With AI-powered identity verification and optical character recognition (OCR), DBS has collapsed timelines “from days to minutes,” improving both efficiency and customer satisfaction. On the servicing side, MSME customers are already interacting with Joy Bot, the bank’s conversational assistant that handles routine queries and increases self-service rates. For internal teams, the GenAI-enabled customer service officer (CSO) assistant acts as a speed layer, shortening response times and improving the quality of customer communications. Personalisation has also matured, Mallya notes. AI models analyse spending and saving behaviour to offer contextual insights that nudge customers towards smarter decisions. It’s part of the bank’s broader aspiration to be a “GenAIenabled bank with a heart”, a phrase the bank uses internally to describe AI that is efficient without being impersonal.
Unlike many organisations that chase use cases first, DBS builds from infrastructure and governance upward. It follows a “Platform–Process–People” approach to AI, underpinned by strong governance and scalable infrastructure to accelerate time-to-value while ensuring responsible deployment. Two internal platforms, ADA (Advancing DBS with Analytics) and ALAN, the bank’s AI protocol and knowledge repository, act as the backbone. ADA ensures quality and discoverability of data, while ALAN embeds governance controls into every model lifecycle.
DBS also draws a clear line between specialised models and general LLMs. A specialised model trained on banking documents, for example, can extract key financial ratios “with near-perfect accuracy,” outperforming general-purpose models on rule-based precision. LLMs such as DBS-GPT are used instead for tasks like summarisation, internal knowledge access, and productivity enhancement. The build-vs-buy philosophy is clear at DBS: build when it offers strategic advantage or demands tight data control (as with DBS-GPT); buy when the market offers mature, cost-effective tools that accelerate deployment.
Underpinning all of this is governance: explainability audits, bias and drift monitoring, human-in-the-loop checks for high-risk decisions, and oversight bodies like the Responsible Data Use Committee and the Responsible AI Taskforce, Mallya points out. Frameworks such as the GenAI Risk Assessment Framework ensure deployments remain “lawful, ethical, and transparent,” he adds. Data, skills, and partners Any bank operating in India understands the messy reality of fragmented data, multi-language user bases, and diverse financial behaviours. DBS Bank India has responded with unified data platforms, automated quality checks, and metadata tools that track lineage for auditability.
According to Mallya, this has materially improved model reliability, an area where DBS has historically invested for over a decade as part of its “digital to the core” mandate. The bank is also re-skilling aggressively. Employees across operations, risk, and product functions are being trained in AI literacy and ML techniques. Tools such as DBS-GPT double up as digital knowledge companions, offering simulations and guidance that sharpen on-the-job decision-making. This capability-building is beginning to show up in software development as well. A GenAI-powered engineering assistant used by over 80 per cent of DBS India’s tech and data employees now supports code completion, debugging, documentation lookup, optimisation suggestions, and personalised learning.
On the business side, GenAI-led content capture has cut data entry time per transaction by nearly 50 per cent, while also improving processing flow and exception handling. For DBS India, fintechs and cloud providers contribute the most value to its AI roadmap as partners, according to Mallya. “Fintechs bring agility and specialised expertise that complement our scale and trust, enabling co-created solutions with faster deployment,” he says. One such collaboration led to Auto-Fin, an AIpowered underwriting tool that speeds up MSME lending and improves credit team productivity.
On cloud providers, Mallya notes that they offer scalable and secure infrastructure for AI-heavy workloads, especially as the bank deepens its GenAI capabilities. The road ahead “Over the next 24 months, AI will reshape operations, customer experience, and risk management,” Mallya notes. Over the next two years, the bank’s key priorities include deepening automation in lending by integrating document processing with AI-driven credit underwriting to shorten loan turnaround times. On the customer experience side, the bank is moving beyond reactive chat interfaces towards proactive, insight-led interactions powered by generative AI, the kind that anticipates intent rather than simply responding to queries. And in risk and ESG, continuous-learning models will sharpen fraud detection while AI-led ESG assessments help strengthen credit decisions for MSMEs, Mallya shares.
“What excites me most is AI’s potential to enhance human capability and create more inclusive financial ecosystems,” he says. Faster decisions, fewer errors, and personalised financial guidance all point to that direction. But the guardrails matter just as much. Mallya’s note of caution as technology accelerates rapidly is that transparency, fairness, and privacy would require constant vigilance.
As featured in ET Enterprise AI
World’s Best Bank 2025, 2021, 2019, Euromoney
World’s Best Bank for Customer Experience 2025, Euromoney
World’s Best Bank for Corporate Responsibility 2025, 2023, Euromoney
Asia’s Safest Bank, 2009 – 2025, Global Finance
Best Bank in the World 2022, Global Finance
India’s Best International Bank 2021, Asiamoney
World’s Best Banks - #1 in India 2021, Forbes
World’s Safest Commercial Bank 2021, Global Finance
