It is indeed a privilege to be here at the Global Fintech Fest at a time when technology is once again reshaping the way we think about banking. In my 38 years as a banker, I've had the opportunity to witness several transformations. I've seen banking evolve from a largely branch-led, paper-intensive model to one where technology has progressively moved the point of banking closer to the customer. But all these changes are underpinned by a fundamental truth, that is trust. Trust at scale and GenAI and the future of banking in a billion-user nation is a topic which I've been given to speak on.
But what are we looking at? If we look at India's financial transformation over the last decade, the starting point was not technology for its own sake. It was inclusion. The Jan Dhan Aadhaar mobile and JAM Trinity created the foundation for bringing millions of Indians into the formal financial system. UPI provided a powerful tailwind to this foundation and fundamentally changed the way millions of Indians transact. That is perhaps India's most important achievement of the last decade. We successfully converted digital infrastructure into trust and trust into scale. And that experience becomes extremely important as we enter the next phase of technological transformation: rapid penetration of artificial intelligence, which is not only about making banking personalized and seamless, but will be about allowing technology to act on our behalf—from digital to intelligent banking.
Artificial intelligence is not new to banking. Banks have been using predictive models and machine learning for quite some time now across many areas. At SBI, we have been early adopters of artificial intelligence. Customer service was one of the early visible applications through chatbots designed to handle routine queries. We were among the pioneers in the industry in using AI and machine learning for pre-approved personal lending. We're using conventional AI-based approaches in credit assessment and cash-flow-based lending, expanded the use of AI in risk management, including early warning systems, fraud detection, anti-money laundering monitoring, and proactive risk management.
And we are now moving into the next generation of AI capabilities, which is going to be different. Agentic AI, the next stage, can potentially move from assisting with a task or carrying out a defined set of tasks and beyond on behalf of a customer and an institution, and respond to changing circumstances with a degree of autonomy. That to me is the difference between a tool that we operate and a capability to which we delegate. The real opportunity lies in deploying agents to areas where scale, speed, and judgment can materially strengthen the financial system. Agents can work across the financial lifecycle of a transaction and a customer, strengthening fraud and mule account detection, supporting KYC, AML, and regulatory processes, accelerating loan appraisal and underwriting, and enabling intelligent reconciliation. On the customer side, they can transform services and complaint management while supporting more proactive and contextual customer lifecycle management. These are not merely opportunities for automation. At India's scale, they can improve the quality of decisions, reduce response times, strengthen controls, and make the financial system more resilient.
Why is India different? We are introducing Agentic AI to a market which is a diverse and complex ecosystem of more than a billion users. India has already demonstrated that technology can be adopted extraordinarily quickly when it is simple, affordable, interoperable, and trustworthy. We have the identity infrastructure, payment infrastructure, consent-based financial data sharing through the account aggregator, and the unified lending interface, resulting in an increasingly sophisticated digital financial ecosystem. And we are now building domestic and sovereign AI capabilities as well. The next frontier for Indian AI is not merely about building larger models, but building models that understand India's diversity and uniqueness. Building foundational AI models in India, for India, is therefore a strategic national capability, which on successful implementation can have relevance across much of the emerging world.
For Bharat, voice-first conversational banking can be particularly a powerful use case. Vernacular models can make financial services more accessible to customers who are more comfortable speaking than typing and in languages other than English. The real promise of agentic AI is demonstrating financial intelligence. The most exciting aspect of agentic AI is not automation, as I said earlier; it is the possibility of democratizing financial intelligence. Agentic AI gives us the opportunity to extend personalized financial assistance to hundreds of millions of customers across every stage of their financial growth. Currently, financial intelligence is restricted to a certain number of savvy customers through wealth management and premier banking, but democratizing this financial intelligence across millions of people is the powerful use case for agentic AI.
Much of this deployment will happen on the mobile phone. This makes affordability, low latency, and efficient use of computing especially important. While the initial fixed cost of agentic AI can be high, the incremental cost can be low, creating the economics of first scale. Newer, scalable, and cost-efficient commercial models will also need to emerge for computing and cybersecurity. This is one of the very important points: how do you make high-cost computing currently available, particularly on the agentic AI and generative AI side, more affordable? When you want to use it at scale, you can't have the high-cost models that are prevailing now. How do we design AI models and AI computing power for India which is affordable is going to be an important element when we talk about scale to be achieved here. In India, this capability through nation-centric customization can be delivered at scale, as UPI has demonstrated, making financial services more accessible, relevant, responsive, and meaningful to the customers. Through business correspondents and kiosk banking, we can extend this intelligence beyond the smartphone as well. Agents can support financial services at the last mile while also helping communities with relevant information with respect to agriculture activities such as weather updates and crop patterns.
AI also should augment human resources. I've always believed that technology succeeds in a bank only when the people using it believe that it makes them better. A significant number of financial activities become increasingly automated. Employees can devote more of their time and capability towards value-driven relationships, problem-solving, and deeper customer engagement. This is where the principle of human in the loop becomes important. Routine, high-volume, data-intensive activities can be handled by AI while complex, high-risk financial decisions continue to receive appropriate human oversight. My optimism about AI is accompanied by an equally strong sense of responsibility. When a system can act autonomously, an error can potentially trigger a sequence of actions across multiple systems and affect customers, counterparties, and institutions at machine speed. The same capability that gives AI its enormous power can therefore amplify the consequences of a mistake. This is why I believe trust cannot be added, but should be embedded in our agentic AI models.
Trust at scale in the context of agentic AI can be expressed through what I call the three A's: accuracy, accountability, and access without asymmetry. An agent operating in financial service needs to be accurate. There is nothing like 99% accuracy; you need to be 100% accurate every time, all the time. At population scale, average accuracy is not enough. It needs to work for customers from all walks of life, and consistent performance at scale itself becomes a test of AI quality. The second dimension is accountability by design. The moment AI moves from recommendation to execution, accountability becomes even more important. There must be an audit trail, traceability, and the ability to understand why an important action was taken. The third dimension is access without asymmetry. AI must be available without bias and must be adoption-blind. The technology should adapt to the customer and not the other way around.
Risk management will have a whole new meaning in an AI-connected world. Banks have spent decades building robust processes about know your customer. As agents begin to participate in financial transactions, we will increasingly need to think about 'know your agent.' We will need clear mechanisms around agent identity, authentication, consent, transaction limits, audit trails, and revocation. We may also find ourselves in an AI versus AI environment, which will require us to combine identity intelligence, behavioral intelligence, transaction intelligence, and real-time risk assessment.
This requires a collective effort. No institution, however large, can build trust and scale alone. Agentic AI will require a coalition. Government and regulators will have to ensure that innovation happens within a framework that protects the customer and the stability of the financial system. Banks will have to become responsible deployers of this technology. Fintechs and technology companies will continue to innovate and build the rails and capabilities on which this system will depend. India's experience has taught us that the biggest digital transformation happens when an entire ecosystem builds infrastructure that everyone can trust. India today is uniquely positioned for the next phase. We have already shown that digital financial infrastructure can serve a billion people. The next opportunity is to make intelligent financial services available to a billion people. I believe this is India's opportunity to move from digital inclusion to intelligent inclusion, from trust in tradition to trust in technology. If we build it with discipline, openness, and a sense of purpose, Agentic AI can become more than the next technological advance in banking. It can become the next enabler of India's financial and economic transformation, worthy of the trust of a billion Indians. Thank you and Jai Hind.
