Banking executives say autonomous AI could accelerate payments, verification and customer service, but lenders must strengthen authentication, monitoring and accountability before handing agents greater control.
Banks exploring artificial intelligence are likely to gain speed and convenience, but industry leaders say the technology’s rise will only be workable if lenders tighten security, define accountability more clearly and keep people in the loop. At the Business Standard BFSI Summit in Mumbai, executives said AI agents could change how banks process payments, verify customers and handle service requests, yet their growing autonomy would force institutions to rethink authentication, monitoring and risk controls.
Ramesh Lakshminarayanan, group head of information technology and chief information officer at HDFC Bank, said agentic banking was no passing experiment and that lenders should prepare for it rather than resist it. He argued that AI agents would need to be authenticated much like human users and said banks would have to build their own safeguards before regulators set out a fuller framework. He also stressed the importance of near real-time transaction monitoring to spot suspicious behaviour quickly.
Accountability was a recurring theme. Lakshminarayanan said if a customer gives an AI agent permission to make a payment or carry out another task and the system later behaves unexpectedly, responsibility would ultimately rest with the bank. Rajesh Prasad, a director at Punjab National Bank, said lenders have always had a duty to manage risk, even when they take precautions, and he likened the issue to lending decisions that can still go wrong despite due diligence.
Sameer Shetty, group head of digital business, transformation and strategic programmes at Axis Bank, said AI represents a more fundamental change than the move from branch-led banking to digital channels because it can make decisions as well as complete tasks. He said that could simplify customer journeys and automate more internal work, but he warned that customer-facing agents raise tougher governance questions. Banks may understand and control their own systems, he said, but it is harder to determine how to oversee agents acting on behalf of customers using their credentials.
Paritosh Wechalekar, managing director of technology services for India business at Accenture, said banks need sharper boundaries around what AI should and should not do. An agent can safely collect documents, review them and verify information within defined limits, he said, but independent loan approval is a different matter altogether. He added that detecting a suspicious transaction is not the same as declaring it fraudulent, because legitimate payments can look unusual. Nishant Singh, founder and chief executive of Business Next, said the long-term destination is “zero touch” banking, but added that agentic AI does not automatically mean autonomy. He said banks will need stronger identity controls, clear authority for customer agents and a reliable way to shut systems down when something goes wrong.
Disclaimer: This article is intended to inform and educate, not to recommend or endorse any financial product, investment or strategy. Please consider your own financial circumstances and seek professional advice where appropriate before making financial decisions.
