GFF 2026: After KYC, SBI chief flags ‘Know Your Agent’ as AI enters banking

As artificial intelligence (AI) evolves from merely recommending actions to independently executing them, the banking sector is poised to face new challenges that necessitate the development of a framework akin to the long-established Know Your Customer (KYC) protocol. This was emphasized by Challa Sreenivasulu Setty, the chairman of the State Bank of India (SBI), during his address at the Global Fintech Fest 2026. Setty proposed the concept of 'Know Your Agent' as a critical component for managing the increasing presence of agentic AI in financial services.

Setty articulated that the integration of agentic AI—AI systems capable of acting autonomously—into banking operations will require institutions to establish clear mechanisms surrounding an AI agent's identity, authentication, consent, transaction limits, audit trails, and the ability to revoke actions. He noted that while banks have dedicated decades to refining processes centered around KYC, the emergence of AI agents in financial transactions compels a reevaluation of these frameworks. "As agents begin to participate in financial transactions, we will increasingly need to think about know your agent," Setty stated.

The Shift from Recommendations to Execution

The transition to agentic AI marks a significant evolution in the capabilities of artificial intelligence. Traditionally, AI has been utilized to assist with tasks or follow predefined instructions, but agentic AI represents a leap forward, allowing systems to operate with a degree of autonomy. Setty highlighted that this shift makes accountability paramount, particularly as AI begins to transition from providing recommendations to executing critical actions. He underscored the necessity for an audit trail, traceability, and an understanding of the rationale behind significant actions taken by AI systems.

The banking industry has already begun to leverage AI and machine learning technologies across various functions, including credit assessment, cash flow-based lending, early warning systems for potential issues, fraud detection, anti-money laundering monitoring, and customer service. With the advent of agentic AI, these capabilities could be further enhanced throughout the financial lifecycle, extending to KYC processes, loan appraisals, underwriting, reconciliation, complaint management, and overall customer servicing.

Navigating the Risks of Autonomous Systems

Setty also raised concerns about the potential risks associated with the growing deployment of autonomous systems, which could lead to an "AI versus AI" scenario. This environment would necessitate the incorporation of various forms of intelligence—identity intelligence, behavioral intelligence, transaction intelligence, and real-time risk assessment—into banking operations. The risks posed by errors in autonomous systems could be exacerbated, as mistakes can trigger a cascade of actions across interconnected systems at machine speed, impacting customers, counterparties, and financial institutions alike. Setty cautioned that "the same capability that gives AI its enormous power can therefore amplify the consequences of a mistake."

In light of these challenges, Setty emphasized that trust must be an integral part of agentic AI models, rather than an afterthought. He outlined what he referred to as the three A's for deploying AI in financial services: accuracy, accountability, and equitable access. An AI agent tasked with handling financial services must demonstrate consistent accuracy at a population scale, while also ensuring accountability for its actions. Furthermore, AI technology should be designed to be accessible without bias, adapting to the needs of customers rather than forcing customers to adapt to the technology.

Building Trust in Intelligent Financial Services

Setty pointed out that India's digital public infrastructure has shown that technology can be rapidly adopted when it is simple, affordable, interoperable, and trustworthy. However, as agentic AI begins to permeate the banking sector, he argued that no single institution can build the necessary trust and scale in isolation. Instead, he called for a collaborative approach involving regulators, banks, fintech companies, and technology providers.

He posited that India's next opportunity lies in transitioning from digital inclusion to what he termed "intelligent inclusion," where intelligent financial services become accessible at a population scale. This vision underscores the importance of creating a robust framework that ensures trust and accountability in the deployment of AI technologies within the banking sector, paving the way for a future where financial services are not only automated but also reliable and trustworthy.