SBI Chief Urges ‘Know Your Agent’ Checks as AI Moves From Advice to Execution
C.S. Setty says autonomous AI will require stronger identity and accountability controls as financial systems move from recommendations to execution.
Topics
News
- Anthropic Discloses Fourth Claude Breach Missed in Earlier Review
- Visa, Mastercard, Ant Push Common Standards for AI Agents Making Payments
- OpenAI Calls for Mandatory US Frontier AI Safety Rules
- NPCI Launches Two Agentic AI Platforms
- SBI Chief Urges ‘Know Your Agent’ Checks as AI Moves From Advice to Execution
- At Fintech Fest, PM Pushes For Global UPI Links to Cut Costs for Indians Abroad
State Bank of India Chairman C.S. Setty said banks will increasingly need to identify and authenticate AI agents acting on behalf of customers as autonomous systems begin executing financial transactions.
Speaking at the Global Fintech Fest 2026 in Mumbai on Thursday, September 10, Setty said agentic AI changes the risk for financial institutions because it can move beyond advising customers to acting for them.
“The moment AI moves from recommendation to execution, accountability becomes even more important,” Setty said. “There must be an audit trail, traceability, and the ability to understand why an important action was taken.”
Setty said AI agents could be deployed in areas including loan appraisal and underwriting, Know Your Customer and anti-money laundering checks, and fraud detection. But he said systems carrying out financial actions would face a much higher standard of reliability.
“There is nothing like 99% accuracy here. You need to be 100% accurate and every time, all the time,” he said.
Setty described accuracy, accountability and “access without asymmetry” as three requirements for building trust in agentic AI at scale.
He said the spread of autonomous agents would also require banks to extend the identity controls they have built around customers.
“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,” Setty said.
The distinction matters because agentic systems can respond to changing conditions and take actions across financial systems rather than simply generate recommendations.
Setty said that raises questions over who or what is acting, what authority an agent has and who is accountable for its decisions.
SBI is already using AI in its operations. The bank has said its pre-approved personal-loan product uses predictive AI and that it has deployed generative-AI tools for employees handling customer-service and compliance queries. It has also been applying AI to fraud and risk-management systems.
Setty also warned of an emerging “AI versus AI” environment as automated systems increasingly interact with other automated systems, raising the importance of real-time identity, transaction and behavioural checks.
The issue is already moving beyond individual banks. Separately on Thursday, Ant International, Mastercard and Visa said they had begun work on a Know-Your-Agent interoperability framework intended to let payment networks, digital wallets, marketplaces and other platforms identify trusted AI agents across different systems.
The companies said their proposed framework would cover operator traceability, common certification requirements and continuous transaction monitoring while allowing each network to retain its own approval and risk controls.
