SBI Uses AI for ₹1 Trillion MSME Loans

State Bank of India (SBI) significantly expanded its lending to micro, small, and medium enterprises (MSMEs) by leveraging artificial intelligence (AI), underwriting approximately ₹1 trillion in loans during the fiscal year 2025-26. This development was revealed by SBI Managing Director Rama Mohan Rao Amara at a FICCI-IBA conference.

AI-Driven Lending Scale and Mechanism

• ₹1 trillion in MSME loans underwritten.
• Target period: fiscal year 2025-26.
• Individual loans offered up to ₹5 crore to new and existing customers.
• Automated cheque processing for amounts up to ₹10,000 using Vision LLMs.

The core underwriting relies on an AI-backed Business Rule Engine (BRE) that integrates various data sources. This includes information from the Goods and Services Tax Network (GSTN), GST filings, account details, credit bureau scores, and other unstructured data. The integration aims to significantly free up relationship managers’ time.

Diverse AI Applications and Benefits

• Lower non-performing assets (NPAs) observed in the AI-underwritten portfolio.
• Early warning systems for identifying vulnerable exposures.
• Fraud detection capabilities.
• Security operations enhancements.
• Generative AI bots handle most customer care calls.

A pilot program is underway for an agentic AI-based digital assistant for corporate banking, embedded in YONO Business. This assistant is designed to streamline data collection, document processing, and risk analysis, allowing relationship managers to focus on validation and assessment.

While the quantifiable impact on the bank’s cost-to-income ratio remains under assessment, SBI highlights intangible benefits. These include enhanced customer satisfaction and optimized utilization of manpower resources.