SBI uses AI to underwrite nearly ₹1 lakh cr. MSME loans

1. At a Glance

• SBI (State Bank of India), India's largest lender, used Artificial Intelligence (AI) to underwrite nearly ₹1 lakh crore (₹1 trillion) in MSME loans (loans up to ₹5 crore each) during FY26 (2025-26). [1]
• Illustrates the growing role of AI/data analytics in banking credit assessment, a live GS-III (Science & Tech / Economy) and GS-II (governance/digital public infrastructure) intersection topic. [1]
• Relevant to UPSC's recurring focus on MSME financing, digital lending, financial inclusion, and AI in governance/finance.

2. Why in the News

• On 12 August 2026, SBI MD Rama Mohan Rao Amara announced at the annual FIBAC (FICCI-IBA) event that SBI underwrote nearly ₹1 lakh crore in MSME loans using AI in FY26, covering both new-to-bank and existing customers. [1]

3. Background & Evolution

• SBI has been expanding digital/SME lending initiatives over recent years, including SME Digital Business Loans launched around mid-2024 to sanction loans within 45 minutes. [2]
• By October 2025, SBI had processed SME digital business loans worth ₹74,434 crore since launch, indicating a scaling trajectory toward the ₹1 lakh crore FY26 figure. [2]
• SBI has separately used credit guarantee schemes for MSMEs (e.g., ₹20,000 crore sanctioned to MSMEs under a credit guarantee scheme, reported 2020), showing a longer institutional history of MSME credit support predating the AI-driven push. [2]
• The bank has combined GSTN/GST filings, bank account information, credit bureau scores, Income Tax Returns (ITR), and other unstructured data through its in-house Business Rule Engine (BRE) to enable automated underwriting. [1]

4. Core Static Facts

Item	Detail
Institution	State Bank of India (SBI) — India's largest public sector lender
Amount underwritten via AI	~₹1 lakh crore (₹1 trillion) in FY26 [1]
Loan ceiling per borrower	Up to ₹5 crore [1]
Beneficiary segment	MSMEs — both new-to-bank and existing customers [1]
Technology/engine used	Business Rule Engine (BRE); integrates GSTN, GST filings, ITR, bank statements, credit bureau data, unstructured data [1]
Announcing official	Rama Mohan Rao Amara, Managing Director, SBI [1]
Platform of announcement	FIBAC (FICCI-IBA Annual Banking Conference), 2026 [1]
Related SBI digital lending product	SME Digital Business Loans (launched 2024; sanctions within 45 minutes) [2]
Fastest AI sanction time (reported)	As fast as ~10 seconds once data is submitted, per data-driven credit assessment engine [1]
Additional AI use by SBI	Deploying AI/Large Language Models (LLMs) to automate cheque processing up to ₹10,000 [2]

5. Multi-Dimensional Analysis

Economic

• MSMEs contribute significantly to India's GDP, exports, and employment; faster AI-driven credit access can ease the chronic MSME credit gap and reduce collateral-heavy, delay-prone traditional lending. [1]
• Reduces turnaround time for loan sanction, potentially boosting MSME liquidity and working-capital cycles.

Scientific/Technological

• Demonstrates practical application of AI/data analytics/LLMs in financial underwriting — using alternate data (GST returns, ITR, bank statements) instead of only traditional collateral-based assessment. [1] [2]
• Signals movement toward India Stack-style data integration (GSTN + banking + credit bureau) enabling near-instant credit decisions.

Administrative/Governance

• Reduces dependency on relationship managers for manual data gathering/analysis, freeing bandwidth for high-value client engagement. [1]
• Raises questions on algorithmic accountability, explainability, and grievance redressal when AI systems deny/approve credit at scale.

Social

• Wider, faster access to formal credit for small/new-to-bank enterprises could aid financial inclusion, though risk of algorithmic bias against thin-file or informal-sector borrowers needs scrutiny.

Ethical/Governance

• AI-based underwriting raises data privacy concerns (use of GST, ITR, banking data) and regulatory oversight questions regarding fairness and auditability of automated credit decisions.

6. Recent Developments (last 12-18 months)

• June 2024: SBI launched SME Digital Business Loans enabling sanctions within 45 minutes. [2]
• October 2025: SBI reported processing ₹74,434 crore worth of SME digital business loans since the product's launch. [2]
• 12 August 2026: SBI MD Rama Mohan Rao Amara disclosed at FIBAC that AI-based underwriting crossed nearly ₹1 lakh crore in MSME loans (up to ₹5 crore each) for FY26. [1]
• SBI is simultaneously deploying AI/LLMs for cheque processing automation (cheques up to ₹10,000), indicating a broader digitisation push across retail/SME banking operations. [2]

7. Prelims Hooks

• SBI underwrote nearly ₹1 lakh crore (₹1 trillion) in MSME loans using AI in FY26 (2025-26). [1]
• Maximum individual loan size under this AI-underwriting drive: ₹5 crore. [1]
• Announcement made at the FIBAC (FICCI-IBA) annual banking conference. [1]
• SBI MD who made the announcement: Rama Mohan Rao Amara. [1]
• Data sources integrated for AI underwriting: GSTN, GST filings, bank account data, credit bureau scores, Income Tax Returns (ITR). [1]
• SBI's internal underwriting technology is called the Business Rule Engine (BRE). [1]
• Fastest reported AI-based sanction decision time: as low as ~10 seconds once documentation is submitted. [1]
• SBI's SME Digital Business Loans product (launched 2024) enables loan sanction within 45 minutes. [2]
• Cumulative SME digital loans processed by SBI (as of October 2025): ₹74,434 crore. [2]
• SBI is also using AI/Large Language Models (LLMs) to automate cheque processing for amounts up to ₹10,000. [2]
• SBI is India's largest public sector lender by assets. [1]
• The AI underwriting covers both new-to-bank and existing customers. [1]

8. Mains Relevance

• GS-III: Indian Economy — Mobilization of resources, growth, development, employment; Role of MSMEs; Science & Technology — AI applications in finance/governance.
• GS-II: Governance — Digital initiatives in banking, transparency, accountability in algorithmic decision-making.
• Possible Mains question stems: 1. "Discuss how Artificial Intelligence is reshaping credit underwriting for MSMEs in India. What are the associated risks and regulatory challenges?" (GS-III) 2. "MSMEs remain credit-starved despite multiple government schemes. Examine how AI-driven digital lending by public sector banks like SBI can bridge this gap." (GS-III) 3. "Algorithmic decision-making in public sector banking raises concerns of accountability and bias. Critically analyse in the context of AI-based loan underwriting." (GS-II/GS-IV)

9. Related Topics to Study Next

• MSME definition & classification (Udyam Registration, investment/turnover criteria) — foundational static topic often tested alongside MSME credit news.
• Credit Guarantee Fund Trust for Micro and Small Enterprises (CGTMSE) — key government MSME credit guarantee mechanism.
• RBI's Priority Sector Lending (PSL) norms — MSME lending is a PSL category; relevant to credit flow discussions.
• Digital Public Infrastructure (India Stack) — GSTN, Account Aggregator framework — underlying data architecture enabling AI underwriting.
• RBI's Framework on Responsible/Ethical AI in financial sector (FREE-AI Committee, 2024-25) — regulatory dimension of AI in banking.
• Emergency Credit Line Guarantee Scheme (ECLGS) — earlier MSME credit support mechanism, useful for comparison.
• Financial Inclusion initiatives (Jan Dhan-Aadhaar-Mobile/JAM trinity) — broader digital financial inclusion context.

10. Common Errors / Trap Areas

• Do not confuse ₹1 lakh crore AI-underwritten MSME loans (FY26) with total MSME credit disbursed by all banks — this figure is SBI-specific, not sector-wide.
• Avoid confusing SBI's Business Rule Engine (BRE) with generic "Business Correspondent" (BC) model used for financial inclusion — different concepts.
• Do not misattribute the announcement to RBI or the Finance Ministry — it was made by SBI's own MD at an industry event (FIBAC), not a government/regulatory pronouncement.
• Loan ceiling under this scheme is ₹5 crore per loan, not to be confused with Mudra loan limits (₹10 lakh–₹20 lakh) or CGTMSE cover limits.
• Note the loan figure is reported both as "₹1 lakh crore" and "₹1 trillion" in different sources — these are numerically identical (1 lakh crore = 1 trillion), not two different figures.