ABOUT THE COMPANY
A legacy bank modernizing SME lending
This Top-5 PSB manages over ₹2 lakh crore in assets with 7,000+ branches across India. Their SME division, though growing at 18% YoY historically, was losing market share to faster fintech lenders. With a ₹400 crore SME book, they wanted to accelerate growth while maintaining RBI compliance. Their IT and credit teams are conservative but pragmatic,they needed a system that fit their VPC-isolated infrastructure and audit requirements.
INDUSTRY CONTEXT & DEMAND
PSBs face the SME lending speed challenge
India's SME credit market is growing 35% annually, but government schemes like ECLGS have created both opportunity and complexity. Lenders must onboard quickly to capture scheme-backed lending before quotas fill, but RBI's scrutiny on PSB credit standards has tightened. PSBs need AI-driven underwriting that passes board audits, integrates with legacy systems, and can run in isolated VPC environments. Speed without compliance is no longer acceptable.
THE CHALLENGE · BEFORE
14 days per SME, legacy systems, low ECLGS uptake
SME loan processing averaged 14 days per applicant, involving manual document collection, GST/ITR verification, credit committee scoring, and compliance checks. The bank's ECLGS utilization was only 54%,well below peer average of 75%,because applicants withdrew after waiting a week. Manual credit analysis required senior officers, bottlenecking throughput. Internal auditors found 8% of files with missing compliance checks.
- 14-day processing cycle vs. 36-hour industry standard for fintechs
- SME book growing at 11% YoY; peers at 20%+
- ECLGS utilization 54%; peer average 75%
- 8% audit findings on credit file completeness per quarter
- Senior credit officers (₹2Cr+ annual cost) doing entry-level verification
HOW AICA HELPED · THE SOLUTION
VPC-isolated, audit-ready AI that runs on PSB infrastructure
AICA was deployed in the bank's Virtual Private Cloud with ISO 27001 encryption and 24-hour audit logs,meeting their IT security requirements. The system integrated with legacy GST and ITR APIs, automated 250+ credit checks, and generated RBI-compliant underwriting memos in parallel. Monitoring flagged borrowers in early warning before default, reducing NPL risk. The entire solution was audit-clean from day one.
- AI Underwriting, 250+ checks (bureau, litigation, financials) in 6.2 seconds, replacing 6 hours
- VPC Integration, Deployed in bank's isolated cloud; zero data egress
- Legacy APIs, Connected to existing GST, ITR, and CIBIL systems
- Continuous EWS, Daily default monitoring with 250+ early warning signals
THE OUTCOME · AFTER
36-hour processing, 22% SME growth, 87% ECLGS utilization
Processing time fell from 14 days to 36 hours, cutting borrower dropout by 65%. SME book growth accelerated to 22% YoY (vs. 11% pre-deployment). ECLGS utilization jumped to 87%, generating ₹160+ crore in subsidized lending in FY25. Internal auditors found zero compliance gaps in the first post-deployment audit. Credit team capacity freed 1.4 FTE for higher-value work.
- Processing time: 14 days → 36 hours (99% reduction)
- SME book growth: 11% → 22% YoY
- Audit findings: 8% of files → 0%
- ECLGS utilization: 54% → 87%
- Credit FTE freed: 1.4
- VPC deployment: ISO 27001, zero audit gaps