Use case #0002

New-to-Credit Borrowers in Tier 2 Cities: A Thin-File AI Case Study

Rajkot. Nashik. Coimbatore. Madurai. Mysuru. These are not emerging markets — they are established, growing economies where first-generation entrepreneurs run businesses generating lakhs of rupees monthly. They are also cities where a substantial proportion of those entrepreneurs have no CIBIL file. This is the case study of how Thin-File AI served one of them.

Rajkot. Nashik. Coimbatore. Madurai. Mysuru. These are not emerging markets — they are established, growing economies where first-generation entrepreneurs run businesses generating lakhs of rupees monthly. They are also cities where a substantial proportion of those entrepreneurs have no CIBIL file. This is the case study of how Thin-File AI served one of them.

The Tier 2 Credit Gap Is Not a Risk Problem — It Is a Distribution Problem

Tier 2 and Tier 3 cities represent India's fastest-growing MSME economy. The National Sample Survey and MSME ministry data consistently show that Tier 2 cities now account for a larger share of new business registrations than Tier 1 cities. Yet the credit penetration in these cities — formal credit from regulated lenders — remains a fraction of what Tier 1 cities receive per unit of economic activity. The gap is not explained by higher risk. It is explained by the fact that Tier 2 credit markets have historically been dominated by informal moneylenders and bank branch infrastructure that required paperwork profiles these borrowers simply do not have.

Thin-File AI is the infrastructure that makes formal credit viable in this market — not by lowering credit standards, but by applying better measurement to creditworthiness that already exists.

"The small business owner in Nashik who has never borrowed formally is not an emerging credit risk. She is an established creditworthy borrower who has been lending to herself — funding growth from cash flow for a decade — because no institution had the tools to see her clearly."
Borrower Profile — Application TF-2025-2241
Nashik · Maharashtra · Pharmaceutical Distributor
Name (masked)Mr V.K. (abbreviated)
Age38 years
LocationNashik, Maharashtra
Business typePharmaceutical distribution
Business vintage9 years
GST registrationSince 2018 (Day 1 of GST)
CIBIL scoreN/A — no bureau file
Prior formal loansNone — always self-funded
Loan requested₹28L working capital LAP
CollateralRegistered commercial premises
GST annual turnover₹1.84Cr (FY25)
Bank avg balance₹4.2L (18-month avg)
Turnover trend+28% CAGR over 6 years
Filing regularityOn time: 27 of 28 quarters
ITC utilisation68% — healthy ratio
B2B revenue share91% — stable, institutional buyers
Seasonality CoV0.14 — very low, consistent
Monthly UPI inflows₹14.2L avg (12 months)
UPI inflow consistencyCoV 0.12 — very stable
NACH / cheque returnsZero in 5 years
SIP / investment deductions₹18,000/month — regular
Insurance premiumsPaid on time — 3 policies

The Bureau-Only Decision vs the Thin-File Decision

Bureau-Only Underwriting (Prior Industry Standard)
CIBIL scoreN/A — no file
Eligibility assessmentIneligible — no bureau data
Application outcomeDeclined — auto-rejection
Reason provided"Insufficient credit history"
Path to eligibilityBuild bureau file via small credit card — 18–24 months
Borrower outcomeApproached informal lender at 24% p.a.
Lender's lost opportunity₹28L LAP at 11.4% — never originated
Thin-File AI Assessment
Thin-File ScoreTFS 748 — equivalent B+ bureau band
Eligibility assessmentEligible — strong alternative data
Application outcomeApproved — ₹24L at 11.4% (LTV 72%)
Reason providedFull score explanation with 6 signal factors
Path to credit improvementActive — first bureau entry created on disbursement
Borrower outcomeLoan disbursed in 7 days. 14 months clean repayment.
Lender's portfolio impact₹24L performing loan + bureau file for future lending

The Journey From Application to Disbursement

Day 1
09:14
Application Received

Application Submitted via Mobile App — Thin-File Pathway Selected

Application received for ₹28L LAP. System detects: CIBIL = N/A, no bureau file. Thin-File AI pathway automatically activated. Consent request sent for GST data pull, bank statement (AA framework), and UPI transaction data (AA aggregator). All three consents received within 40 minutes.

Day 1
11:38
Data Ingestion Complete

28 Quarters of GST, 18 Months of Bank Statements, 12 Months UPI

Thin-File AI pulls 7 years of GST filing data via GSTN API (consent-based). Bank statement data retrieved via Account Aggregator framework — 18 months, 2,841 transactions classified. UPI transaction history from linked account — 12 months, 4,218 transactions. Property valuation request dispatched to empanelled valuer.

Day 1
11:52
Score Computed — TFS 748

28 Metrics Computed · Score Band B+ · Recommendation: Approve at ₹24L

Thin-File Score computed: 748 (B+ equivalent). LTV constraint applied: property valuation estimate ₹33.2L → maximum loan ₹24L at 72% LTV. Recommended rate 11.4% (standard B+ pricing). FOIR check: ₹24L loan EMI ₹24,800 against estimated monthly income ₹1.18L → FOIR 21.0% — well within 45% limit. Policy gate: all checks passed. Decision: Approve ₹24L.

Day 3
Property Valuation Received

Registered Commercial Property Valued at ₹34.1L — LTV 70.4%

Empanelled valuer confirms ₹34.1L (revised upward from estimate). LTV recalculated: ₹24L / ₹34.1L = 70.4% — within 80% maximum for commercial LAP. Title search initiated: RERA check, registered deed search, encumbrance certificate. Title clear — no disputes found.

Day 5
V-KYC Completed

Video KYC — Identity Confirmed · Business Address Verified

V-KYC session completed. Aadhaar OTP verified. PAN match confirmed. Business address verified against GST registration address — match confirmed. Borrower explains use of funds: pharmaceutical inventory build-up for Q4 festival demand season. Stated purpose aligns with business seasonality data in GST filing history.

Day 7
Disbursed

₹24L Disbursed. NACH Activated. First Bureau Entry Created.

Loan disbursed to borrower's linked current account. NACH mandate activated — EMI ₹24,800, 15th of each month. CIBIL reporting initiated: borrower now enters the formal credit system for the first time. 14 months later: all payments on time, zero DPD. Borrower returned for a second loan — this time with a CIBIL score of 742.

The Portfolio-Level Case for Tier 2 Thin-File Lending

The case study above is illustrative — but the portfolio-level data tells the more compelling story. Across a sample of 2,400 thin-file loans originated in Tier 2 cities using the Thin-File AI model (GST + UPI + bank statement scoring) in FY2024, the 12-month default rate was 3.2%. Across a comparable population of bureau-scored borrowers with CIBIL 700–740 originated in the same period, the 12-month default rate was 2.9%. The difference — 0.3 percentage points — is more than offset by the rate premium that thin-file borrowers carry (50 to 75 basis points above bureau-equivalent pricing) and by the portfolio diversification value of Tier 2 geographic exposure.

The institution that originates quality thin-file loans in Tier 2 cities does not just add a socially valuable credit product — it builds a portfolio of first-time borrowers who will, within 12 to 18 months, have a CIBIL score they did not have before, a repayment record that validates the alternative data prediction, and a demonstrated loyalty to the institution that gave them their first formal credit relationship.

7 daysApplication to disbursement — same TAT as bureau-scored LAP pipeline
3.2%12-month default rate across Tier 2 thin-file cohort — within 0.3pp of bureau-scored equivalent
+75bpsRate premium on thin-file loans — compensates for slightly higher default rate with margin
742CIBIL score of this borrower on their second application — 14 months after first loan disbursement

The First Loan Is the Most Important Loan

The value of originating a borrower's first formal credit relationship is not captured in the economics of the first loan alone. The institution that gives a creditworthy Tier 2 entrepreneur their first formal credit relationship gains a borrower who, when they need their next loan — larger, longer, more complex — has a reason to return. Thin-file lending in Tier 2 cities is not philanthropy and it is not risk management compromise. It is the most efficient form of loan book growth available in a market where bureau-scored borrowers in Tier 1 cities are already multiply-approached by every lender with a digital origination channel.

← Back to Thin-File Credit Agent AI