Part 05 — The Lending Lifecycle

Pre-Screening

Purpose

Pre-screening is the controlled reject-or-route stage before a lender spends money on full underwriting, field investigation (FI), legal or valuation. It protects customer experience and credit cost by catching obvious ineligibility early: duplicate exposure, wrong pincode, insufficient vintage, poor bureau, prohibited business, over-leverage or mismatch between requested amount and observable cash flow. The output is not a sanction. It is an eligibility disposition that decides whether the file proceeds to KYC onboarding and application documents.

Actors

Actors are sales operations, credit policy, bureau operations, risk analytics, fraud/RCU, compliance, branch/RM and sometimes partner API operations. A credit underwriter may not touch every pre-screened case, but policy owns the rule set and exception path. Sales should not be able to override negative matches without documented authority.

Inputs & documents

Inputs are PAN, GSTIN, CIN/limited liability partnership identification number (LLPIN), Udyam Registration Number, mobile, date of birth/incorporation, pincode, business vintage, industry/NIC code, requested loan amount, product, existing lender exposure if declared, consent, source, bank name and rough turnover. Documentary inputs are usually not required yet, except where the lender needs GST certificate, Udyam certificate or basic bank statement to avoid wasting a bureau pull.

Step-by-step workflow

Perform internal dedupe using PAN, GSTIN, CIN/LLPIN, mobile, email, bank account, device id, address, promoter names and fuzzy business names.
Check existing customer exposure: live loans, past write-off, settlement, fraud mark, open service complaints, unresolved document deficiency and group exposure.
Apply pincode and branchability rules: serviced geography, negative areas, distance from branch/collection hub and collateral jurisdiction if secured.
Apply industry filters: banned activities, high-cash or high-fraud segments needing senior approval, seasonal businesses needing special underwriting.
Run basic policy: age 21-65/70 depending product, business vintage 2-3 years for normal unsecured business loans, minimum turnover, minimum bank credits, GST filing availability, no severe recent delinquency.
Pull bureau only after consent. For proprietorships and many micro enterprises, pull both proprietor consumer bureau and commercial bureau where available.
Estimate indicative eligibility using turnover, fixed obligation to income ratio (FOIR), debt service coverage ratio (DSCR), average bank balance (ABB), bank credits and existing obligations.
Return one of four outcomes: pass, reject, refer-to-credit or data-insufficient.

Exceptions & edge cases

Dedupe is rarely exact. A proprietorship may use the proprietor PAN, a trade name in GST, a shop electricity bill in the spouse’s name and a current account with a shortened name. Good systems score dedupe confidence rather than forcing exact matches: PAN/GSTIN exact match is hard stop; mobile and address match may be soft; device and bank-account match may be fraud-review.

Negative industry rules also need nuance. A liquor retailer may be allowed in some lenders’ LAP policy but not unsecured policy. A contractor with government receivables may look strong on invoices but weak on bank balance because payments are lumpy. A seasonal school-uniform manufacturer should not be rejected solely for three low-credit months if annual GST and bank credits reconcile.

For bureau pre-screening, the lender must not treat CIBIL score or CMR as the whole credit decision. A thin-file borrower with secured cash flows may be referable; a high score with recent over-utilisation, cheque bounces or many enquiries may still be risky.

Systems touched

LOS, internal customer master, group exposure system, bureau gateway, fraud negative list, watchlist/sanctions screening, pincode master, product-policy engine, BRE, consent store, partner API gateway and audit log. The BRE should version every rule and store the rule-result tree so a later audit can explain why an application was rejected or routed.

TATs/SLAs

Digital pre-screening should return in 30 seconds to 5 minutes if all identifiers are available. Assisted sales pre-screening is usually same day. Manual refer cases should be cleared in 4-8 working hours for unsecured tickets up to ₹50 lakh and within 1-2 working days for larger secured proposals. If a bureau or GST service is down, the case should enter a retry queue rather than being silently rejected.

Stage metrics

Track pre-screen pass rate, hard-reject reasons, dedupe hit rate, false-positive dedupe reversals, bureau consent drop-off, bureau no-hit/thin-file rate, policy-refer rate, rule override rate, subsequent approval rate of pre-screen pass cases, early delinquency by pre-screen band and TAT by external dependency. Policy teams should review rules that reject many files but later show low bad rates when manually approved.

Co-lending/partner-origination variant

In partner journeys, pre-screening may run twice: first at the partner or originating entity, then at the funding partner. The 2025 co-lending directions define a co-lending arrangement as an ex-ante agreement between an originating regulated entity and partner regulated entity to jointly fund loans in a pre-agreed proportion with revenue and risk sharing (RBI Co-Lending Arrangements Directions, 2025). The system must distinguish “partner eligible” from “lender eligible”; a partner pass cannot automatically become a lender sanction.

For LSPs, RBI’s 2025 digital lending rules require the regulated entity to assess borrower creditworthiness and retain the necessary economic-profile information before extending a loan (RBI Digital Lending Directions, 2025). Therefore, even if a platform provides a risk grade, the regulated entity must store its own pre-screen result and data snapshot.