FinTech Lending Case Study | Kawach Technology

Client Overview

QuickFund Financial Services

QuickFund Financial Services had grown its loan book past ₹200 crore largely on the strength of its underwriting team's judgment — but that same manual process had become the ceiling on how much further the business could grow. Loan applications were reviewed by hand, and even straightforward cases took five to seven business days to get a decision, mostly because underwriters were manually cross-...

Industry

FinTech / NBFC Lending · FinTech Software Development

Business Size

Mid-size NBFC, ₹200 Cr+ loan book

Location

Pune, India

Business Model

Digital Lending / NBFC

Project Duration

10 months

Existing Challenges

Loan applications were reviewed manually by underwriters, taking 5-7 business days per decision even for straightforward cases.
KYC verification relied on physically collected documents, creating bottlenecks and a poor applicant experience.
Credit risk assessment varied significantly between underwriters, since there was no standardized, data-driven scoring model.
Leadership had no real-time dashboard of loan book health — NPAs, disbursement trends, and collections were compiled manually every month.
RBI-mandated regulatory reports were assembled by hand from multiple spreadsheets every quarter, consuming days of the finance team's time.

The Challenge

What We Were Up Against

QuickFund Financial Services had grown its loan book past ₹200 crore largely on the strength of its underwriting team's judgment — but that same manual process had become the ceiling on how much further the business could grow. Loan applications were reviewed by hand, and even straightforward cases took five to seven business days to get a decision, mostly because underwriters were manually cross-referencing bureau reports, income documents, and bank statements one application at a time.

KYC verification relied on physically collected documents, which meant delays before an application could even enter the review queue. Worse, credit risk assessment varied meaningfully between underwriters — there was no standardized, data-driven scoring model, just individual judgment shaped by each underwriter's own experience. That inconsistency was a real risk, not just an efficiency problem.

Leadership had no real-time view into the health of the loan book. NPAs, disbursement trends, and collections were compiled by hand into monthly reports, meaning problems could go unnoticed for weeks. And every quarter, the finance team spent days manually assembling RBI-mandated regulatory reports from a patchwork of spreadsheets — time that could have gone toward actually managing risk instead of documenting it after the fact.

Our Solution

How We Built It

Because this was a regulated lending business, we started with QuickFund's risk and compliance teams, not the engineering backlog — working through RBI's digital lending guidelines together before designing a single workflow. That sequencing mattered: the credit scoring model we eventually built came directly from translating QuickFund's own underwriters' existing manual assessment logic into a structured, data-driven scoring system, rather than importing a generic off-the-shelf model.

The platform integrates Aadhaar and PAN verification APIs so identity checks that used to take days now complete in minutes, alongside a hybrid rules-based and machine-learning credit scoring engine, e-signature for loan agreements, a real-time portfolio dashboard, and automated RBI regulatory reporting generated directly from live loan book data.

We didn't let the automated scoring engine make a single live decision until it had earned that trust: for six weeks, it ran in parallel with manual underwriting, and we compared every automated recommendation against what an experienced underwriter actually decided. Only once that comparison showed consistent accuracy did QuickFund's compliance team sign off on a phased go-live — starting with smaller loan tickets and expanding to larger ones only after the system built a track record.

The result respects a principle QuickFund was firm about from day one: automation should handle the clear-cut cases, and route anything ambiguous to a human underwriter rather than force a decision either way.

Key Modules Delivered

Digital Loan Origination

End-to-end online application flow from form submission through disbursement, replacing the paper-based process.

Automated KYC Verification

Integrated Aadhaar and PAN verification APIs to confirm applicant identity in minutes.

Credit Scoring Engine

A rules-based and machine-learning hybrid model that scores applications using bureau data, income signals, and repayment history.

E-Sign Loan Agreements

Legally compliant digital signing of loan agreements, eliminating physical paperwork and courier delays.

Portfolio Dashboard

Real-time visibility into disbursements, collections, and NPA trends for leadership and the risk team.

Regulatory Reporting

Automated generation of RBI-mandated reports directly from live loan book data.

Goals & Objectives

What Success Looked Like

Automate Underwriting

Replace manual document review with an automated, data-driven credit decisioning engine.

Speed Up KYC

Verify applicant identity digitally in minutes instead of days.

Standardize Risk Scoring

Apply a consistent, auditable credit scoring model to every application.

Give Leadership Real-Time Visibility

Replace manual monthly reporting with a live portfolio health dashboard.

Features Developed

What We Built

Automated Underwriting

Data-driven credit decisions in minutes instead of days.

Digital KYC

Aadhaar and PAN verification completed in under 3 minutes.

E-Sign Agreements

Legally compliant digital signing, no physical paperwork.

Real-Time Portfolio Dashboard

Live visibility into disbursements, collections, and NPAs.

Automated Regulatory Reports

RBI-mandated reports generated directly from live data.

Technology Stack

Built With the Right Tools

We selected every technology based on this project's real requirements: compliance obligations, scalability needs, and long-term maintainability. No trend-chasing, only battle-tested solutions.

Backend

Laravel 10 PHP 8.2 PostgreSQL

Credit Scoring

Python scikit-learn credit bureau APIs

Integrations

Aadhaar eKYC API PAN verification API e-Sign gateway

Infrastructure

AWS (isolated VPC) encrypted RDS S3

Development Process

How We Delivered It

Agile delivery with regular demos and continuous deployment. Full transparency at every stage.

Total Timeline

10 months

Started → Ongoing

1

01

Regulatory & Risk Discovery

Worked closely with QuickFund's risk and compliance teams to understand RBI lending guidelines before designing any workflow.

2

02

Credit Model Design

Collaborated with QuickFund's underwriters to translate their existing manual assessment logic into a structured, data-driven scoring model.

3

03

Core LOS Development

Built the loan origination workflow and KYC integrations first, since every other module depended on a verified applicant record.

4

04

Parallel-Run Validation

Ran the automated scoring engine alongside manual underwriting for 6 weeks, comparing outcomes before trusting it with live decisions.

5

05

Compliance Review & Sign-Off

Had the full system, including data handling and audit logging, reviewed by QuickFund's compliance team before go-live.

6

06

Phased Go-Live

Rolled out automated underwriting for smaller loan tickets first, expanding to larger ticket sizes only after building a track record of accuracy.

Security & Compliance

Built for the Strictest Standards

RBI Lending Guidelines

The credit decisioning and disclosure workflows were designed in direct consultation with QuickFund's compliance team to align with RBI digital lending guidelines.

Data Localization & Encryption

All applicant and financial data is stored on infrastructure within India, encrypted at rest and in transit.

Audit Trail

Every underwriting decision, automated or manually reviewed, is logged with a full audit trail for regulatory inspection.

Consent-Based Data Use

Applicant data is only used for credit assessment with explicit consent captured during the application flow.

Results / KPIs

Measurable Impact

Numbers measured at 6 months post-launch, independently verified by the client's operations team.

5-7 days → 12 min

Loan Approval Time

< 3 min

KYC Verification Time

+3x applications/day

Underwriting Team Capacity

-90%

Regulatory Report Prep Time

Before vs. After

Before	After
5-7 business days to approve a loan application	~12 minutes for automated decisions on standard applications
Physical KYC document collection	Digital Aadhaar/PAN verification in under 3 minutes
Underwriter judgment varied case to case	Standardized, auditable credit scoring for every application
Manual monthly portfolio reports	Real-time portfolio dashboard for leadership

"

Lending is a business of trust and speed, and we were struggling on both fronts with a manual process. Kawach didn't just automate our underwriting — they built it in a way our compliance team could actually stand behind. We went from a 5-7 day approval cycle to minutes, without cutting a single corner on due diligence.

VI

Vikram Deshmukh

Chief Risk Officer, QuickFund Financial Services

★★★★★

Key Achievements

Why This Project Matters

Beyond the numbers: what this project changed day-to-day for QuickFund Financial Services and the people who rely on what we built.

Approval Time Cut from Days to Minutes

Straightforward loan applications now receive a decision in about 12 minutes instead of 5-7 business days.

3x Underwriting Capacity Without New Hires

The existing underwriting team now handles roughly 3x the application volume by focusing only on edge cases the automated engine flags for review.

Passed Compliance Review on First Submission

The new regulatory reporting module passed QuickFund's internal compliance audit on its first submission, with no findings.

FAQ

Common Questions

Have more questions? Book a call with our team.

How did you validate the automated credit scoring model before trusting it with real decisions?

We ran the model in parallel with QuickFund's manual underwriting process for 6 weeks, comparing every automated decision against what an experienced underwriter would have decided, before allowing it to make live decisions.

Does the system still involve human underwriters?

Yes. The automated engine handles clear-cut approvals and rejections, and routes borderline or high-value applications to human underwriters — it's designed to augment the team, not replace judgment on edge cases.

How is applicant data kept compliant with RBI regulations?

All data handling, consent capture, and disclosure workflows were built in direct consultation with QuickFund's compliance team and reviewed against RBI's digital lending guidelines before go-live.

What happens if the automated system is unsure about an application?

Applications that don't clearly meet or fail the scoring criteria are automatically flagged and routed to a human underwriter rather than being auto-approved or auto-rejected.

Related Case Studies

How QuickFund Financial Services Automated Loan Underwriting and Cut Approval Time from Days to Minutes