Case Study

How a ₹1,000 Crore NBFC Cut Underwriting Time by 75% with an AI BRE & Co-Pilot

LendingIQ redesigned underwriting for a large Indian NBFC, deploying an AI-powered Business Rules Engine (BRE) and underwriter co-pilot - enabling credit teams to configure live policy rules and evaluate every application against them in real time.

Client

Leading ₹1,000 Crore Indian NBFC

Domain

MSME Lending

Function

Underwriting Automation

Role

AI Underwriting System Design & Build

75%

Reduction in underwriting time

3-4x

Increase in underwriter productivity

100%

Applications evaluated against live BRE policy rules

A fast-growing NBFC processing large MSME loan volumes was constrained by two compounding problems: analysts were spending most of their time extracting and normalizing data before they could begin credit judgment - and the credit policies guiding those decisions lived in spreadsheets and institutional memory, applied differently by different people under volume pressure.

The Problem: Manual Underwriting & No Policy Enforcement Layer

The NBFC needed comprehensive analysis across statements, GST returns, transaction histories, and bureau inputs for each application. The process was manual, time-heavy, and impossible to scale as volumes grew.

Beneath the speed problem was a structural policy problem: credit rules existed in documents and in people's heads. There was no enforcement layer - no system that applied the same rules, in the same way, to every application in real time. Risk variance was invisible until it became a problem.

No live policy enforcement

Credit policies existed as documentation and spreadsheets - not as a system that evaluated applications in real time against current rules.

Rules too slow to update

When credit policy changed, communicating and applying that change across the team was manual, delayed, and impossible to verify at scale.

Manual financial analysis

Underwriters spent hours parsing raw bank statements, GST data, and financial reports with no automated extraction layer.

Slow decision turnaround

Each loan consumed several analyst hours before a credit recommendation could be produced.

Inconsistent risk assessment

Different analysts interpreted the same policy differently under pressure, creating variance in credit outcomes and audit exposure.

Limited team capacity

Underwriters were stretched thin, limiting daily application throughput and the NBFC's ability to grow its MSME book.

"The bottleneck was not a talent problem. It was a workflow and policy enforcement problem."

The Solution: AI BRE + Underwriting Co-Pilot

LendingIQ built a two-layer AI underwriting system integrated directly into the NBFC's existing LOS. The Dynamic AI BRE evaluates each application against dynamic ruleset in real time. The underwriter co-pilot then surfaces those BRE outputs alongside AI-generated risk summaries so analysts can make faster, better-informed decisions.

The design principle: automate data extraction, enforce policy systematically via BRE, and preserve human judgment for final credit decisions.

Capability 1

AI Dynamic BRE

Every application is evaluated against the ruleset appropriate for that application. Decisions are consistent, instant, and grounded in the most current credit policy - not a static snapshot.

Capability 2

Automated document processing

Unified ingestion and AI parsing for all incoming financial document formats - statements, GST returns, bureau reports - with no manual preprocessing.

Capability 3

AI financial analysis

Automated extraction of revenue trends, cash flow patterns, and behavioral risk markers across all document inputs.

Capability 4

Underwriting co-pilot

BRE outputs and AI-generated risk narratives delivered inside the analyst workflow, keeping underwriters focused on judgment rather than data assembly.

System Architecture

Six connected layers drive underwriting from document ingestion through real-time BRE evaluation and analyst decision support.

Document ingestion

Bank statements, GST returns, financial reports, and bureau data are pulled directly from LOS workflows - no manual handoff required.

AI document processing

Structured and unstructured document formats are parsed automatically, eliminating manual preprocessing as a dependency.

Financial signal extraction

Revenue patterns, expense behavior, cash-flow stability, and anomalies are surfaced automatically across all input sources.

AI BRE - Real-Time Policy Evaluation

Extracted signals are evaluated instantly against the NBFC's live credit ruleset. The credit team configures eligibility thresholds, risk triggers, and decisioning logic directly - no engineering dependency. Every application hits the same rules, at the same time, with zero interpretation variance.

AI credit scoring

Borrower health is scored within the BRE policy framework, producing an explainable risk output grounded in live, enforced policy logic.

Underwriting co-pilot interface

Analysts receive BRE decision outputs, policy flags, and AI risk summaries inside their existing workflow - ready for final human review.

The Impact

Within weeks of deployment, underwriting speed, analyst capacity, and policy consistency all improved materially - with the credit team in direct control of the rules driving every decision.

Underwriting time dropped by 75% across core MSME loan workflows.

Each underwriter processed 3 to 4 times more applications per day.

100% of applications are now evaluated in real time against live BRE policy rules - zero manual interpretation.

Credit team can update and enforce new policy rules immediately, with no engineering dependency or rollout delay.

Risk variance across analysts eliminated - every decision reflects the same live ruleset.

Credit outputs are fully standardized, policy-traceable, and audit-ready.

What This Means for MSME Lending

When credit policy is configurable, enforced in real time, and no longer dependent on individual interpretation, lenders can scale underwriting without scaling risk.

Credit team controls policy rules directly - no engineering dependency

Real-time BRE evaluation on every application

Zero policy interpretation variance across the team

Higher underwriting throughput with existing headcount

Faster borrower decisions with consistent risk standards

Scalable MSME lending without proportional team growth