Use case #0002

How Early Warning AI Reduces Gross NPA by Alerting Collections 60 Days Early

The difference between a 3.8% GNPA ratio and a 2.6% GNPA ratio is not luck, not economic conditions, and not a better underwriting model. It is the 60-day advantage that early warning gives the collections team. When a borrower is contacted before their first missed payment rather than after it, the resolution rate is between 2 and 3 times higher. That resolution rate difference is what moves the NPA needle.

The difference between a 3.8% GNPA ratio and a 2.6% GNPA ratio is not luck, not economic conditions, and not a better underwriting model. It is the 60-day advantage that early warning gives the collections team. When a borrower is contacted before their first missed payment rather than after it, the resolution rate is between 2 and 3 times higher. That resolution rate difference is what moves the NPA needle.

The Resolution Rate Difference — Why 60 Days Is the Decisive Variable

Collections recovery rates are not uniform across the delinquency lifecycle. They follow a predictable decay curve: the probability of resolution drops sharply as the borrower moves from pre-delinquency through the DPD buckets. A borrower contacted proactively while still current — before any payment has been missed — resolves at approximately 68 to 74% (restructuring, commitment, or self-cure). A borrower contacted at DPD 1–30 resolves at approximately 51 to 58%. At DPD 31–60, the resolution rate is 38 to 44%. By DPD 61–90, it has fallen to 22 to 31%. And once an account crosses into NPA at DPD 90, the expected recovery through resolution (not legal) falls below 20%.

The 60-day early warning advantage does not just move the collection action earlier — it moves the borrower to a different point on the resolution rate curve. The same borrower, contacted at the Amber alert stage rather than the DPD 1 stage, is 1.5 to 2x more likely to resolve their situation before it becomes a permanent credit mark. At portfolio scale, this difference translates directly to GNPA percentage points.

"The 60-day early warning advantage is not an operational improvement. It is a financial outcome — measured in GNPA percentage points, provision charges avoided, and recovery costs that never had to be incurred." — Early Warning AI · NPA Impact Module

Before and After: Portfolio Impact of EWS Deployment

Without Early Warning AI
GNPA ratio3.8%
First contact: DPDDPD 1–15 (post-bounce)
Pre-delinquency resolution rateNot measured
DPD 1–30 resolution rate52%
NPA entry rate (% of active book)1.8% per quarter
Provision charge (annual)₹48Cr
Collection cost per account₹4,840
Avg DPD at first contact8 days
With Early Warning AI (12 months post-deployment)
GNPA ratio2.6% (−1.2pp)
First contact: DPDDPD −60 (pre-delinquency)
Pre-delinquency resolution rate71%
DPD 1–30 resolution rate58% (+6pp)
NPA entry rate (% of active book)1.1% per quarter (−38%)
Provision charge (annual)₹31Cr (−₹17Cr saved)
Collection cost per account₹2,920 (−40%)
Avg DPD at first contact−58 days (pre-delinquency)

The Portfolio Stress Heatmap: Where the NPA Risk Lives Right Now

The GNPA reduction is not a future aspiration — it is driven by the current portfolio stress heatmap that the Early Warning AI produces. This view shows exactly where in the active portfolio the stressed accounts are concentrated — by product, geography, vintage, and segment — so that the collections team can prioritise its intervention resources toward the accounts where early action will have the highest impact.

Portfolio Stress Distribution — Current Month
Active Book · EWS Score Distribution by Segment
Segment Normal (0–30) Watch (31–55) Amber (56–74) Red (75–89) Critical (90+) Total AUM at Risk Priority Action
Home Loan — Salaried 88% 8% 3% 0.8% 0.2% ₹8.4Cr Standard monitoring
LAP — Self-Employed 72% 14% 8% 4% 2% ₹42.1Cr RM outreach — Amber accounts
MSME — Manufacturing 68% 16% 9% 5% 2% ₹61.3Cr Sector stress — escalate to CPO
Personal Loan — Digital 76% 12% 7% 3.5% 1.5% ₹28.4Cr Velocity intervention on Red band
Affordable Housing 82% 10% 5% 2.2% 0.8% ₹18.7Cr Monsoon geography — geo-flag review

The Financial Case: What ₹17Cr in Saved Provision Means

The provision charge reduction in the before/after comparison — ₹17Cr annually for a mid-tier NBFC — is not the ceiling of the EWS financial benefit. It is the most directly measurable component. The full financial impact also includes: reduced collection cost (₹2,920 vs ₹4,840 per account — a 40% reduction across the collections cost base); reduced legal and recovery cost for accounts that resolve pre-delinquency versus those that go to NPA; improved NIM as the provision coverage ratio is maintained at lower actual NPA levels; and reduced regulatory scrutiny for institutions whose GNPA trajectory shows consistent improvement.

The GNPA ratio improvement of 1.2 percentage points — from 3.8% to 2.6% — has implications beyond the balance sheet. In a market where lenders are benchmarked by asset quality, a sustained 2.6% GNPA is a competitive and regulatory asset. It enables better credit ratings, lower cost of borrowing, and the institutional confidence to grow the loan book faster — because the risk management infrastructure is demonstrably working.

−1.2ppGNPA reduction: 3.8% → 2.6% in 12 months post-deployment
71%Pre-delinquency resolution rate — vs 52% at DPD 1–30
−38%NPA entry rate reduction — fewer borrowers crossing into 90 DPD
₹17CrAnnual provision charge saved at benchmark mid-tier NBFC portfolio size

The Best Collection Is the One That Never Has to Happen

A borrower who receives a proactive restructuring call at EWS Amber stage — before any payment failure — is not just 71% more likely to resolve than a borrower who receives a collections call at DPD 15. They are a borrower who never appears in the NPA register, never requires provisioning, never costs a legal fee, and never becomes a write-off. The Early Warning AI's financial case is not built on recovering more from bad loans. It is built on preventing good loans from becoming bad ones — which is always cheaper, faster, and less damaging to the institutional balance sheet and borrower relationship simultaneously.

← Back to Early Warning Agent AI