Case Study of SRG HOUSING FINANCE

Turning EMI Collections Into A Continuous Recovery Journey.

SRG Housing Finance partnered with Fundamento to move beyond generic collection calling, creating bucket-specific AI journeys that adapt to borrower behaviour—from pre-due reminders through deeper delinquency and NPA intervention.

Written By

Team Fundamento

Published

Managing Repayment Across A Long-Cycle Housing Finance Portfolio

SRG Housing Finance operates in a lending environment where borrower repayment extends across a longer lifecycle and collection intervention must change as an account ages.

Fundamento’s deployment focused on making each stage of that lifecycle operationally distinct.

A Borrower Who Promises To Pay Is Not The Same As A Borrower Who Has Paid.

SRG needed a collection system capable of distinguishing early reminders from deeper recovery situations while continuously verifying whether promised payments had actually happened.

The challenge was compounded by wrong numbers, refuse-to-pay behaviour and customers claiming that payment had already been made.

  • PTP without payment
  • Refuse-to-pay responses
  • Blank / ambiguous responses
  • Wrong-number records
  • “Already paid” disputes
Stage Operational Need
Pre-Due Gentle repayment reminder
Post-Due 0–30 Settlement / payment nudge
Post-Due 30–90 Stronger recovery intervention
NPA 90+ Consequence education + branch / field action
Turn every borrower response into the next operational action—not simply another dial.

Signal OS™ Turned Borrower Responses Into Collection Decisions.

The deployment used DPD position, PTP status and interaction outcomes to determine whether the account should close, be reworked, reclassified or escalated.

Signal What It Revealed
DPD / Bucket Current repayment position
PTP Status Promise made vs no promise
Payment Status Whether the promise translated into payment
Refuse-to-Pay Separate behavioural cohort
Blank Response Requires reclassification
Wrong Number Requires data correction
Already Paid Claim Requires ledger verification
Architecture Layer Deployment Capability
Customer Signals DPD • PTP • payment status • response disposition
Decision Intelligence Bucket-specific next action
Journey Intelligence Pre-Due • Post-Due • NPA
Execution Layer AI Voice + WhatsApp / PayU
Human Intelligence Branch / field routing + ledger operations
Account Enters Bucket → AI Outreach → Borrower Response → PTP / No PTP / Refuse / Blank → Next Action → Payment Verification → Close / Rechurn / Escalate

Every Collection Bucket Gets Its Own

Journey AI Treatment Next State
Pre-Due Gentle reminder Payment / PTP
Post-Due 0–30 Settlement nudge Payment / PTP / Rechurn
Post-Due 30–90 Stronger recovery intervention Payment / PTP / Escalation
NPA 90+ Consequence education + routing Branch / field recovery
Borrower Response Next Action
PTP Given Verify payment / close
PTP Not Given Rechurn
Blank Reclassify
Refuse To Pay Separate recovery track
Already Paid Ledger verification
Wrong Number Correction within SLA
Pre-Due → Post-Due → Deeper Delinquency → NPA → Field / Branch Recovery

The Agent Changes Its Conversation As The Borrower Moves Through Risk.

Agent / Layer Role
PTP Logic Captures and classifies payment commitment
Rechurn Logic Re-engages borrowers requiring another intervention
WhatsApp / PayU Enables live payment capture
Verification Workflow Checks payment status after interaction
Routing Layer Pushes deeper recovery cases to branch / field teams
Bucket + Behaviour → Conversation → Disposition → Next Action

AI Handles The Conversation. Operations Validate And Complete The Recovery.

Trigger Human Team Outcome
NPA / Deep Overdue Branch / field collections In-person recovery
Already Paid Claim Operations Ledger reconciliation
Physical Recovery Requirement Field team Same-day action
Payment Dispute Operations Verified account status
Human intervention is triggered by operational necessity—not by every borrower interaction.

Capabilities Are Not Projected. They Are Measured.

Outcome Area After Fundamento
Collection Model Bucket-specific AI journeys
Response Handling PTP / no PTP / refuse / blank differentiated
Payment Capture WhatsApp / PayU integrated into journey
Data Correction Wrong-number SLA built into operations
Recovery Handoff Branch / field routing for NPA
The deployment changed collections from repeated calling into a response-driven recovery system.

Every Borrower Response Created A Better Next Action.

Intelligence Created What It Enables
Bucket Intelligence Different intervention by delinquency stage
PTP Intelligence Separates promises from realised payment
Refuse-to-Pay Cohort Dedicated recovery treatment
Blank Reclassification Prevents ambiguous responses from polluting outcomes
Wrong-Number Intelligence Faster data correction
Roll-Forward Signals Earlier identification of deeper recovery risk
Collections performance improves when the system learns from what the borrower actually says—not simply whether the borrower answered.

Related Articles