Product Manager · Fintech · Data Analytics

Gurpreet Singh Padam

7+ Years  ·  LOS · Collections · PD App · Analytics  ·  Delhi, India

Micro LAP PM Collections PM Data-Driven gurpreetsinghpadam.com
View Products Case Studies Download CV ↗
Scroll
7+
Years Experience

About Me

Builder. Analyst. PM.

Product Manager with 7+ years designing data-driven products in fintech and marketplace. Currently leading two product lines at Capri Global Capital — Micro LAP (LOS & PD App) and Collections (Field App & Web Portal).

Supporting 1000+ loan journeys monthly alongside 1,200+ field agents managing ~28,000 collection cases. Previously drove cost reduction, booking automation, and cohort-based CAC optimization at Zomato, Stanza Living, and Cars24.

PRD WritingWorkflow DesignStakeholder MgmtRoadmap Planning SQLFunnel AnalysisCohort AnalysisPower BI · Looker

What I Own

Products I Build

◆ Capri Global · Micro LAP
Loan Origination
System (LOS)
End-to-end loan platform managing 1000+ monthly journeys — from application through document validation, credit assessment, PD, and disbursement.
Multi-stage workfloweSign integrationPD Audio AITAT tracking
◆ Capri Global · Field App
PD Application
(Athema)
Field app for sales & credit officers — structured PD forms, audio recording with AI analysis, GPS verification, photo capture, and direct LOS integration.
Audio recordingAI transcriptGPS verifiedOffline capable
◆ Capri Global · Collections
Collections
Field App
1,200+ field agents · 28,000 cases/month. AI route optimization, digital receipts, trail capture, and real-time DPD tracking.
AI route optimizerDigital receiptBucket tracking
◆ Capri Global · Collections
Collections
Web Portal
Role-based hierarchy (AM→RM→RCM→ZSM→NSM) with MTD/YTD dashboards, bucket movement tracking, and performance visibility at every level.
6-level hierarchyMTD / YTDBucket tracking
◆ Impact at Scale
1000+
Loan journeys/month
1,200+
Field agents
28K
Collection cases/month
30%
TAT reduction

In-Depth Work

Case Studies

⚠ The Problem

PD officers conducting verbal discussions had no accountability for topic coverage at scale — 1000+ PDs per month impossible to manually audit.

  • No verification that income, property, liabilities, loan purpose were discussed
  • Risk of undisclosed money demands by field officers
  • Credit team couldn't efficiently review PD quality
  • Compliance exposure — no audit trail for regulatory review
⚙ The Solution

Integrated audio recording in Athema PD App with AI-powered analysis pipeline.

  • PD session recorded directly in the app during the officer's visit
  • Audio auto-uploads to cloud on case submission
  • AI model: speech-to-text → NLP analysis → structured summary
  • Auto-checklist: income ✓ · property ✓ · liabilities ✓ · loan purpose ✓
  • Flags any segment with money demand language detected
  • Full transcript + summary auto-attached to LOS case
✓ The Result

Compliance and quality control automated at scale — no manual review needed for standard cases.

  • 100% PD topic coverage verification (from ~0% previously)
  • Money demand incidents detectable and actionable
  • Credit team PD review time reduced 80%+ (summary vs full audio)
  • Full regulatory audit trail — transcript on record per case
  • Officer accountability improved — behaviour tracked objectively
◆ How the AI Pipeline Works
1
PD conducted — officer records session using in-app mic during borrower visit
2
Case submitted — audio auto-uploads to cloud, linked to case ID in LOS
3
Speech-to-text — AI transcribes full conversation, timestamped segments
4
NLP analysis — topic model checks coverage · flags demand language · generates structured summary
5
Results to LOS — checklist, transcript, summary auto-attached to case. Credit team sees it instantly.
AI Output — Sample
PD SUMMARY · MIC-2024-3847
✓ Income discussed — Self-employed, ₹42K/month
✓ Property — 1BHK owned, Rohini sector 14
✓ Liabilities — One existing loan (HDFC)
⏳ Loan purpose — briefly mentioned
✓ No money demand detected
100% Coverage ✓ Compliant ✓ 80% Audit Time ↓
⚠ The Problem

Loan documents required physical wet signatures from borrower, co-applicant, and property owner — before disbursement could proceed.

  • Physical presence at branch or courier added 3–5 days to final TAT
  • Multi-party coordination (borrower + co-applicant + property owner) was chaotic
  • Risk of document loss, tampering, or illegibility
  • Compliance team needed audit-proof signature records
  • Paper-based process couldn't scale with loan volume growth
⚙ The Solution

Integrated an Aadhaar OTP-based eSign platform directly into the LOS workflow.

  • After credit approval, eSign request auto-triggered from LOS
  • Borrower receives signing link on mobile — signs via Aadhaar OTP
  • Sequential multi-party workflow: borrower → co-applicant → property owner, each gets their own OTP link
  • Signed document is tamper-proof with hash verification
  • Audit trail: who signed, when, IP, device — auto-logged in LOS
  • RBI/NBFC compliant — legally equivalent to wet signature
✓ The Result

Signing became a mobile-first, async process — zero physical handling required.

  • TAT reduced by 3–5 days per case (signing no longer a bottleneck)
  • 100% compliance with RBI digital signature guidelines
  • Near-zero document loss or tampering risk
  • Multi-party signing coordinated automatically via LOS triggers
  • Disbursement speed increased — more cases closed within SLA
◆ eSign Flow — Before vs After
✗ Before — Manual Process
1
Credit approved — RM calls borrower to visit branch
2
Borrower, co-applicant, property owner each sign physical docs — coordination takes 2–4 days
3
Docs couriered to HQ — 1–2 more days, risk of damage/loss
4
Manual filing, no digital record — audit is manual and slow
✓ After — eSign Flow
1
Credit approved → LOS auto-triggers eSign request instantly
2
Each party receives mobile OTP signing link — signs from anywhere in minutes
3
Sequential completion auto-tracked — LOS proceeds to disburse on all parties signed
4
Tamper-proof signed doc + full audit trail auto-saved to case in LOS
-3 to 5 Days TAT RBI Compliant Zero Paper
⚠ The Problem

Campaigns targeted cold users indiscriminately — ignoring warm existing users who hadn't transacted yet.

  • No differentiation between previous & new users
  • Same channel and frequency for all segments
  • Budget wasted on low-intent users
⚙ The Solution

Built a two-layer data model — IV for feature selection, XGBoost for propensity scoring.

  • Identified previous users who hadn't sold their car yet
  • XGBoost propensity score assigned per user
  • Cohorts bucketed by score band
  • High score → WhatsApp · Mid → RCS · Low → SMS
✓ The Result

Precision targeting dramatically improved campaign efficiency.

  • CAC reduced by 15–20%
  • Targeting accuracy +20–30%
  • Model refreshed monthly — repeatable engine
  • Communication frequency optimised per cohort
◆ Model Pipeline & Channel Strategy
Cohort Propensity Funnel
All prev users
100%
280K
IV-filtered
72%
200K
High propensity
30%
85K
Converted
18%
50K
IV Feature Selection XGBoost Scoring Channel Assignment
Channel × Cohort Matrix
Cohort
Channel
Freq/wk
High
85K
WhatsApp
2×
Mid
56K
RCS
1×
Low
59K
SMS
0.5×
↓ CAC -15–20% vs unmodeled outreach
⚠ The Problem

First Insights team spent majority of time on basic qualification questions — wasting agent capacity on routine queries.

  • Agents repeatedly answering duration, budget, location queries
  • High lag between lead creation and first contact
  • Ground staff received unqualified, incomplete handoffs
⚙ The Solution

Implemented a WhatsApp chatbot flow replacing the initial qualification stage entirely.

  • Lead created → automated WhatsApp message instantly sent
  • User qualifies via chatbot: budget, duration, city, property type
  • Chatbot shortlists & presents matching properties
  • User picks date/time for property visit
  • Visit notification sent directly to property staff on confirmation
✓ The Result

Automation replaced repetitive human work at scale.

  • Sales dependency -60%
  • Conversion +20%
  • First contact time: hours → <5 seconds
  • Ground staff receive fully qualified, scheduled visits
  • Standardised data captured for every lead
◆ New WhatsApp Flow
Before vs After — Agent Touchpoints
Before
Qualify
Schedule
Notify
After
Bot handles all 3 steps
Auto

Career Timeline

The Road So Far

Oct 2017 — Jan 2021
Sr. Data Analyst
Zomato Media Pvt Ltd
  • Reduced CTO from 13% → 6.5% — ~50% cost reduction
  • Built automated data pipelines & visualization systems
  • Root-cause analysis on data quality and systems
~50% cost reduction achieved
Feb 2021 — Aug 2023
Asst. Manager — Product & Analytics
Stanza Living
  • +20% lead-to-booking conversion via journey redesign
  • -60% sales dependency via WhatsApp chatbot automation
  • ETL pipelines & Looker BI infrastructure
60% sales automation unlocked
Aug 2023 — Dec 2024
Manager — Analytics & Growth
Cars24
  • -15–25% CAC via IV + XGBoost cohort targeting
  • Built automated cohort models; +20–30% accuracy
  • Led API integrations (Zoho, CleverTap); -40–50% manual effort
~25% CAC reduction
Dec 2024 — Present ●
Product Manager — Micro LAP & Collections
Capri Global Capital Ltd · Gurgaon
  • Micro LAP: LOS & PD App — 1000+ loan journeys/month
  • eSign integration — signing TAT reduced 3–5 days/case
  • PD Audio + AI analysis — 100% coverage verification
  • -20–30% overall loan processing TAT
  • Collections: Field App & Web Portal — 1,200+ agents
  • AI multi-stop route optimization — +20% agent visit capacity
  • ~28,000 collection cases managed monthly
  • Role-based portal (AM→RM→RCM→ZSM→NSM) with MTD/YTD dashboards
2 products · 1,200+ agents · 28,000 cases/month

Data Work

Analytics
& Impact

Every product decision I make is backed by data. Loan funnel analysis, cohort targeting, TAT reduction — the visualisations below reflect the data work I drive across the product lifecycle.

◆ Loan Origination Funnel — Sankey Flow
Volume drop-off at each loan stage. Width proportional to case volume passing through.
PM-ownedCreditDisbursed
◆ TAT Reduction — Monthly (Days)
Loan processing turnaround over 8 months. eSign + automation drove sustained reduction.
◆ Conversion Funnel — Stage by Stage
Volume as % of total leads at each loan stage.
Leads
100%
1000
Qualified
75%
750
Applied
50%
500
PD Done
38%
380
Credit OK
26%
260
Disbursed
18.5%
185

Capabilities

Skills & Tools

◆ Product Management
PRD Writing & Documentation95%
Workflow & Process Design93%
Stakeholder Management90%
Feature Prioritization (RICE / MoSCoW)88%
User Journey Mapping92%
Agile / Scrum85%
◆ Data & Analytics
SQL90%
Funnel & Cohort Analysis88%
Power BI / Looker / Tableau85%
A/B Testing & Experimentation82%
KPI Definition & Root Cause Analysis90%
Python72%
SnowflakeRedshiftREST APIsPostmanCleverTapZoho CRMGoogle SheetsJiraFigmaMixpanelPythonLooker

Measured Impact

Numbers That Matter

0+
Years Experience
0%
Cost Reduction — Zomato
0%
Sales Automation — Stanza
0+
Field Agents — Collections
0K
Cases/Month — Collections
◆ Key Impact Metrics by Initiative
50%
Zomato
Cost reduction
60%
Stanza
Sales automation
20%
Stanza
Conversion ↑
25%
Cars24
CAC reduction
30%
Capri LAP
TAT reduction
20%
Collections
Agent capacity ↑

Thoughts & Writing

Blog & Content

Coming Soon
Product Management
How I Write PRDs That Actually Get Built
A structured breakdown — from discovery questions to functional specs, validation logic, and backward compatibility. Real examples from fintech LOS workflows.
Gurpreet Singh Padam · PM
Coming Soon
Data & Analytics
Reducing TAT in Loan Processing: A Data PM's Playbook
How eSign integration, PD audio automation, and workflow mapping cut loan TAT by multiple days per case at Capri Global Capital.
Gurpreet Singh Padam · Fintech
Coming Soon
Content & Growth
Product Thinking Applied to Instagram Growth
How I apply retention loops, hook frameworks, and A/B-style experimentation to grow a faceless content account without showing my face.
Gurpreet Singh Padam · Content

Let's Connect

Open to Opportunities

Senior PM / Business Analyst roles in fintech & marketplace. Open to product-led, data-driven organisations.

gurpreet.zeom@gmail.com
LinkedIn ↗ GitHub ↗ +91 92509 46303