Made byBobr AI

Finance Saathi: Agentic AI for Financial Literacy in India

Explore Finance Saathi, an AI-powered financial companion delivering multilingual guidance, fraud protection, and scheme navigation for the Indian market.

#fintech#artificial-intelligence#india#financial-literacy#agentic-ai#banking#startup#inclusive-finance
Watch
Pitch
Nomura KakushIN 2026
Finance Saathi
— Agentic AI for Financial Literacy
"One AI Companion. Every Indian's Financial Language."
Team: Tech Titans  |  Rahul Sharma, Priya Patel
User Personas
Nomura KakushIN 2026
Made byBobr AI
02
The Problem
₹20.6T
UPI transactions in FY24
Record-breaking digital payment adoption
80M+
New retail investors since 2020
India's fastest-growing investing class
"
Financial ACCESS has outpaced Financial UNDERSTANDING
Millions transact digitally but cannot read a credit score, compare insurance, or spot a predatory loan.
Can't decode credit scores
Misled by ULIP mis-selling
Fall into predatory loan traps
Finance Saathi
2 / 15
Made byBobr AI
03
Who We're Solving For
Four real Indians who need Finance Saathi
👩‍🏫
Priya
32 • Thane School Teacher • Marathi
₹35,000/month (fixed)
Pain Point
Overwhelmed by jargon & ULIP mis-selling
Core Need
"Understand my insurance before I buy it"
🛵
Rajesh
24 • Mumbai Gig Worker • Hindi
₹15K–40K/month (variable)
Pain Point
No emergency fund, falls for loan apps
Core Need
"Save money despite irregular income"
🧑‍🌾
Kisan
45 • Karnataka Farmer • Kannada
3 acres, seasonal income
Pain Point
Confused by PM-KISAN & crop insurance
Core Need
"Know which schemes I'm eligible for"
👩‍🦯
Divya
28 • Pune Content Writer • Visual Impairment
₹25,000/month
Pain Point
No voice-first, screen-reader tools
Core Need
"Full financial access without barriers"
Finance Saathi
3 / 15
Made byBobr AI
04
Why Existing Solutions Fail
Generic Financial Apps
Assume digital literacy + connectivity
English-only or Hindi-only
No context for irregular income or disability
Bank Advisors
Gatekept by ticket size
"Come back with ₹5 lakhs"
Mass-market users are invisible
YouTube / Online Content
Generic, not income-calibrated
Not in user's native language
Can't answer follow-up questions
NGOs / Educators
Cannot scale 1:1 counseling
Limited geographic reach
No real-time scheme updates
No existing solution is persona-aware, multilingual, AND accessible at scale.
Finance Saathi
4 / 15
Made byBobr AI
05
Our Solution
Finance Saathi
= A Persona-Aware, Multilingual, Multi-Agent AI System
Detects WHO the user is Routes to the RIGHT SPECIALIST AGENT Delivers PERSONALIZED financial guidance
🎤 Voice
💬 Chat
📱 WhatsApp
📟 SMS/USSD
♿ Screen Reader
📚 Financial Education
🏛️ Scheme Navigation
🛡️ Fraud Protection
💰 Budget Coaching
Finance Saathi
"One AI Companion. Every Indian's Financial Language."
5 / 15
Made byBobr AI
06
System Architecture — High-Level Design
10-Layer Agentic AI Flow
① Input Layer
🎤 Voice / Whisper ASR
💬 Text/Chat
📱 WhatsApp Business API
📟 SMS/USSD
② Persona Intelligence Engine
🌐 Language Detector (Bhashini/IndicTrans)
👤 Profile Inference
🧠 Context Memory
♿ Accessibility Layer
③ Orchestrator Agent
LangChain + Groq Llama3-70B
🔀 Intent Classifier
🛣️ Agent Router
💾 Session State Manager
④ Specialist Agents
📚 Education Agent
🏛️ Scheme Navigator
🛡️ Fraud Guard
💰 Budget Coach
⑤ RAG & Knowledge Layer
ChromaDB Vector Store
📄 RBI/SEBI Docs
🏛️ Govt Scheme DB
⚠️ Fraud Patterns DB
⑥ Response Synthesis
🗣️ Language Adapter
📐 Format Adapter
✅ Confidence Filter
⑦ Output Layer
📊 Personalized Guidance
📋 Scheme Report
🚨 Fraud Alerts
📈 Budget Plan
⭐ Impact Score
⑧ Behavioral Nudge System
Goal Tracker
Proactive Nudges
Gamification
⑨ Human Oversight & Responsible AI
Escalation Trigger
AI Disclaimer
Bias Monitor
Finance Saathi
6 / 15
Made byBobr AI
07
Specialist Agents
Four AI experts, each purpose-built
📚 Education Agent
Financial Literacy
  • Explains financial concepts in plain language
  • Calibrated to user's income level & literacy
  • Covers: SIPs, credit scores, insurance types, EMIs
  • Output: Jargon-free explanations with analogies
Personas served: Priya, Divya
🏛️ Scheme Navigator
Govt Schemes
  • PM-KISAN, EPF, PMSBY, PMJDY eligibility checks
  • State + Central scheme filtering
  • Step-by-step application guidance
  • Generates eligibility reports with next steps
Personas served: Kisan, Rajesh
🛡️ Fraud Guard Agent
Fraud Protection
  • Detects predatory lending patterns
  • Real-time scam signature matching
  • India-specific: gig/rural loan app fraud
  • Sends instant alerts before user commits
Personas served: Rajesh, Kisan
💰 Budget Coach
Budgeting
  • Built for irregular/variable income
  • Seasonal cash flow planning
  • Weekly micro-savings goal builder
  • Emergency fund roadmap in ₹ steps
Personas served: Rajesh, Kisan, Priya
Finance Saathi
7 / 15
Made byBobr AI
08
Sample Output — UI Mockup
What users actually see
Budget Coach — Rajesh (Mumbai, Hindi)
FS
Finance Saathi
● Online
I earn different amounts every month, paisa kaise bachau?
Rajesh • Variable income • Hindi
नमस्ते Rajesh! 🙏 आपकी income variable है — इसलिए मैं आपको एक flexible micro-savings plan देता हूँ:
• Week 1-2: ₹500 emergency jar
• Week 3: ₹300 SIP (when income > ₹25K)
• Week 4: ₹200 buffer
Total: ₹1,000/month minimum savings ✅
🌟
Financial Health: 34/100 → Target: 55 in 90 days
⚠️ This is financial education, not advice. Consult a SEBI-registered advisor for investments.
Scheme Navigator — Kisan (Karnataka, Kannada)
PM-KISAN ಅರ್ಹತೆ ವರದಿ
Finance Saathi
✅ ನೀವು ಅರ್ಹರು (You are eligible)
1
ಆಧಾರ್ ಕಾರ್ಡ್ ಮತ್ತು ಭೂ ದಾಖಲೆ ತಯಾರಿ ಮಾಡಿ
(Prepare Aadhaar + land records)
2
pmkisan.gov.in ಗೆ ಹೋಗಿ
(Visit pmkisan.gov.in)
3
Farmer Corner → New Registration ಕ್ಲಿಕ್ ಮಾಡಿ
(Click New Registration)
₹6,000/year benefit • 3 installments
Generated by Finance Saathi • Verified against official portal data
Finance Saathi
8 / 15
Made byBobr AI
Nomura KakushIN 2026
09
Data Model — Simplified
Core entities powering Finance Saathi
FINANCE SAATHI
9 / 15
initiates (1:many) generates (1:many) queries (many:many) tracked in (1:1) updates (many:1)
👤 User Profile
persona_type (teacher/gig/farmer/pwd)
preferred_language
income_pattern (fixed/var/seasonal)
literacy_level (1–5)
accessibility_needs
location_state
💬 Conversation Session
session_id
user_id (→ User Profile)
intent_detected
agent_routed_to
conversation_history[]
timestamp
📚 Knowledge Base
record_type (scheme/circular/fraud)
content_text (embedded)
category
state_applicability
eligibility_criteria
last_updated
📋 Interaction Log
log_id
session_id (→ Session)
query_text
response_text
confidence_score (0–1)
escalation_flag (bool)
📈 Impact Tracker
tracker_id
user_id (→ User Profile)
financial_health_index (0–100)
goals_set / goals_achieved
nudges_sent / nudges_acted_on
snapshot_date
Made byBobr AI
10
What's Genuinely New Here
Original thinking — not derivative of existing products
Novel Architecture
Persona-Inference-First Routing
NOT
A generic chatbot with one fixed persona
YES
Detects income pattern, language, literacy, accessibility → routes to the RIGHT agent automatically
Domain-Specific Intelligence
Fraud Guard Trained on Indian Predatory Patterns
NOT
Generic scam detection
YES
Trained on gig-worker loan app fraud + rural chit fund schemes specific to Indian fintech context
Outcome-Oriented Metric
Financial Health Impact Score™
NOT
One-off Q&A chatbot
YES
Longitudinal metric tracking user's financial wellbeing over time — from first interaction to goal achievement
Responsible AI by Design
Bias Monitor Across Personas
NOT
Assume AI treats all users equally
YES
Active fairness check — low-income/low-literacy users get SAME quality answers as high-income users. Monitored, not assumed.
None of these four features exist in combination in any current Indian financial app.
Finance Saathi
10 / 15
Made byBobr AI
11
Responsible AI & Trust
Built-in safeguards — not bolted on
AI Always Disclosed
Users know they're talking to an AI — upfront, every session. No deceptive personas.
Human Escalation Triggers
Low confidence score OR high-stakes query (large investments) → automatically escalated to human SEBI-registered advisor.
Data Anonymized & Local
No PII shared with third parties. Stored regionally per Indian data protection norms. User controls their data.
Bias Monitoring
Answer quality compared across all persona types and languages. Low-literacy users must receive equivalent quality guidance.
No Disguised Advice
Financial education ≠ financial advice. Every response includes a clear SEBI disclaimer when applicable.
"Trust is not a feature — it is the foundation."
Finance Saathi
11 / 15
Made byBobr AI
12
Key Assumptions
Clearly stated — not hidden
1 Device Access
Users have basic smartphone OR feature phone with SMS/USSD capability — not full data connectivity required.
2 Bhashini/IndicTrans Availability
Government's Bhashini API and IndicTrans translation models are accessible for 12 Indian languages under open-access terms.
3 Scheme Data Accessibility
Government scheme data sourceable via public APIs or periodic scraping of official portals (pmkisan.gov.in, epfindia.gov.in, etc.)
4 Groq Inference Cost-Viability
Groq's Llama3-70B inference is cost-viable at target scale — ₹X per 1,000 queries stays within viable unit economics.
5 WhatsApp Business API Approval
WhatsApp Business API approval is obtainable for a fintech/social-impact use case within a reasonable timeline.
6 User Consent to Personalization
Users provide informed consent to data storage for personalization purposes, with option to delete at any time.
These assumptions are validated against publicly available information as of June 2026.
Finance Saathi
12 / 15
Made byBobr AI
13
Implementation Plan
8-week phased delivery
WEEKS 1–2
P1
Core Foundation
Orchestrator agent setup
Education Agent + Scheme Navigator
Text/chat input only
English + 1 regional language
Basic RAG pipeline (ChromaDB)
MVP Ready
WEEKS 3–4
P2
Expand Agents & Languages
Fraud Guard Agent + Budget Coach
12 languages via Bhashini
WhatsApp Business integration
Full RAG knowledge base
Multi-language
WEEKS 5–6
P3
Voice & Accessibility
Voice input via Whisper ASR
Screen reader accessibility layer
SMS/USSD fallback
Low-connectivity optimization
Inclusive Access
WEEKS 7–8
P4
Trust & Impact Layer
Behavioral nudge system
Financial Health Impact Score
Human escalation workflow
Bias/fairness testing across personas
Production Ready
Tech Stack Summary
LangChain
Groq Llama3-70B
ChromaDB
Bhashini
Whisper ASR
Streamlit
WhatsApp API
Finance Saathi
13 / 15
Made byBobr AI
14 Roles & Responsibilities
Two members, full-stack ownership
AI / Backend Lead
Member 1
🔧
Orchestrator agent design & LangChain routing logic
🧠
RAG pipeline — ChromaDB, embeddings, vector search
✍️
Prompt engineering for all 4 specialist agents
🛡️
Fraud-pattern model training & tuning
🔌
API integrations: Groq, Bhashini, WhatsApp Business
🏗️
System architecture implementation
LangChain
Groq
ChromaDB
Python
Bhashini API
Product / Frontend Lead
Member 2
🎨
Persona research & UX flow design
📱
Streamlit / WhatsApp / voice interface
Accessibility implementation (screen reader, USSD)
📊
Data model design & impact-scoring logic
⚖️
Responsible AI: disclaimers, escalation UX
🧪
Testing across all four personas
Streamlit
Whisper
UX Design
Impact Score
Fairness Testing
Combined: Full-stack AI system from data model to voice interface — delivered in 8 weeks.
Finance Saathi
14 / 15
Made byBobr AI
15
Impact & Closing
Four Personas. Four Transformations.
👩‍🏫
Priya
"Gets jargon-free retirement guidance in Marathi — finally understands her ULIP"
📈 Financial literacy: Low → Medium
🛵
Rajesh
"Gets a workable ₹1,000/month micro-savings plan despite variable income"
💰 Emergency fund: ₹0 → ₹12,000 in 12 months
👨‍🌾
Kisan
"Gets a 3-step PM-KISAN application path in Kannada — no middlemen"
🌾 Scheme access: None → ₹6,000/year
Divya
"Gets a fully voice-first, screen-reader-compatible financial experience"
♿ Barrier-free: First time accessing financial tools
Make your mark.
Empower millions.
Build the future of financial literacy.
Finance Saathi — Nomura KakushIN 2026
Made byBobr AI
Bobr AI

DESIGNER-MADE
PRESENTATION,
GENERATED FROM
YOUR PROMPT

Create your own professional slide deck with real images, data charts, and unique design in under a minute.

Generate For Free

Finance Saathi: Agentic AI for Financial Literacy in India

Explore Finance Saathi, an AI-powered financial companion delivering multilingual guidance, fraud protection, and scheme navigation for the Indian market.

Finance Saathi

— Agentic AI for Financial Literacy

One AI Companion. Every Indian's Financial Language.

<span style="font-weight: 600; color: #FFFFFF;">Team:</span> Tech Titans &nbsp;|&nbsp; Rahul Sharma, Priya Patel

Nomura KakushIN 2026

02

The Problem

₹20.6T

UPI transactions in FY24

Record-breaking digital payment adoption

80M+

New retail investors since 2020

India's fastest-growing investing class

Financial ACCESS has outpaced Financial <span style="color: #D4AF37;">UNDERSTANDING</span>

Millions transact digitally but cannot read a credit score, compare insurance, or spot a predatory loan.

Can't decode credit scores

Misled by ULIP mis-selling

Fall into predatory loan traps

Finance Saathi

2 / 15

03

Who We're Solving For

Four real Indians who need Finance Saathi

Priya

32 &bull; Thane School Teacher &bull; Marathi

&#8377;35,000/month (fixed)

Overwhelmed by jargon & ULIP mis-selling

Understand my insurance before I buy it

Rajesh

24 &bull; Mumbai Gig Worker &bull; Hindi

&#8377;15K&ndash;40K/month (variable)

No emergency fund, falls for loan apps

Save money despite irregular income

Kisan

45 &bull; Karnataka Farmer &bull; Kannada

3 acres, seasonal income

Confused by PM-KISAN & crop insurance

Know which schemes I'm eligible for

Divya

28 &bull; Pune Content Writer &bull; Visual Impairment

&#8377;25,000/month

No voice-first, screen-reader tools

Full financial access without barriers

Finance Saathi

3 / 15

Why Existing Solutions Fail

Generic Financial Apps

Bank Advisors

YouTube / Online Content

NGOs / Educators

No existing solution is persona-aware, multilingual, AND accessible at scale.

Finance Saathi

4 / 15

05

Our Solution

Finance Saathi

= A Persona-Aware, Multilingual, Multi-Agent AI System

Detects

WHO the user is

Routes to the

RIGHT SPECIALIST AGENT

Delivers

PERSONALIZED financial guidance

🎤 Voice

💬 Chat

📱 WhatsApp

📟 SMS/USSD

♿ Screen Reader

📚 Financial Education

🏛️ Scheme Navigation

🛡️ Fraud Protection

💰 Budget Coaching

One AI Companion. Every Indian's Financial Language.

Finance Saathi

5 / 15

06

System Architecture — High-Level Design

10-Layer Agentic AI Flow

① Input Layer

② Persona Intelligence Engine

③ Orchestrator Agent

④ Specialist Agents

⑤ RAG & Knowledge Layer

⑥ Response Synthesis

⑦ Output Layer

⑧ Behavioral Nudge System

⑨ Human Oversight & Responsible AI

Finance Saathi

6 / 15

Specialist Agents

Four AI experts, each purpose-built

Finance Saathi

7 / 15

08

Sample Output — UI Mockup

What users actually see

Budget Coach — Rajesh (Mumbai, Hindi)

I earn different amounts every month, paisa kaise bachau?

Rajesh • Variable income • Hindi

नमस्ते Rajesh! 🙏 आपकी income variable है — इसलिए मैं आपको एक flexible micro-savings plan देता हूँ:

• Week 1-2: ₹500 emergency jar

• Week 3: ₹300 SIP (when income > ₹25K)

• Week 4: ₹200 buffer

Total: ₹1,000/month minimum savings ✅

Financial Health: 34/100 → Target: 55 in 90 days

⚠️ This is financial education, not advice. Consult a SEBI-registered advisor for investments.

Scheme Navigator — Kisan (Karnataka, Kannada)

PM-KISAN ಅರ್ಹತೆ ವರದಿ

✅ ನೀವು ಅರ್ಹರು (You are eligible)

ಆಧಾರ್ ಕಾರ್ಡ್ ಮತ್ತು ಭೂ ದಾಖಲೆ ತಯಾರಿ ಮಾಡಿ<br><span style="font-size: 15px; color: #55565A; font-weight: 500; margin-top: 4px; display: block;">(Prepare Aadhaar + land records)</span>

pmkisan.gov.in ಗೆ ಹೋಗಿ<br><span style="font-size: 15px; color: #55565A; font-weight: 500; margin-top: 4px; display: block;">(Visit pmkisan.gov.in)</span>

Farmer Corner → New Registration ಕ್ಲಿಕ್ ಮಾಡಿ<br><span style="font-size: 15px; color: #55565A; font-weight: 500; margin-top: 4px; display: block;">(Click New Registration)</span>

₹6,000/year benefit • 3 installments

Generated by Finance Saathi • Verified against official portal data

Finance Saathi

8 / 15

Data Model — Simplified

Core entities powering Finance Saathi

Nomura KakushIN 2026

What's Genuinely New Here

Original thinking — not derivative of existing products

Persona-Inference-First Routing

A generic chatbot with one fixed persona

Detects income pattern, language, literacy, accessibility → routes to the RIGHT agent automatically

Novel Architecture

Fraud Guard Trained on Indian Predatory Patterns

Generic scam detection

Trained on gig-worker loan app fraud + rural chit fund schemes specific to Indian fintech context

Domain-Specific Intelligence

Financial Health Impact Score™

One-off Q&A chatbot

Longitudinal metric tracking user's financial wellbeing over time — from first interaction to goal achievement

Outcome-Oriented Metric

Bias Monitor Across Personas

Assume AI treats all users equally

Active fairness check — low-income/low-literacy users get SAME quality answers as high-income users. Monitored, not assumed.

Responsible AI by Design

None of these four features exist in combination in any current Indian financial app.

Finance Saathi

10 / 15

11

Responsible AI & Trust

Built-in safeguards — not bolted on

AI Always Disclosed

Users know they're talking to an AI — upfront, every session. No deceptive personas.

Human Escalation Triggers

Low confidence score OR high-stakes query (large investments) → automatically escalated to human SEBI-registered advisor.

Data Anonymized & Local

No PII shared with third parties. Stored regionally per Indian data protection norms. User controls their data.

Bias Monitoring

Answer quality compared across all persona types and languages. Low-literacy users must receive equivalent quality guidance.

No Disguised Advice

Financial education ≠ financial advice. Every response includes a clear SEBI disclaimer when applicable.

"Trust is not a feature — it is the foundation."

Finance Saathi

11 / 15

Device Access

Users have basic smartphone OR feature phone with SMS/USSD capability — not full data connectivity required.

Bhashini/IndicTrans Availability

Government's Bhashini API and IndicTrans translation models are accessible for 12 Indian languages under open-access terms.

Scheme Data Accessibility

Government scheme data sourceable via public APIs or periodic scraping of official portals (pmkisan.gov.in, epfindia.gov.in, etc.)

Groq Inference Cost-Viability

Groq's Llama3-70B inference is cost-viable at target scale — ₹X per 1,000 queries stays within viable unit economics.

WhatsApp Business API Approval

WhatsApp Business API approval is obtainable for a fintech/social-impact use case within a reasonable timeline.

User Consent to Personalization

Users provide informed consent to data storage for personalization purposes, with option to delete at any time.

These assumptions are validated against publicly available information as of June 2026.

Finance Saathi

12 / 15

13

Implementation Plan

8-week phased delivery

WEEKS 1–2

Core Foundation

<div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #2EC4B6; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Orchestrator agent setup</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #2EC4B6; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Education Agent + Scheme Navigator</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #2EC4B6; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Text/chat input only</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #2EC4B6; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">English + 1 regional language</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #2EC4B6; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Basic RAG pipeline (ChromaDB)</div></div>

MVP Ready

WEEKS 3–4

Expand Agents & Languages

<div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #D4AF37; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Fraud Guard Agent + Budget Coach</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #D4AF37; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">12 languages via Bhashini</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #D4AF37; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">WhatsApp Business integration</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #D4AF37; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Full RAG knowledge base</div></div>

Multi-language

WEEKS 5–6

Voice & Accessibility

<div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #BA8EE4; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Voice input via Whisper ASR</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #BA8EE4; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Screen reader accessibility layer</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #BA8EE4; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">SMS/USSD fallback</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #BA8EE4; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Low-connectivity optimization</div></div>

Inclusive Access

WEEKS 7–8

Trust & Impact Layer

<div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #E57373; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Behavioral nudge system</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #E57373; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Financial Health Impact Score</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #E57373; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Human escalation workflow</div></div> <div style="display: flex; align-items: flex-start; gap: 12px;"><div style="color: #E57373; font-size: 18px; font-weight: 800;">✓</div><div style="opacity: 0.9;">Bias/fairness testing across personas</div></div>

Production Ready

<div style="background: rgba(255,255,255,0.06); border: 1px solid rgba(255,255,255,0.15); padding: 6px 20px; border-radius: 20px; color: #E0E0E0; font-size: 14px; font-weight: 500; box-shadow: 0 3px 8px rgba(0,0,0,0.2);">LangChain</div> <div style="background: rgba(255,255,255,0.06); border: 1px solid rgba(255,255,255,0.15); padding: 6px 20px; border-radius: 20px; color: #E0E0E0; font-size: 14px; font-weight: 500; box-shadow: 0 3px 8px rgba(0,0,0,0.2);">Groq Llama3-70B</div> <div style="background: rgba(255,255,255,0.06); border: 1px solid rgba(255,255,255,0.15); padding: 6px 20px; border-radius: 20px; color: #E0E0E0; font-size: 14px; font-weight: 500; box-shadow: 0 3px 8px rgba(0,0,0,0.2);">ChromaDB</div> <div style="background: rgba(255,255,255,0.06); border: 1px solid rgba(255,255,255,0.15); padding: 6px 20px; border-radius: 20px; color: #E0E0E0; font-size: 14px; font-weight: 500; box-shadow: 0 3px 8px rgba(0,0,0,0.2);">Bhashini</div> <div style="background: rgba(255,255,255,0.06); border: 1px solid rgba(255,255,255,0.15); padding: 6px 20px; border-radius: 20px; color: #E0E0E0; font-size: 14px; font-weight: 500; box-shadow: 0 3px 8px rgba(0,0,0,0.2);">Whisper ASR</div> <div style="background: rgba(255,255,255,0.06); border: 1px solid rgba(255,255,255,0.15); padding: 6px 20px; border-radius: 20px; color: #E0E0E0; font-size: 14px; font-weight: 500; box-shadow: 0 3px 8px rgba(0,0,0,0.2);">Streamlit</div> <div style="background: rgba(255,255,255,0.06); border: 1px solid rgba(255,255,255,0.15); padding: 6px 20px; border-radius: 20px; color: #E0E0E0; font-size: 14px; font-weight: 500; box-shadow: 0 3px 8px rgba(0,0,0,0.2);">WhatsApp API</div>

Finance Saathi

13 / 15

Roles & Responsibilities

Two members, full-stack ownership

AI / Backend Lead

Member 1

Product / Frontend Lead

Member 2

Combined: Full-stack AI system from data model to voice interface — delivered in 8 weeks.

Finance Saathi

14 / 15

15

Impact & Closing

Four Personas. Four Transformations.

Priya

Gets jargon-free retirement guidance in Marathi — finally understands her ULIP

📈 Financial literacy: Low → Medium

Rajesh

Gets a workable ₹1,000/month micro-savings plan despite variable income

💰 Emergency fund: ₹0 → ₹12,000 in 12 months

Kisan

Gets a 3-step PM-KISAN application path in Kannada — no middlemen

🌾 Scheme access: None → ₹6,000/year

Divya

Gets a fully voice-first, screen-reader-compatible financial experience

♿ Barrier-free: First time accessing financial tools

Make your mark.

Empower millions.

Build the future of financial literacy.

Finance Saathi — Nomura KakushIN 2026