AI Expense Buddy

Overview

An AI-powered personal finance app that automatically reads and parses transaction SMS messages from UPI apps — GPay, PhonePe, Paytm, BHIM, Amazon Pay, and bank notifications — and turns them into structured, categorized expense insights without any manual entry.

The Problem It Solves

India processes over 12 billion UPI transactions monthly. Every transaction generates an SMS or app notification — but that data sits siloed in your inbox, unstructured and unactionable. Users who want to track spending either use manual tools (painful) or bank statement exports (delayed, cumbersome). There’s no zero-friction, real-time layer that connects the dots.

Core Features

1. SMS Ingestion & Parsing

The app requests SMS read permission and continuously monitors for new UPI/bank messages. A local NLP parser (or on-device LLM) extracts:

• Transaction amount
• Merchant/recipient name
• UPI ID or VPA (Virtual Payment Address)
• Date and time
• Transaction type (debit/credit/refund)
• Reference ID

It handles the wide variation in message formats across HDFC, SBI, ICICI, Axis, Kotak, and dozens of other banks — each with their own SMS templates.

2. AI Categorization Engine

Each parsed transaction is classified into spend categories:

• Groceries (BigBasket, Zepto, DMart)
• Food & Dining (Swiggy, Zomato, restaurants)
• Transport (Ola, Uber, BMTC, fuel)
• Shopping (Myntra, Flipkart, Amazon)
• Healthcare (PharmEasy, Apollo)
• Utilities & Recharges (Jio, Airtel, electricity)
• Rent & Housing
• Entertainment (BookMyShow, OTT subscriptions)
• Transfers (P2P UPI payments)

The AI learns from user corrections over time, improving accuracy for local merchants and regional vendor names.

3. Smart Dashboard

• Monthly spend summary with category breakdown
• Daily/weekly trend charts
• Merchant-level drill-downs (e.g., “You’ve spent ₹4,200 on Swiggy this month”)
• Budget tracking with overspend alerts
• Peer comparison benchmarks (anonymized, opt-in)

4. Conversational Query Layer

Users can ask natural language questions like:

• “How much did I spend on food last month?”
• “Show me all Paytm payments above ₹500”
• “What’s my biggest expense this week?”

The AI responds with data-backed summaries, no manual filtering needed.

5. Recurring Subscription Detector

Automatically flags recurring charges — OTT platforms, gym memberships, SaaS tools — and tracks them separately with renewal reminders.