FinAgent — Autonomous AI Personal Finance Assistant

What is FinAgent?

FinAgent is an autonomous AI-powered personal finance assistant that helps users understand and manage their everyday finances without requiring them to manually perform financial analysis.

Users can upload their transaction data as a CSV file and interact with FinAgent using natural-language questions. The agent analyzes the financial data, selects the appropriate analysis tools, identifies important patterns, and generates clear and actionable financial insights.

FinAgent is built using the Strands Agents SDK and Gemini, with a Streamlit interface for an interactive user experience.

The Problem

Managing personal finances can be time-consuming. Users often need to manually:

• Calculate income and expenses
• Track savings
• Categorize spending
• Identify unusually large transactions
• Understand their overall financial health
• Decide where they can reduce unnecessary spending

Traditional dashboards mainly display information. They do not always understand what the user is asking or decide what analysis should be performed.

FinAgent addresses this problem by combining an interactive financial dashboard with an autonomous AI agent.

Our Solution

FinAgent converts raw transaction data into understandable financial insights.

After uploading a transaction CSV, users can immediately see their income, expenses, savings, savings rate, spending categories, unusual transactions, financial health score, and recommendations.

Users can also ask questions such as:

• "How much am I spending on food?"
• "Do I have any unusual expenses?"
• "How healthy are my finances?"
• "Where can I reduce my spending?"

Instead of running every analysis for every question, FinAgent determines which specialized finance tool is relevant and uses that tool to obtain the required information.

Why is it Agentic?

The key feature of FinAgent is autonomous tool selection.

The agent has access to specialized financial analysis tools:

1. Transaction Analysis — calculates income, expenses, savings, and transaction count.
2. Spending Category Analysis — analyzes spending across categories.
3. Unusual Transaction Detection — identifies unusually large expenses.
4. Financial Health Assessment — calculates financial health and savings rate.
5. Recommendation Generation — produces personalized recommendations based on the user's financial situation.

When a user asks a question, FinAgent decides which tool or combination of tools is necessary to answer it.

For example:

User: "How much am I spending on food?"

FinAgent: Selects the spending category analysis tool → retrieves the actual transaction data → returns the food spending amount.

For a broader question such as "Analyze my finances",

Built With

• agents
• gemini
• pandas
• python
• sdk
• strands
• streamlit