📊 LangGraph AI Agent for Business Insights

📌 Overview

This project is a simple AI agent that analyzes basic business data (e.g., daily revenue, costs, and customers) using the LangGraph framework.

The agent can:

• 📈 Calculates daily profit
• 📉 Detects changes in revenue, cost, and CAC (Customer Acquisition Cost)
• 🤖 Provides actionable recommendations to improve business decisions

Built using:

• 🧠 LangGraph for graph-based workflow
• 🔁 Multi-step agent structure
• ✅ UV for fast package management and reproducible environments

---

🎯 Objectives

• Understand how to use LangGraph to build modular AI agents
• Apply basic business metrics and logic (profit, CAC, trends)
• Generate practical, interpretable output
• Write test cases to ensure correctness

---

🧩 Features

• 🔢 Profit Calculation: Computes daily profit using revenue and cost
• 📊 Trend Detection: Compares today’s vs yesterday’s data
• 💰 CAC Alerting: Warns if customer acquisition cost increases
• 💡 Smart Recommendations: Offers suggestions like reducing costs or boosting marketing
• 🧪 Test Coverage: Pytest used to validate outputs

---

📉 Graph

---

🧪 LangGraph Studio Integration

Easily test and visualize your AI agent using LangGraph Studio — a web interface for debugging and observing LangGraph-based workflows.

🌐 How to Use with Studio

1. Add langgraph.json to your repo:

   json
   {
   "dependencies" : [
   "ipykernel>=6.29.5",
   "langchain-groq>=0.3.6",
   "langgraph>=0.5.2",
   "langgraph-cli[inmem]>=0.3.4",
   "pytest>=8.4.1",
   "python-dotenv>=1.1.1"
   ],
   "graphs": {
       "my_agent": "./agent.py:app"},
   "env": "./.venv"
   }
   
   

2. Provide any required environment variables like your API key.

3. Trigger your agent with sample data and view node-by-node execution.

4. Open Run a local server on LangGraph Studio website

---

🎞️ Studio Walkthrough (Demo GIF)

---

💡 Why Use LangGraph Studio?

• 🧭 Visual debugging — See exactly how data flows between nodes
• 🚀 Rapid iteration — Run and test inputs live, no CLI needed
• 🔒 Secure — Use environment variables without hardcoding secrets

---

📂 Project Structure

.
├── src/
│   ├── main.py           # LangGraph agent logic
│   └── test_agent.py     # Pytest test cases
├── .gitignore
├── .env.example          # Safe env config example
├── pyproject.toml        # UV + dependencies
└── README.md

---

🧪 Testing

Run tests with:

uv pip install pytest
pytest src/