📊 Business Intelligence LangGraph Agent

This project implements a simple AI agent using LangGraph that analyzes basic business data (daily revenue, costs, customers) and generates automated insights and recommendations.

🚀 Features

🧮 Calculates:

• Profit/Loss
• Customer Acquisition Cost (CAC)
• Revenue per Customer
• Day-over-day changes (revenue, cost, CAC)

📢 Generates actionable recommendations:

• Reduce costs if profit is negative
• Alert if CAC increases by more than 20%
• Suggest marketing adjustments based on trends
• Identify shrinking customer base or falling revenue per customer
• Highlight low margins or high cost ratios

🧠 Built using LangGraph for agent-based flow and state management

📦 Installation

1. Create and activate virtual environment (recommended)

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

2. Install dependencies

pip install -r requirements.txt

▶️ Usage

Run the agent on sample input:

from agent_graph import graph

sample_input = {
    "data": {
        "today": {"revenue": 2000, "cost": 1200, "customers": 40},
        "yesterday": {"revenue": 1500, "cost": 1100, "customers": 35}
    }
}

result = graph.invoke(sample_input)
print(result["recommendations"])

Output Format:

{
  "recommendations": {
    "status": "Profit",
    "alerts": ["CAC increased by more than 20%."],
    "advice": [
      "Review marketing campaigns due to increased CAC.",
      "Consider increasing advertising budget due to growing sales."
    ]
  }
}

✅ Testing

Run tests to validate various business scenarios:

python test_agent.py

Test cases include:

• Profit and growth
• Loss and rising CAC
• Shrinking customer base
• Low margin alerts
• Falling revenue per customer
• Stable scenario