Retail Intelligence Multi-Agent System

A sophisticated multi-agent workflow built with LangGraph and Google Gemini to perform automated competitive market research. This system simulates a team of specialized retail analysts to help businesses evaluate new store locations.

Overview

This project uses a directed acyclic graph (DAG) to pass a business "state" through five specialized AI agents. Each agent performs a specific task, enriching the data before passing it to the next.

• Location Analyst: Extracts and validates the geographical target from raw user prompts.
• Competitor Researcher: Identifies top-tier local and national competitors in the area.
• Visitor Analyst: Estimates visitor patterns, comparing weekday vs. weekend dynamics.
• Strategy Consultant: Develops actionable business recommendations based on competitor weaknesses.

pip install -r requirements.txt

• Set up Environment Variables:
  Create a .env file in the root directory and add your keys:

  GOOGLE_API_KEY=your_api_key_here

How to Run

Run the following command:

python marketing.py

Usage

input_state = {
"input_prompt": "I want to open a new clothing retail store in downtown San Francisco"
}

# Execute the graph
result = app.invoke(input_state)
print(result["report"])