🚀 Agentic AI Automation Platform

📌 Overview

This project is a multi-agent AI system that autonomously plans, executes, and verifies tasks using Large Language Models (LLMs).

It simulates a real-world AI Business Analyst, capable of breaking down complex queries and generating structured outputs.

---

🧠 Architecture

The system is built using a 3-layer agentic architecture:

• Planner Agent

  • Breaks user query into actionable steps

• Executor Agent

  • Executes each step using tools

• Verifier Agent

  • Validates outputs and retries if incorrect

---

🔥 Key Highlight

This project implements a verification loop to reduce hallucinations in LLM outputs.

---

⚙️ Tech Stack

• LangChain
• OpenAI
• FastAPI
• Streamlit
• Python

---

🚀 Features

• Multi-agent workflow
• Tool-based execution
• Verification mechanism (retry logic)
• API + UI integration
• Modular code structure

---

📂 Project Structure

agentic_ai_platform/ │ ├── agents/ │ ├── planner.py │ ├── executor.py │ ├── verifier.py │ ├── tools/ │ └── search_tool.py │ ├── utils/ │ └── prompts.py │ ├── app/ │ └── main.py │ ├── frontend/ │ └── streamlit_app.py │ ├── requirements.txt └── README.md

---

▶️ How to Run

git clone https://github.com/yourusername/agentic_ai_platform.git
cd agentic_ai_platform

python -m venv venv
venv\Scripts\activate

pip install -r requirements.txt

uvicorn app.main:app --reload
streamlit run frontend/streamlit_app.py

https://github.com/anshgagai-dotcom/agentic_ai_platform.git


<img width="1365" height="762" alt="screenshot" src="https://github.com/user-attachments/assets/390f0ff2-618e-4b11-843a-e18f81ed1f63" />