🤖 Hugo — Reactive Intelligence for Procurement

Hugo is an agentic AI co-pilot designed to help procurement and operations teams detect risks, reason over operational data, and take evidence-backed actions with speed and clarity.

It simulates a real enterprise workflow where supplier emails act as operational signals that trigger AI-driven reactive intelligence.

---

🚀 What Hugo Does

• 📩 Simulates supplier emails (delay, quality issue, cancellation, price change)
• 👀 Watches incoming .eml files in real time
• 🤖 Uses Gemini AI to extract structured intelligence
• 🧠 Reasons about risk, impact, and recommendations
• 📊 Renders results in a human-in-the-loop dashboard

---

🏗️ Architecture Overview

This architecture mirrors how real procurement systems ingest and react to supplier communications.

---

📂 Project Structure

Hugo/ ├── frontend/ # React + Vite dashboard ├── backend/ # Python watcher & AI agent ├── assets/ # Architecture diagrams, screenshots ├── README.md

---

⚙️ Setup & Installation

Follow these steps to run the project locally.

---

1️⃣ Clone the Repository

git clone <your-repository-url>
cd Hugo

Create a Python Virtual Environment

Create the environment at the project root:

python -m venv venv

Activate it:

macOS / Linux

source venv/bin/activate


Windows

venv\Scripts\activate

3️⃣ Set Gemini API Key

Set your Gemini API key as an environment variable:

macOS / Linux

export GEMINI_API_KEY=<YOUR_API_KEY>


Windows

set GEMINI_API_KEY=<YOUR_API_KEY>

4️⃣ Install Backend Dependencies

pip install -r requirements.txt

5️⃣ Start the Frontend

cd frontend
npm install
npm run dev

Frontend will start at:

http://localhost:5173

6️⃣ Start the Backend

Open a new terminal, activate the virtual environment again, then:

cd backend
python server.py

Open a new terminal:

cd backend
python main.py

▶️ How to Use the Application

Open http://localhost:5173

in your browser

Navigate to the Send tab

Choose one of the predefined email templates or create a fresh custom email

Click Simulate Email

Wait a few seconds while the AI processes the email

View the results in the Dashboard panel

📊 What the Dashboard Shows

Each alert generated by Hugo includes:

Event Category (e.g. quality issue, delay)

Risk Level (high / moderate / low)

AI Reasoning

Operational Evidence

Impact Analysis

Recommended Actions

Affected Departments

All insights are explainable and traceable to data.

🎥 Demo Video

A walkthrough video demonstrating the full flow and UI is provided below:

📎 Add demo video link here

🛠️ Tech Stack

Frontend: React, Vite, Tailwind CSS

Backend: Python

AI: Google Gemini

Data: CSVs + simulated .eml ingestion

Architecture: Event-driven, human-in-the-loop

🧠 Why Hugo Matters

Hugo demonstrates how agentic AI can:

React to real operational signals

Reason across messy enterprise data

Support decision-making instead of replacing humans

This is Reactive Intelligence, not just analytics.

👥 Team

Maaz2212 : Frontend, UX, Dashboard

taha-rizvi : Backend, AI, Architecture

📌 Notes

mockdata.ts is used as a temporary persistence layer for hackathon speed

In production, this would be replaced by APIs or event streams