Inspiration

Big companies depend on many suppliers. When one fails, the damage spreads to stock, deliveries, cost and sales. Managers often decide using spreadsheets and guesswork, and testing a plan on the real business is risky. We asked: what if a company had a "time machine" to test a disruption before it happens?

What it does

Digital Twin is a GenAI-for-enterprise tool that creates a digital twin of a business, a pretend company where anyone can break something and see what happens, with no real loss.

Example: we built a fictional phone shop. It sells 100 phones a day and keeps 600 in stock. It has a main supplier and a backup supplier who charges a little more. The manager asks: "What if the main supplier shuts down for 14 days?"

Scenario controls: the manager picks the supplier, the 14 days, and any change in demand.
Simulation engine: Python counts day by day. Stock drops from 600 to 0 in about 6 days, and after that customers cannot buy phones.
Recovery comparison: the dashboard shows three versions side by side: normal shop, shop with the problem, and shop with the backup supplier.
AI advisor: explains the risk in simple words, for example "You will run out of stock in 6 days. First action: contact the backup supplier today."

So the manager sees the risk and the best first step without risking the real business.

How we built it

Python and Pandas for the day-by-day simulation (stock, orders, cost, revenue)
Streamlit for the interactive dashboard with sliders
Plotly for live comparison charts
AI advisor with a built-in explanation now, and LLM API integration ready (it switches to AI-written advice when an API key is added)

Challenges we ran into

Turning a real business problem into simple simulation rules
Keeping the simulator simple but still realistic (backup supplier delay, extra cost)
Setting up Python, VS Code and Streamlit, and fixing folder and installation errors
Making the advice clear enough for people who are not experts

Accomplishments that we're proud of

A working prototype that runs on a laptop, built as second-year students
A live dashboard where moving one slider instantly changes stock, cost and sales charts
A clear three-way comparison: normal, disrupted and recovery plan