The Project — From Storage Software to RAG Chatbot

When we had identified the pain points, I started thinking about what we could actually build. My first instinct was to tackle the storage logistics problem, and my initial idea was to use Claude Code to spin up a simple storage management application — something where the company could list their inventory, keep track of items, and manage their stock.

The appeal of the idea was its flexibility. I wanted to make it malleable enough that they could adapt it to whatever scale they needed, without us having to anticipate every detail of how they work. The thought was to give them a platform where they could define and arrange their own entities — essentially just handing them the structure and letting them fill it in themselves.

But I had missed the point. The solution had to be AI-driven.

So I teamed up with a classmate and we went back to the drawing board. That’s when we found out something useful about how the company currently manages their storage: it lives in an Excel file. That detail changed everything.

We ended up settling on a RAG chatbot. The idea is straightforward — the company uploads their existing storage file, either as an .xlsx or .csv, and the application reads and chunks the data so it can be queried through a chat interface. No need to migrate to a new system or learn new workflows. They keep working the way they already do, and the chatbot sits on top of it as a natural language layer.

The end result would be a simple web app where they can upload their file and just ask questions about their inventory. Things like what’s in stock, how many of something they have, or whether a specific item is available — answered instantly, in plain language, without having to dig through a spreadsheet.

It’s a solution that’s simple enough to actually build, but genuinely useful. And this time, it has AI at the centre of it.