Inspiration
Small businesses often struggle to track spending patterns and inventory needs from piles of work order receipts. We wanted to build a smart assistant that not only reads those receipts but also provides actionable insights — helping owners make data-driven decisions about what to buy more of and where to cut costs.
What it does
Popo AI analyzes uploaded work order receipts and generates real-time advice on inventory management and spending optimization. It suggests which items should be restocked, which ones to reduce spending on, and produces a generated receipt summary that visually lists all recommendations.
How we built it
We built Popo AI using React and JavaScript for the frontend and Node.js for the backend. The NVIDIA Nemotron model powers the AI’s analytical engine, interpreting text from receipts and generating intelligent spending recommendations. Git and GitHub were used to manage collaboration across our team — creating branches, reviewing pull requests, and pushing commits frequently to ensure everyone was updated and able to contribute simultaneously.
Challenges we ran into
One of the biggest challenges was connecting the AI model to read and understand uploaded receipt files and outputting its responses in a structured format through the React interface. We also faced challenges in synchronizing work among multiple developers, handling merge conflicts, and keeping all branches aligned with the latest updates.
Accomplishments that we're proud of
We successfully built a working prototype that ties together multiple AI systems and a clean user interface. Seeing the AI analyze real receipts, provide accurate inventory suggestions, and automatically generate a polished summary felt like a huge step toward real-world automation.
What we learned
We learned how to integrate AI models into full-stack applications, manage real-time updates in React, and coordinate a collaborative development workflow using Git branching strategies. We also gained a deeper understanding of how to handle AI-generated output and present it in an intuitive, user-friendly format.
What's next for Popo AI
Next, we plan to enhance Popo AI’s analytics by adding trend tracking over time, automatic expense categorization, and integration with point-of-sale systems. We also want to optimize the AI’s accuracy and improve collaboration tools to make the project even more scalable for larger teams and businesses.
Log in or sign up for Devpost to join the conversation.