Nessence

Inspiration

We were inspired after hearing about the Nessie API at the opening ceremony on Friday, and we felt that we could make something effective with it.

What it does

This project functions as a budget-management program for companies. It contains three different user levels: Employee, Manager, and Finance. Employees can submit expense reports and reimbursement requests, uploading receipts or invoices for proof. Managers can approve reimbursement requests and submit budget increase requests for their department. Finance can see an overview of all Employees and Managers, their requests, and their transactions. Additionally, they can edit policies regarding auto-approval of requests. Some additional features include the ability to set auto-approval thresholds for reimbursement requests, and an LLM that parses uploaded receipts/invoices and makes sure that they are accurate to what is being requested.

How we built it

We built this project using Python. We used Flask for the frontend, rendering Html templates with Jinja, and SQLite for our database on top of the Nessie API in the backend, using python to make requests to the API.

Challenges we ran into

We initially misunderstood the exact behavior of the Nessie API. We originally thought it simulated performing transactions, when in reality it is more of a ledger that keeps track of transactions. To navigate this, we implemented our own database and code on top of the API that simulates performing transactions.

Accomplishments that we're proud of

We are proud of the AI integration we implemented. We believe that it significantly enhances the overall project.

What we learned

We strengthened our knowledge of version control with git, learned how to utilize AI into a project, and how to build a database that integrates with an API.

What's next for Nessence

The next steps would be to add more advanced policy options, more robust receipt/invoice parsing, more granular control of individual employees and employee budgets, and deploying the project for widespread use at scale.