Inspiration

SMBs should not need to suffer through messy Excel files to understand their spending. We built Finance Command Center to make expense data speak clearly, catch problems faster, and help finance teams make smarter decisions ๐Ÿ’ก

What it does

Finance Command Center meets all 4 required features. It lets managers ask spending questions in plain English, checks every transaction against Brimโ€™s policy, sends risky spend to an AI approval queue, and creates clean expense reports automatically ๐Ÿš€

How we built it

We built it with React, TypeScript, FastAPI, Pandas, and Gemini through OpenRouter. Gemini plans and explains, while Pandas does the real calculations, charts, policy checks, grouping, and exports so the results stay accurate ๐Ÿ› ๏ธ

Challenges we ran into

The hardest part was making the app powerful without making fake assumptions. Some Excel files may miss names, receipts, approvals, or departments, so we built data honesty into the system. If a column is missing, the app says so instead of inventing it โš ๏ธ

Accomplishments that we're proud of

We built more than the basic brief. Along with the 4 required features, we added risk scores, fraud patterns, duplicate detection, split-charge detection, vendor insights, spend leaks, peer benchmarks, forecasting, editable policy rules, and a full audit trail ๐Ÿ†

What we learned

We learned that strong finance AI must be trustworthy. Rules should detect problems, Pandas should calculate the facts, and AI should explain the story. This makes the system useful, clear, and safe for real finance teams ๐Ÿง 

What's next for Mpc hacks

Next, we want live Brim card data, multi-user accounts, saved reports, real-time alerts, stronger approval workflows, and deeper budget forecasting. The goal is to turn Finance Command Center from a hackathon prototype into a real SMB finance intelligence platform ๐Ÿ”ฎ

Built With

Share this project:

Updates