Case Study • Agentic AI Development

Budget Forecasting PWA

Designed and built for a School Division, this client-side Progressive Web App puts powerful budget forecasting and variance analysis directly in the hands of principals and budget owners — with no server infrastructure, no login, and full offline capability.

The Challenge

Budget management across a school division is no small undertaking. Principals, department leads, and budget owners each carry responsibility for their own slice of the division's finances, yet the tools available to them were largely manual, spreadsheet-driven, and disconnected from the division's financial system.

There was no purpose-built way to import real financial data, track actuals against budget, visualize spending trends, or model what the rest of the year might look like — without a significant amount of manual effort and a high risk of error.

The Solution

Using an agentic AI-assisted development approach, I designed and built a client-side Progressive Web App that puts powerful budget forecasting tools directly in the hands of the people who need them, with no server infrastructure required.

The application integrates directly with the division's Atrieve financial system, allowing users to import budget data and transaction details via CSV export. From there, the app automatically matches transactions to GL accounts, calculates actuals, and surfaces variance analysis across all budget periods. A PREPAID transaction handling system ensures that full-year prepayments are correctly identified and flagged rather than distorting monthly spending figures.

Beyond tracking, the tool offers scenario modelling, trend-based forecasting, and year-over-year comparison when prior year data is available. An analytics dashboard delivers visual insights through charts and graphs, and all data can be exported to CSV, JSON, Excel, or PDF for reporting and backup purposes.

Because all data is stored locally in the browser using IndexedDB, there is no authentication required, no data ever leaves the user's device, and the app works fully offline as an installable PWA.

The Development Approach

This project was built using agentic AI-assisted development, with Claude acting as a collaborative coding partner across a structured 12-phase build plan. Each phase delivered a working, testable version of the application, with user feedback incorporated before advancing to the next stage.

This approach allowed for rapid, high-quality delivery while maintaining full technical ownership and decision-making throughout.

The Result

Budget owners across the division now have a purpose-built tool that connects directly to their financial data, surfaces meaningful insights, and works reliably on Chromebooks without requiring an internet connection. What previously lived in disconnected spreadsheets is now a structured, visual, and exportable forecasting platform built specifically for the way the school division manages its finances.

12 Structured build phases delivered
0 Server infrastructure required
4 Export formats supported

Need a purpose-built tool that works with your data and your workflow?

I build practical, deployable applications for real operating environments — including client-side tools that require no infrastructure, no login, and no ongoing maintenance overhead.