Jeremy Neely Text Jeremy

AI & Agents · Production

Hatz Usage Watch

Tray app that tracks every Hatz AI client's monthly credit use and warns early.

Built for: The AME Group's Secure AI admins and account managers

One client's credits this period, with pace and limit lines (sample data).
One client's credits this period, with pace and limit lines (sample data).

The problem

The AME Group resells Hatz AI to many client tenants, each with a monthly credit limit. The only way to see who was close to running out was to check each tenant by hand, so overages were found after the fact.

What I built

I built a Windows tray app that polls the Hatz admin API for every tenant and colors its icon by how clients are tracking. It forecasts each client, opens a dashboard with drill-down, and emails a branded PDF report when a threshold is crossed.

How it works

  1. Poll the Hatz admin usage endpoint for every tenant.
  2. Forecast each client from pace, history and usual monthly shape.
  3. Update the tray icon color, tooltip and menu.
  4. Open the dashboard for per-client drill-down on demand.
  5. Email the client a report once per period when a threshold is crossed.

Built with

  • C#
  • .NET 8
  • WPF
  • WebView2
  • Chart.js 4.4
  • Serilog

More of Jeremy's work