Skip to main content
Track and control AI service costs across the platform in real time.

​
How cost tracking works

Every AI action follows a lifecycle:
  1. Budget check — verify the user has sufficient credits
  2. Cost reservation — reserve the estimated cost before execution
  3. Processing — the AI action runs
  4. Finalization — actual cost is recorded and any difference is reconciled

​
Daily reports

Admins receive daily summaries including:
  • Total AI spend
  • Number of unique users
  • Average and 95th percentile cost per user
  • Failed action count
  • Fallback and refund totals

​
Alerts

Automated alerts notify admins of:
  • Budget threshold breaches
  • Unusual usage patterns
  • Service failures
  • Cost anomalies

​
Rate management

AI service pricing is managed with versioned rates per provider, model, and action type — including resolution-specific pricing and effective date tracking.