Monthly AI Spend

Monthly AI Spend is the overall AI budget for this tenant for the current calendar month: a usage meter against a cap. It is independent of Total AI Spend (subscription lifetime) and of Per-scan AI budget (a ceiling on a single Constantine-style scan).

How the cap is applied

Usage resets on a calendar month. When the tenant has not been given a customer budget yet, the card shows Not provisioned yet — for customer users, that cap is set by the Guard Intelligence tier, not from this screen.

Scheduled Hannibal hunts already skip when remaining headroom is $0. You do not have to turn Automatic Penetration Tests off for that.

Who can edit the cap

Only a super-admin. Customer admins can still change Category Allocation (how the cap is split) but cannot raise or lower the cap. Non-super-admins see Edit locked; customers are pointed at support@praetorian.com.

Lowering the cap below allocations

Category percentages are stored as dollar limits that must sum to the cap. If you save a new cap below the sum of those category limits, Guard clears the category allocation before saving the cap. The UI warns first: Saving a lower budget clears the category allocation. After that, inference uses the shared platform budget until someone sets allocation again.

The API rejects a cap below the configured allocation total. You cannot keep a $10k Hannibal slice on a $5k monthly cap.

What this card does not do

It does not edit per-category percentages (Category Allocation). It does not set the per-scan ceiling (Per-scan AI budget). It does not enable or disable hunts or Cato.