Finance · All retail

Compute Spend Governance

AI spend climbing by team. Caps per department, routes to the cheapest model.

Target KPI AI cost / department
Executes in ≤ 1 hour
System of record Ward routing layer
Trigger condition

A department passes 80% of its monthly compute budget before 80% of the month has elapsed.

Write-back

Routing policy updated in the Ward layer for the named workloads; budget alert and cap recorded against the department.

Every write is gated on an approver role you name. Nothing runs unattended.

Procedure
  1. Attribute spend to department, user and query pattern, so the conversation is about workloads rather than a single total.
  2. Identify the queries paying for a premium model without needing one, which is where most of the overage sits.
  3. Model the routing change: which workloads move tier, and what it does to answer quality on the eval set.
  4. Alert the department owner at 80% with the options: reroute, raise the cap, or stop. Never silently degrade quality.
  5. Close when the department lands inside budget with eval scores held.
Outcome metric

Cost per query for the department over 30 days with answer quality held on the eval set. Both, or the change is reverted.

The case closes on this number, not on the action being taken. A playbook without a close condition is a dashboard.

Run Compute Spend Governance on your data.

Pick three playbooks from the catalog. We wire them against your system of record for the pilot.

Read-only to start · your LLM keys · SOC 2 Type II underway · or book a call directly

Find out what your data has been hiding.

Tell us about your operation. We’ll show you the problems Ward catches, and the ones your current tools miss.

Step 1 of 3
What are your goals?
Step 2 of 3
About your operation
Step 3 of 3
Your contact info