Advisory
AI cost analysis and reduction
AI spend grows quietly, per seat, per tool, on several different cards.
What is in scope
- Full spend inventory across tools, APIs and per-seat subscriptions
- Usage against licences — the unused-seat count is reliably higher than anyone expects
- Model routing analysis: which work is running on an expensive model that a cheaper one handles fine
- Caching and prompt efficiency
- Build-versus-subscribe comparison at your actual volume
- A reduction plan, ordered by saving against effort
How the engagement runs
- Inventory. Every AI line item, including the ones in other departments' budgets.
- Usage analysis. Who uses what, how much, for what.
- Routing review. The largest single saving in most businesses is work running on a frontier model that a mid-tier model does equally well.
- Plan. Ranked actions with the expected saving against the effort to get it.
What determines the price
Fixed fee for the analysis. We publish our own engine token economics at AI engine pricing, which is the same data this analysis is built on.
When this is the wrong service
Stated plainly, because an engagement that should not have happened costs us more than it costs you.
- Your total AI spend is small. The analysis would cost more than it saves, and we will say so on the first call.
- You want us to cut spend regardless of effect. Some of that spend is earning its keep.
Bring us the process that keeps breaking
A short call is usually enough to tell you whether this is the right service.
