How to measure AI agent ROI

AI agent ROI is not just model cost. A useful measurement links all attributable costs to a defined business outcome, then states how that outcome was attributed.

The practical model

• Define one outcome before the run: revenue, savings, or time saved.
• Track model, tool and external-service costs, not only token spend.
• Separate authorization from settlement: approved spend is not the same as actual spend.
• Record the evidence and attribution method for the outcome.
• Publish the methodology next to the number.

How Alem approaches it

Alem records costs separately from outcomes. The return multiple is attributed value divided by measured cost; ROI percentage is (value − cost) / cost. Outcomes may be owner- or integration-attributed, so the result should be labelled directional unless independently verified.

What is live today

Alem's sandbox includes company workspaces, agent registry, budgets, spend policies, human approvals, economic ledger, ROI, HTTP/OpenAPI and MCP. Live payment execution remains private beta / integration-stage.