AI spend per employee is one of the most useful numbers a finance leader can track, because it normalizes for company size. A fifty-person company and a five-hundred-person company cannot compare raw AI totals, but they can compare cost per employee, and that comparison links AI spend directly to headcount planning.
What the metric captures
Per-employee tracking works best for the productivity layer of AI spend: the tools individuals use to do their own work, from coding assistants to writing and research tools. This spend is headcount-driven, so it should scale with hiring and shrink with attrition in a predictable way. When it does not, that is a signal worth investigating.
How to track it
Map every AI tool subscription and seat to a named user, then roll up by headcount. For consumption-based tools, attribute usage to individuals through per-user keys or metadata where possible. Divide total productivity AI spend by active employees to get the per-employee figure, and trend it over time rather than reading it as a single snapshot.
What the number tells you
A per-employee figure that is climbing faster than adoption suggests tool duplication, where different teams pay for overlapping tools nobody consolidated. A figure that is flat while adoption grows suggests good efficiency. A figure that is high with a clear, high-value use case is not a problem at all. The number surfaces the question; the investigation answers it.
Use it for planning
Per-employee spend is what makes AI budgeting predictable. Once you know your steady-state cost per employee, projecting next year's productivity AI budget is a matter of multiplying by expected headcount, adjusted for how adoption is trending. That turns a volatile-feeling line item into a plannable one.
Per-employee tracking is the benchmark that keeps AI spend legible as the company grows. Track it, trend it, and use deviations as prompts to look closer rather than as verdicts on their own.