Tired is data.

Most AI dashboards measure speed: throughput, time-to-task, completion rate. The metrics that match what we already wanted were measured before AI arrived.

What rarely gets measured is the cost of holding pace. The person who used to write three reports a week is now editing twelve. The customer-service rep who used to handle nine cases an hour is at fourteen. The throughput line on the slide goes up; the human side of the line — fatigue, vigilance, the residual stress of always-checking — is rarely on the same chart.

Tired is data. Not a complaint, not a soft-side afterthought, not a quarterly engagement-survey footnote. Data — as load-bearing as throughput — about whether the gains are repeatable next quarter.

Calibration matters more than throughput because it is the only metric that asks: at what cost? Without it, you optimise for a six-month win and absorb a two-year liability.