The agent metrics that predict performance, and the ones that just look busy

Twenty-four metrics that show up on contact center scorecards, each with a verdict, how to measure it properly, and how it gets gamed. Mark the ones on your scorecard and the audit on the right tells you what it is really rewarding. AI-to-human handoff metrics included, because that is where 2026 scorecards are blind.

Predicts performance: put it on the scorecard Diagnostic only: measure it, never target it Drop it from the agent scorecard

The handoff rule: segment every agent metric by origin

A contact that reaches a human after an AI agent tried first is harder by selection. The easy ones were resolved upstream. Blended AHT, CSAT and FCR punish the agents who take your hardest work. Report every agent metric twice, once for direct contacts and once for AI handoffs, and set targets separately. Then measure the handoff itself with the six metrics below.