Microsoft Copilot Studio: Cost visibility in Agent Evaluations

Now makers will have clear visibility into their consumption costs directly associated with their evaluations, including a breakdown of usage for eval generation, test runs and model grading/judgement. Makers can see the costs as they test, iterate, and review conversations, making usage transparent throughout the development lifecycle.

Added to roadmap: 15 sep 2026 | Last modified: 15 sep 2026

This is in the weekly digest for week 38 · What changed this week? →