Jun 11, 2026: Cortex Agent tool evaluation metrics (Preview)

With this preview release, Cortex Agent evaluations include two new system metrics that measure how your agent uses tools:

  • Tool selection accuracy (TSA) – Whether the agent’s orchestration layer invokes the tools you expect for a given query.
  • Tool execution accuracy (TEA) – Whether each tool that runs receives appropriate input and returns output that meets your requirements.

You provide expected tool calls in your evaluation dataset through the ground_truth_invocations key, and enable the metrics in the Agent Evaluation YAML or with the system metric toggles in Snowsight.

For more information, see Tool metric ground truth (TSA and TEA).