Improve agent quality with Insights Agent and Multi-turn Evals, now in LangSmith

LangSmith has released new capabilities to help monitor and improve AI agents in production. The new Insights Agent automatically analyzes production traces to discover and categorize common usage patterns, agent behaviors, and failure modes. A second feature, Multi-turn Evals, allows for scoring complete agent conversations to determine if the user's overall goal was accomplished, rather than just evaluating individual steps. These tools aim to give AI teams better visibility into how their agents are performing and where to focus improvements.