Overview
Agent monitoring & analytics helps your team see where agents resolve customer requests, where service breaks down and what it costs to operate. Use Analytics to track outcomes and spend, Evaluations to assess conversations against your criteria, and Transcripts to investigate the messages and execution logs behind a result.
Open Analytics, Evaluations or Transcripts from your project sidebar.
Choose a starting point
- How is my agent performing? Compare trends in Analytics.
- How do I build my dashboard? Choose and arrange dashboard cards.
- Can I track my own quality criteria? Define an Evaluation, then add its metric to Analytics.
- What happened in a conversation? Review its Transcript and logs.
- What is driving operating cost? Inspect the spend breakdown.
Explore Agent monitoring & analytics
See resolution, customer feedback and operating spend together. Arrange dashboard cards and track your own quality metrics to decide what to improve.
Check conversations against your business criteria. Define metric types, test judgments and score new or historical conversations.
Find what happened in a customer interaction. Filter saved conversations and inspect messages, execution logs, metadata and Evaluation results.
From a finding to a tested change
Find a result to investigate
In Analytics, choose the Environment and period you want to review. Open a custom Evaluation from Quality to inspect its scored conversations.
Check the conversation behind the score
In the Evaluation’s Results view, recheck the dates and Environment: its filters are independent of the dashboard. Compare the score and reasoning with the Transcript’s messages and execution logs to identify the request, reply or tool result that needs attention.
Make and test a change
Use Agent studio to edit the relevant instructions, Playbook, Workflow or tool. Use Agent testing & QA to exercise the same situation in a Test or Simulation and review the result before publishing the change.