Skip to main content
Chat History gives builders and data teams a complete view of every conversation the agent has had — both sessions run in Agent Studio and live chats from deployed end users. It’s where you go to understand how the agent is actually being used, catch answers that need improving, and access detailed observability traces.

​
The chat history table

The Agent Chat History view lists all conversations with this agent across all users. Each row shows:
  • Date — When the conversation took place
  • Duration — How long the session ran
  • Tenant — The organization the user belongs to
  • Agent — Which agent handled the conversation
  • View Type — Whether the user saw an End User View (clean, no technical details) or Admin View (full builder view with tool calls visible)
  • Embedded — Whether the agent was accessed through an embedded integration
  • Feedback — Any thumbs up/down submitted by the user
You can export the full table as a CSV using the Export CSV button.
Agent Chat History table showing all conversations with date, duration, view type, and feedback columns

​
Viewing a conversation

Click any row to open a slide-in panel showing the full conversation thread and its metadata — Agent Name, View Type, Embedded status, Feedback, and Feedback Comment. The Chat Messages section shows the full exchange exactly as it occurred. The Show Details toggle at the top of the panel controls how much technical context is surfaced. When off, you see the conversation as the end user saw it. When on, you see additional trace information alongside the messages.
Chat History panel open on the right with a conversation selected, showing metadata and chat messages

​
Accessing observability from a conversation

From any open conversation, you can drill into the full observability trace for any response. With Show Details toggled on, click View Observability at the bottom of the relevant exchange to open the step-by-step breakdown of what the agent did. See Agent Observability for a full explanation of what the trace contains and how to use it to diagnose and improve agent behavior.

​
Next steps

Agent Observability

Inspect the step-by-step trace of any agent response.

Golden Assets

Add example queries and charts to fix patterns you identify in past conversations.