Skip to main content
Conversations give you full visibility into every chat happening across your agents and deployments. Use them to review response quality, investigate specific interactions, and export conversation data for compliance or reporting purposes.
Conversations display in real time. As new conversations occur, they appear in the list automatically — use the Refresh button to pull in the latest.

​
Accessing conversations

Navigate to Monitor → Conversations in the left sidebar. The page uses a two-panel layout: a list of conversations on the left, and the full conversation thread on the right.

​
Filtering conversations

Use the filter bar at the top to narrow down the conversation list:
  • Agent Type — Show conversations for a specific agent type, such as Patient Navigator or HCP Clinical Decision Support.
  • Deployment — Narrow to a specific widget deployment.
  • Time Range — Select 7, 30, 60, or 90 days.
Use the deployment filter to compare conversation quality across different widget deployments. This is especially useful when you’re running the same agent on multiple websites and want to see if the audience differs.

​
Conversation list

Each entry in the list shows:

​
Reviewing a conversation

Select any conversation to view the full thread in the right-hand panel:
  • Message bubbles — User messages and agent responses in a familiar chat layout, clearly distinguished by color, with formatted (markdown) agent responses.
  • Citations — Agent responses show the knowledge base sources they drew from.
  • Follow-up suggestions — Any suggested follow-up questions offered to the user.
Conversations link directly to other areas of the platform. For example, adverse events detected in a conversation link back to the relevant conversation thread for full context.

​
Exporting conversations

Click the Export CSV button to download all conversations matching your current filters. The export includes:
  • Conversation ID
  • User identifier
  • Deployment name
  • Timestamps
  • Full message content (user and agent messages)
CSV exports are generated based on your active filters. Apply the filters you need before exporting to ensure the data is relevant and manageable.

​
Common use cases

Quality assurance

Review recent conversations to check response quality and accuracy. To find conversations about a specific medicine or therapy area, set up Topics and use the Content tab on the Activity page.

Compliance review

Export conversation data for internal compliance audits. Use time range filters to pull data for specific reporting periods.

Knowledge gap identification

Look for conversations where the agent struggled to provide useful responses. These gaps indicate areas where your knowledge base could be strengthened.

Deployment comparison

Filter by deployment to compare conversation patterns across different widget placements. This helps you understand how different audiences interact with your agent.