Skip to main content
Every conversation between a visitor and your chatbot is logged in the Conversations section. Use it to review what visitors are asking, check how well your bot is performing, and take action when needed.

​
Conversation List

The main view shows all conversations across your workspace, sorted by most recent activity. Each row shows:
  • Visitor name (or “Anonymous” if unidentified) and reference ID
  • Company name, if available
  • Contact badge if the visitor has been captured as a contact
  • Verified identity badge (shield icon) for identified users
  • Conversation summary excerpt
  • Topic tags (up to 3, with a count for additional topics)
  • Satisfaction indicator (thumbs up, thumbs down, or mixed) — reflects either message reactions or the visitor’s own end-of-conversation CSAT rating, whichever was submitted
  • Spam classification dot
  • Message count and last activity time
  • Status indicator (New, Open, Awaiting reply, Snoozed, or Closed)

​
Searching and Filtering

Use the controls at the top of the list to narrow down conversations.

​
Status Filter

Filter by conversation state: Escalation to a human is tracked separately from status — escalated conversations show a flag and appear in the Live Inbox, which is also where you snooze conversations and set snooze presets. Type in the search box to find conversations by message content or visitor name. Results update as you type.

​
Bulk Actions

Select rows with the checkboxes to Archive, Tag, or Export the selected conversations from a toolbar at the bottom of the screen. For real-time actions — claiming, assigning, replying, snoozing, or closing — use the Live Inbox, which offers a fuller bulk-action set (up to 100 conversations per call).

​
Identified Users Only

Toggle this on to see only conversations from visitors who have been verified through user verification. Useful for filtering out anonymous traffic.

​
Satisfaction Health Banner

When your satisfaction scores drop below a healthy threshold, a banner appears at the top of the list with a warning. This helps you catch quality issues early.

​
Conversation Detail

Click any conversation to open the full thread.

​
Message Thread

The main panel shows the complete message history between the visitor and your bot. Each message shows:
  • Who sent it (visitor or bot)
  • The message content
  • Visitor reactions (thumbs up or thumbs down) on bot responses
Reactions give you direct feedback on which responses worked and which didn’t. Use this to identify where your knowledge base needs improvement.

​
Contact Sidebar

When the visitor has been captured as a contact, a sidebar shows their details:
  • Name, email, and company
  • Contact status (New, Active, Qualified, Unqualified, Spam, Archived)
  • Team member assignment dropdown
  • Tags
  • Spam classification with override toggle
  • AI-generated insight about the contact
  • Verified identity badge
You can update the contact’s status and assignment directly from the conversation view.

​
Conversation Metadata

Below the sidebar, you’ll see:
  • Conversation status and resolution reason
  • Created and last updated timestamps
  • Total message count
  • Whether the conversation was escalated

​
Collapsible Sections

Three expandable sections appear below the conversation thread:
  • Form Submissions — If the visitor filled out any forms during the conversation, submissions are shown here with all field data.
  • KB Retrieval Logs — Shows which knowledge base documents the AI used to generate its responses. Useful for debugging incorrect answers — you can see exactly what content the bot referenced.
  • Conversation Notes — Internal notes added by you or your team. Notes are never visible to visitors. Flag a note as Note, Issue, or Reviewed to organize them, and @mention a teammate to notify them directly — they get an in-app notification and, unless they’ve opted out, an email.

​
Understanding Escalation

A conversation is escalated when the bot determines it cannot help the visitor or the visitor explicitly asks for human assistance. Escalation is tracked separately from conversation status (there’s no “Escalated” status filter) — escalated conversations show a flag and are best worked from the Live Inbox, which lets you filter to unassigned/escalated conversations directly. When you see escalated conversations, review them to understand:
  • What the visitor needed that the bot couldn’t provide
  • Whether adding knowledge base content would prevent similar escalations
  • If the visitor’s question was outside the scope of your chatbot

​
Stats Bar

At the top of the conversation list, a stats bar shows counts:
  • Open — Workspace-wide count of active conversations (not affected by the current filter)
  • Needs attention — Workspace-wide count of fresh conversations not yet picked up (only shown when greater than zero); this is distinct from escalation
  • Total — Count of conversations matching your current filter and search
These update automatically as new conversations come in. For SLA breach and at-risk badges on individual conversations, see SLA Targets — those are tracked and surfaced in the Live Inbox.

​
Tips

  • Review escalated conversations first. They represent the biggest opportunities to improve your bot.
  • Check KB retrieval logs when the bot gives a wrong answer. The log tells you exactly which documents were used, so you know what to update.
  • Use conversation notes for team communication. If you spot something that needs follow-up, leave a note so the next person who reviews the conversation has context.
  • Monitor satisfaction trends. If thumbs-down reactions increase, check recent conversations to understand what changed.