Skip to content

Gen3 Conversations

GT AI OS Release edited this page Jun 11, 2026 · 3 revisions

Conversations

Start Here

  1. Open Conversations from the tenant sidebar.
  2. Search or filter for the thread you need.
  3. Reopen a conversation into GT Chat or export when supported.
  4. Use exports for records management after significant workflows.

Why this matters

Conversation history is the system of record for what the agent said, when, and under which agent configuration.

Details

Conversations is the active Gen 3 history workspace. Use it when you need to search across stored threads, filter by agent or dataset, inspect a conversation outside the live chat route, or export history in bulk.

What the page supports

  • standardized filter toolbar (same agent-style pattern as Users and Model Catalog): search with icon, agent/dataset/time dropdowns, match-count pill, and Reset all
  • searching titles and message content
  • filtering by agent, dataset, and time window
  • reviewing the selected conversation transcript
  • reopening a conversation in GT Chat
  • exporting filtered history as JSON or DOCX
  • exporting an individual conversation as JSON or DOCX

How it differs from GT Chat

Use GT Chat for active work. Use Conversations when the work is complete and you need to find, compare, review, or export prior threads.

Standard review workflow

  1. Open Conversations.
  2. Enter a search term if you know part of the title or a message phrase.
  3. Narrow the list with agent, dataset, or time filters.
  4. Select the conversation you want to inspect.
  5. Export it or reopen it in chat if you need to continue the thread.

Filters and exports

Search

Search checks conversation titles and message text, so it is the best first step when you remember content but not the exact thread name.

Agent and dataset filters

These filters help when you want to review only the work associated with a specific agent or the dataset-backed conversations for a particular knowledge source.

Time filter

Use time filtering to isolate recent activity such as Today, This week, or This month before exporting or reviewing.

Export formats

  • JSON is best for machine-readable archival or downstream processing
  • DOCX is best for a human-readable document handoff

Message-level actions

Within the selected conversation detail view, Gen 3 keeps the same practical response actions exposed in live chat:

  • copy individual message content
  • download the conversation
  • move back into the live chat workspace when the thread should continue

Best practices

  • Apply filters before bulk export so the output matches a real business question.
  • Prefer reopening the exact conversation in chat instead of starting over when context continuity matters.
  • Export a single conversation when you need one clean handoff, and use filtered exports for reporting or audit packets.

Related pages

Clone this wiki locally