> ## Documentation Index
> Fetch the complete documentation index at: https://docs.cogniagent.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Conversations history

> Browse every conversation your flow has handled — transcripts, routing traces, tool calls, and delivery status.

Every conversation your flow handles — test sessions, real inbound traffic, outbound campaigns — is stored and browsable. The **Conversations** drawer is where you go to see what's happening on a live flow, debug a misbehaving conversation, or audit historical activity.

<Frame caption="Conversations history with one test session.">
  <img src="https://mintcdn.com/glorium/eqWkhSfBUec9afZU/images/conversation-flows/conversations-drawer.png?fit=max&auto=format&n=eqWkhSfBUec9afZU&q=85&s=2ad23b0427829f46bcdf8f3982bd4681" alt="Conversations drawer" width="1471" height="1225" data-path="images/conversation-flows/conversations-drawer.png" />
</Frame>

## Opening the drawer

In the bottom toolbar of a flow's editor, click **Conversations**. The drawer opens on the right.

## The list view

Each row is one conversation. You see:

* **AI-generated title** — a short summary written by the platform (e.g. "Inquiry regarding SSO support in Acme Pro vs. Enterprise").
* **Participant** — who's on the other end (email, Slack handle, phone number, or "Test user" for test sessions).
* **Status** — `in-progress` or `completed`.
* **Test badge** — if the conversation came from the [Test drawer](/conversation-flows/operate/test-flow).
* **Message count** — how many turns.
* **Last activity** — when the most recent turn happened.

### Filters and search

* **All / In Progress / Completed** — switch which conversations show.
* **Search** — filter by participant name or identifier.

## The detail view

Click a conversation to see the full transcript.

<Frame>
  <img src="https://mintcdn.com/glorium/eqWkhSfBUec9afZU/images/conversation-flows/conversation-detail.png?fit=max&auto=format&n=eqWkhSfBUec9afZU&q=85&s=c3160467f69180d69055aab9996886a5" alt="Conversation detail view" width="1471" height="1225" data-path="images/conversation-flows/conversation-detail.png" />
</Frame>

What you see per turn:

* The user's message (left-aligned).
* The actor's reply (right-aligned, with actor name and avatar).
* **Routing trace** below the actor reply — same decisions you see in the [Test drawer](/conversation-flows/operate/test-flow).
* **Tool call badges** — when the actor called an integration action or custom tool, a small badge shows the tool name, duration, and token count.
* **Delivery status** — for outbound messages, whether the platform actually delivered the reply on the channel.

## Common things you'll do here

<AccordionGroup>
  <Accordion title="Debug a conversation that went sideways" icon="bug">
    Open the conversation, scroll through the turns, read the routing traces. Most "why did the actor do X?" questions are answered by the trace.

    Look for:

    * Routing decisions that picked the wrong actor (fix activation prompts).
    * Tool calls with wrong arguments (fix actor instructions).
    * Done() calls that fired prematurely (fix Definition of Done).
  </Accordion>

  <Accordion title="Confirm outbound delivery" icon="paper-plane">
    For Email/Slack/Teams channels, each outbound message has a delivery status badge. If it shows "failed", the integration connection probably hit an error — check the connection's status.
  </Accordion>

  <Accordion title="End a stuck conversation" icon="square-x">
    Some conversations linger — the user never replied, the session timed out, but it's still marked `in-progress`. Use the **End session** button to close it manually.
  </Accordion>

  <Accordion title="Replay a problem case in Test mode" icon="rotate-right">
    Found a real conversation that exposed a routing bug? Copy the user's messages and replay them in the [Test drawer](/conversation-flows/operate/test-flow). Edit prompts, retest, deploy when fixed.
  </Accordion>
</AccordionGroup>

## Privacy and retention

* Conversations are stored per workspace and visible to workspace members.
* Phone calls are stored as **text transcripts only** — no audio is retained.
* Test sessions live alongside real ones with the TEST badge. They're never delivered to real channels.

## Next

<CardGroup cols={2}>
  <Card title="Test a flow" icon="flask" href="/conversation-flows/operate/test-flow">
    Replay a problem case in the sandbox.
  </Card>

  <Card title="Deploy a flow" icon="rocket" href="/conversation-flows/operate/deploy">
    Take edits live after debugging.
  </Card>
</CardGroup>