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

# AI performance

> Track resolutions, volume, CSAT, and transfers over time

Open **Analytics > Chats**. This is the **AI agent performance** dashboard: the first place to look when you want to know whether the agent is actually resolving conversations, and whether that is trending up or down.

Pick a date range at the top of the page. Every metric shows the current period against the previous one, so a change reads as a comparison rather than a bare number.

## Headline metrics

| Metric                   | What it tells you                                          |
| ------------------------ | ---------------------------------------------------------- |
| **AI agent resolutions** | Conversations the agent resolved on its own in this period |
| **AI agent chats**       | Total conversations the agent handled                      |
| **AI agent CSAT score**  | Positive feedback share among rated messages               |
| **Transferred chats**    | Share of conversations escalated to a human                |

Below the headline cards, the **AI agent resolutions** chart plots resolutions over time, with a **See resolved chats** link that jumps to the underlying conversations in **Chat logs**. Further down, breakdowns show where conversations come from by channel and by geography.

## How to read it

* **Resolutions rising, transfers flat or falling** is the healthy pattern: the agent is absorbing more volume without pushing more work to your team.
* **CSAT dropping while volume grows** usually means new question types the knowledge base does not cover yet. Use the [Topic Explorer](/user-guides/analytics/topic-explorer) to find them.
* **Transfers rising** means escalation triggers are firing more. Check whether that is real customer frustration or over-eager keyword rules under **Agent settings > Security**.

Trends tell you *that* something changed. To find out *why*, drill into the conversations themselves.

<CardGroup cols={2}>
  <Card title="Topic Explorer" icon="tags" href="/user-guides/analytics/topic-explorer">
    What customers are actually asking about.
  </Card>

  <Card title="Conversation Evals" icon="flask" href="/user-guides/analytics/conversation-evals">
    Score real conversations for quality.
  </Card>

  <Card title="Activity" icon="inbox" href="/user-guides/agent-management/activity">
    Read the underlying transcripts.
  </Card>

  <Card title="Guardrails" icon="shield" href="/user-guides/settings/guardrails">
    Tune the escalation triggers behind transfer rates.
  </Card>
</CardGroup>