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

# Agent analytics

> How a single chat agent is performing — volume, resolution and answer confidence.

Each chat agent has its own analytics page, reachable from the agent's detail view. It covers that agent only; the workspace [dashboard](/workspace/dashboard) aggregates across all of them.

## Headline metrics

| Metric                    | What it tells you                          |
| ------------------------- | ------------------------------------------ |
| **Total conversations**   | Volume over the selected period            |
| **Resolution rate**       | Share of conversations that ended resolved |
| **Avg response time**     | How quickly the agent replies              |
| **Customer satisfaction** | Feedback from visitors, where given        |

## Answer quality

Every answer is classified by how well the agent's knowledge covered the question:

| Bucket                | Meaning                                | Action                         |
| --------------------- | -------------------------------------- | ------------------------------ |
| **Confident answers** | Well covered by your knowledge sources | Nothing                        |
| **Tentative answers** | Partially covered — the agent hedged   | Strengthen the relevant source |
| **Unanswered**        | Not covered at all                     | Add a source or an FAQ         |

Alongside these: **total queries**, **support requests**, **leads**, **bookings**, **tickets** and **resolved** counts.

<Tip>
  Unanswered questions are the highest-value list in Botric. Every one is a visitor who wanted something you didn't have — and each is a one-minute fix as an [FAQ](/workspace/knowledge-source).
</Tip>

## Using the numbers

<Steps>
  <Step title="Work the unanswered list weekly">
    Add an FAQ or knowledge source for each recurring question.
  </Step>

  <Step title="Convert tentative to confident">
    Tentative answers usually mean the content exists but is vague. Tighten it.
  </Step>

  <Step title="Watch resolution rate after each change">
    It's the cleanest measure of whether your knowledge base is improving.
  </Step>

  <Step title="Tie leads and bookings back to channels">
    If leads are flat, revisit the form on the [Lead Generation channel](/support/chat-agent/channels).
  </Step>
</Steps>

## Message history

How far back conversation history is retained is set by your plan. See [Plans](/account/plans).

<CardGroup cols={2}>
  <Card title="Conversations" icon="list" href="/support/conversations">
    The individual sessions behind these numbers.
  </Card>

  <Card title="Knowledge Source" icon="book-open" href="/workspace/knowledge-source">
    Close the gaps the analytics reveal.
  </Card>
</CardGroup>