---
title: "View ARGENT usage and chat history"
description: "Read the question count, token count, answer speed and satisfaction on the Dashboard, and review each user question and agent reply in Chat History"
canonical_url: "https://docs.ardge.com/en/knowledge/argent/usage-analytics"
---
# View ARGENT usage and chat history

> Read the question count, token count, answer speed and satisfaction on the Dashboard, and review each user question and agent reply in Chat History

## Overview

Once an agent is live, ARGENT records how it's actually used:

- The **Dashboard** page sums up usage metrics across all of ARGENT, such as the total number of questions, tokens consumed, average answer time and user satisfaction
- The **Chat History** page keeps every user's conversations, which you can view by session or by individual message

You can use the chat history to check whether the agent's answers meet your expectations, and to find questions people actually ask that your current file sources don't cover yet.

## What you'll learn

- Read the usage metrics on the **Dashboard** page
- Search and review user questions and agent replies in **Chat History**
- Use feedback from users to find answers that need improving

**Estimated time: about 5 minutes**

## Applies to

- ARGENT 1.0.1 and later

## Before you begin

- Your account has ARGENT administrator permissions
- Users have already asked the agent questions. With no records yet, **Chat History** shows **No chat history found**

## Step 1: Read the metrics on the Dashboard

1. Sign in to the ARGENT admin interface and go to the **Dashboard** page
2. The top of the page shows a tutorial on creating an agent. If you already know the flow, click the close icon at the top right to collapse it. You can click **Show tutorial** later to bring it back

The cards below the tutorial show the following metrics, counted across all of ARGENT:

<table>
<thead>
  <tr>
    <th>
      Metric
    </th>
    
    <th>
      What it means
    </th>
  </tr>
</thead>

<tbody>
  <tr>
    <td>
      <strong>
        Total File Sources
      </strong>
    </td>
    
    <td>
      The number of file sources created
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Total AI Agents
      </strong>
    </td>
    
    <td>
      The number of agents created
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Total Questions Asked
      </strong>
    </td>
    
    <td>
      The total number of questions all agents have received
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Tokens · Input
      </strong>
    </td>
    
    <td>
      The total tokens sent to the model, including users' questions and the retrieved document chunks
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Tokens · Output
      </strong>
    </td>
    
    <td>
      The total tokens the model generated — the answers
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Average Answer Time
      </strong>
    </td>
    
    <td>
      The average time from question to finished answer
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Average Search Time
      </strong>
    </td>
    
    <td>
      The average time per file content search
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Total Searches
      </strong>
    </td>
    
    <td>
      The total number of file content searches
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Satisfaction
      </strong>
    </td>
    
    <td>
      The share of user ratings on answers that are <strong>
        Likes
      </strong>
      
      , with the counts of likes and dislikes below it
    </td>
  </tr>
</tbody>
</table>

![The metric cards on the Dashboard](https://docs.ardge.com/screenshots/knowledge/en/argent/usage-analytics/step-1.webp)

## Step 2: Review the chat history

1. Go to the **Chat History** page
2. Use the toggle to the right of the search box to choose **Session** or **Message**
3. To find a particular record, enter a session name, user, agent or message text in the search box
4. Click any row to see the full conversation

The columns for each view are:

<table>
<thead>
  <tr>
    <th>
      View
    </th>
    
    <th>
      Each row is
    </th>
    
    <th>
      Columns
    </th>
  </tr>
</thead>

<tbody>
  <tr>
    <td>
      <strong>
        Session
      </strong>
    </td>
    
    <td>
      One complete conversation
    </td>
    
    <td>
      <strong>
        Session Name
      </strong>
      
      , <strong>
        Agent Name
      </strong>
      
      , <strong>
        User
      </strong>
      
      , <strong>
        Messages
      </strong>
      
      , <strong>
        Feedback
      </strong>
      
      , <strong>
        Last Message
      </strong>
      
      , <strong>
        Created At
      </strong>
    </td>
  </tr>
  
  <tr>
    <td>
      <strong>
        Message
      </strong>
    </td>
    
    <td>
      One question and its reply
    </td>
    
    <td>
      <strong>
        User Prompt
      </strong>
      
      , <strong>
        Agent Response
      </strong>
      
      , <strong>
        Feedback
      </strong>
      
      , <strong>
        Session Name
      </strong>
      
      , <strong>
        Agent Name
      </strong>
      
      , <strong>
        User
      </strong>
      
      , <strong>
        Created At
      </strong>
    </td>
  </tr>
</tbody>
</table>

Records are sorted newest first. You can change the number of rows per page at the bottom of the table.

![Chat History in the Session view](https://docs.ardge.com/screenshots/knowledge/en/argent/usage-analytics/step-2-1.webp)

![Chat History in the Message view](https://docs.ardge.com/screenshots/knowledge/en/argent/usage-analytics/step-2-2.webp)

![The full content of one session](https://docs.ardge.com/screenshots/knowledge/en/argent/usage-analytics/step-2-3.webp)

<tip>

**Use feedback to find answers that need improving:**

The **Feedback** column shows how many **Likes** and **Dislikes** users gave the answers. Switch to the **Message** view to go through the answers that got dislikes one by one, then add to your file sources or adjust the prompt accordingly.

</tip>

## Related articles

- [Create your first AI agent](https://docs.ardge.com/en/knowledge/argent/first-ai-agent)
- [Embed an AI agent on your website](https://docs.ardge.com/en/knowledge/argent/embed-on-website)


## Sitemap

See the full [sitemap](https://docs.ardge.com/sitemap.md) for all pages.