> For the complete documentation index, see [llms.txt](https://docs.supademo.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.supademo.com/ai-demo-agent/agent-analytics.md).

# Agent Analytics

The Monitor section gives you visibility into how your agent is performing. This page covers the **Insights** dashboard and the **Sessions** table.

{% @supademo/embed demoId="cmp78yujp1ub3qmq66yk9f2qa" url="<https://app.supademo.com/demo/cmp78yujp1ub3qmq66yk9f2qa>" %}

### Insights

The Insights dashboard provides a high-level view of agent performance over a configurable date range (default: last 30 days).

<figure><img src="/files/eI7sZiU3DFtLwVKZa3Ht" alt=""><figcaption></figcaption></figure>

#### KPI Cards

| Metric                   | Description                                                                                              |
| ------------------------ | -------------------------------------------------------------------------------------------------------- |
| **Sessions**             | Total number of conversations started.                                                                   |
| **Avg Duration**         | Average conversation length in seconds.                                                                  |
| **High Intent Sessions** | Number of sessions where the visitor showed high buying intent. Click to jump to filtered Sessions view. |

Each KPI shows a period-over-period comparison so you can track trends.

#### Visualizations

* **Sessions over time** -- Stacked bar chart showing conversations per day/week/month, broken down by intent level (high, medium, low, unscored).
* **Intent breakdown** -- Donut chart showing the distribution of buying intent across all conversations.
* **Top content** -- Ranked list of your most-recommended content sources with recommendation counts. Shows which demos, videos, and documents are resonating.

***

### Sessions

The Sessions tab shows every visitor conversation in a searchable, filterable table. The table auto-refreshes every 30 seconds to surface active conversations.

<figure><img src="/files/HC2afK3UcwDE04OJ01m2" alt=""><figcaption></figcaption></figure>

#### Session Data

| Column       | Description                                                                                                |
| ------------ | ---------------------------------------------------------------------------------------------------------- |
| **Visitor**  | Name, email, role, company, and primary use case. A green pulse dot indicates an active (ongoing) session. |
| **Messages** | Total message count in the conversation.                                                                   |
| **Duration** | How long the conversation lasted.                                                                          |
| **Credits**  | Credits consumed by this session.                                                                          |
| **Time**     | When the conversation occurred (relative time with absolute tooltip).                                      |
| **Intent**   | Buying intent score: **High** (green), **Medium** (yellow), or **Low** (gray).                             |

#### Filters

| Filter         | Options                                                      |
| -------------- | ------------------------------------------------------------ |
| **Search**     | Filter by visitor name or email.                             |
| **Intent**     | All / High / Medium / Low                                    |
| **Status**     | All / Active / Ended                                         |
| **Duration**   | Any / Quick (< 2 min) / Short (2-10 min) / Engaged (10+ min) |
| **Date range** | Configurable, defaults to last 90 days.                      |

#### Conversation Detail

Click any session row to view the full conversation detail, including:

* Complete message transcript (every visitor and agent message).
* Content recommendations the agent made during the conversation.
* Visitor information (name, email, company, role, use case).
* Session metadata (duration, device type, referrer, UTM parameters).
* Buying intent assessment and summary.

## Knowledge Gaps

Knowledge Gaps surfaces questions your agent struggled with, detected automatically from conversation analysis. This is your improvement roadmap -- a continuous feedback loop between what visitors ask and what your knowledge base can answer.

<figure><img src="/files/LMCXhEsTx6gXJDbEKBMr" alt=""><figcaption></figcaption></figure>

***

### How It Works

After each conversation, the system analyzes the agent's responses to identify questions where it had insufficient content. Similar questions are grouped together under a normalized/canonical form so you see patterns, not noise.

***

### Active vs. Resolved

| Tab          | Description                                     |
| ------------ | ----------------------------------------------- |
| **Active**   | Gaps that need resolution. Shows a count badge. |
| **Resolved** | Previously addressed gaps. Shows a count badge. |

***

### Gap Details

Each knowledge gap shows:

* **Question** -- The canonical form of what visitors asked (similar questions are grouped).
* **Severity** -- How badly the agent struggled:
  * `Unanswered` -- The agent had no relevant content at all.
  * `Low Confidence` -- The agent had some content but wasn't confident in its answer.
* **Recommendation** -- What to do about it:
  * `Create a demo` -- Build an interactive demo covering this topic.
  * `Upload a document` -- Add a PDF, deck, or case study that addresses it.
  * `Add text context` -- Write a text snippet with the missing information.
  * `None` -- No specific recommendation (may require custom judgment).

***

### Resolving Gaps

Each gap card includes a CTA that navigates directly to the relevant Knowledge Base tab:

1. **Create a demo** takes you to the Demos tab.
2. **Upload a document** takes you to the Documents tab.
3. **Add text context** takes you to the Text tab.

After adding the missing content, return to Knowledge Gaps and mark the gap as resolved.

***

### Best Practices

* **Review weekly** -- Check active gaps at least once a week to stay ahead of visitor needs.
* **Prioritize by severity** -- Address `Unanswered` gaps first, then `Low Confidence`.
* **Watch for patterns** -- Multiple gaps around the same topic area signal a significant content hole.
* **Use Text for quick fixes** -- While creating a demo or uploading a document is ideal, adding a Text snippet is the fastest way to close a gap.
* **Track your readiness score** -- As you resolve gaps and add content, your Agent Readiness score should climb. Aim for 80%+.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.supademo.com/ai-demo-agent/agent-analytics.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.