> For the complete documentation index, see [llms.txt](https://mind-expression-docs.gitbook.io/home/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://mind-expression-docs.gitbook.io/home/how-to-guides/audit-ai-activities.md).

# Audit AI Activities

You can audit every AI activity on Conversation History. You can access the page on the Scope details page, and you will see the lists of all conversations. There are two types of sessions: Live sessions which are the actual conversations between AI and your customers, and Sandbox sessions from Sandbox tests. To learn more about Conversation History, see [Concepts: Conversation History](https://me-dev.mind.ai/docs/concepts/conversation-history).

## Audit sessions

To audit a session, click one you'd like to see the details of. If the session includes fallback, you will see the list of causes in the Points of Failure panel. In the Conversation panel, you can see every customer message and AI response. To see the detailed information about an AI response, click on the AI message. The details will appear in the AI Action details panel.

#### Filter Sessions

For a custom list view, you can apply filtering conditions.

* **Filter by session types**

  You can view the sessions with customers in the Live tab, and test sessions from Sandbox are available for the review in the Sandbox tab.
* **View sessions with exception**

  You can filter out and see sessions with fallback or escalation errors only in the list.
* **Filter by dates**

  Add a filter by dates to view sessions from yesterday, the last 3 days, the last 7 days, or manually set the dates as you see fit.

#### Check AI Action Details

AI action details consist of two parts: How AI Reasons and Conversation Model. To check AI Action Details, simply click an AI message in the Conversation panel.

* **How AI Reasons**

  How AI Reasons shows the customer input message and how AI interprets it.
* **Conversation Model**

  The Conversation Model shows the current Component of the message and the Component structure, webhook details for the webhook involved actions, how AI responds, and reference information if applicable.

### Fallback and escalation error

Sessions with fallback or an escalation error are displayed with a red error icon as well as the specific message including the issue on the details page. If a fallback occurs, you can see the reason for the fallback in the AI Action Details. For escalation errors, you can see the webhook request and response details and report the error to the Mind team if needed.

### Manage sessions

#### Delete Sandbox Sessions

Only Sandbox sessions can be deleted from the list.

![](https://mind-expression.s3.ap-northeast-2.amazonaws.com/dev/static/docs/media/how-to-guides/03-delete_sessions.png)

1. Hover over the session you want to delete from the Sandbox session list.

&#x20; 2\. Click the delete button.

&#x20; 3\. In the popup, confirm and click **Delete**.


---

# 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, and the optional `goal` query parameter:

```
GET https://mind-expression-docs.gitbook.io/home/how-to-guides/audit-ai-activities.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

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.
