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

# AI Interviewer

### AI Interviewer

The **AI Interviewer** enables automated, human-in-the-loop validation of investigation findings by conducting intelligent interviews with users involved in an alert.

This capability helps confirm or refute AI-derived conclusions using direct human input, improving accuracy while reducing analyst workload.

### The AI Interviewer Page

The AI Interviewer page is located in the left navigation under AI Setup & Tuning.

<figure><img src="https://435022081-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FmBVcKuiytGCsIDpL70BC%2Fuploads%2Fgit-blob-25afb433bb70fb6bf0fbaacee344fd415802970e%2Fui-settings-interviewer-config.png?alt=media" alt=""><figcaption><p>The AI Interviewer Configuration Settings</p></figcaption></figure>

Here you can:

1. See a complete list of all existing interviewer rules
2. Enable/Disable the Interviewer
3. Select the [Communicator integration](https://docs.dropzone.ai/integrations/communicator) you wish to use to query your personnel
4. Create new interview rules
5. Manage automatic interviewing for individual rules

#### How the AI Interviewer Works

The AI Interviewer is programmed with Rules.

A Rule has two parts:

1. **Scenario**\
   The behavior flagged in Investigation that triggers a follow-up interview. These typically include actions that are suspicious but could plausibly be benign, e.g. logging in to work from a new device or country.
2. **Conclusion Triggers**\
   The Investigation outcomes that a rule applies to, e.g. `Inconclusive`, `Suspicious`, `Malicious`.

Rules can be either Active (beginning interviews automatically when an investigation matches their scenario and conclusion trigger) or Inactive (requiring manual approval to send messages). You can assign this in the AI Interviewer Page.

The process of the AI Interviewer is as follows:

1. **Initial Investigation Completion**\
   The AI completes an investigation and determines that human verification is required.
2. **Interview Proposal**\
   The system generates a proposed interview with targeted, context-aware questions.
3. **Auto-Approval Check**\
   Rules determine whether the interview is automatically approved or requires manual approval. If a scenario does NOT cover auto approval then the interview will go into **pending** state until it is approved.
4. **Interview Execution**\
   The AI conducts the interview with the identified user using the configured communication channel.
5. **Result Integration**\
   Interview responses are incorporated directly into the investigation’s findings and final conclusion.

{% hint style="info" %}
Any changes to conclusion will be noted in the **change log** of the investigation and the entire interview will be available in the [Investigation Interview Tab](https://docs.dropzone.ai/best-practices/deep-dive-into-investigation-reviews#interviews-tab) in detail for audit or review.
{% endhint %}