> ## Documentation Index
> Fetch the complete documentation index at: https://docs.goguardian.com/llms.txt
> Use this file to discover all available pages before exploring further.
# How Beacon Works
> GoGuardian Beacon uses machine learning to review student activity on school devices and alert staff to signals of suicide risk, self-harm, or harm to others.
GoGuardian Beacon uses machine learning (ML) to analyze student online activity on school-issued devices. When the ML models detect activity aligned with signals of suicide risk, self-harm, or harm to others, Beacon generates an alert for designated school staff to review.
## Understand the Detection Process
Beacon uses supervised learning — an ML technique where models are trained on labeled datasets to make predictions about new outcomes. Beacon's models are trained on datasets labeled by individuals with experience in recognizing and responding to mental health risks.
When deployed by a school, Beacon's ML models continuously analyze student online activity. If the models detect activity that matches the patterns they were trained to identify, Beacon generates an alert.
Designated school staff are always the ultimate decision-makers as to whether a Beacon-generated alert requires human intervention.
## Review Alert Contents
Each alert contains the following information to help the reviewing staff member make an informed, rapid decision about student safety:
* Date and time of the event
* 30-minute browsing history window (±15 minutes from the event)
* Student name and email address
* Guardian contact information (if available)
* Risk category
* URL
* Relevant content
* Screen captures (if available)
## Understand Alert Phases
Beacon categorizes each alert into one of five phases:
| Phase | Description |
| -------------------- | --------------------------------------------------------------------------------- |
| **Active Planning** | Signals indicating a student may be actively planning self-harm or harm to others |
| **Suicide Ideation** | Signals indicating a student may be experiencing thoughts of suicide |
| **Self Harm** | Signals indicating a student may be engaging in or researching self-harm |
| **Help & Support** | Signals indicating a student may be seeking help or support resources |
| **Suicide Research** | Signals indicating a student may be researching suicide-related topics |
Notifications for Active Planning alerts are required. Notifications for the remaining four phases are optional and can be configured by a SuperUser or SuperCounselor.
## Resources
Overview of what Beacon is and how schools use it.
Definitions for key terms used throughout Beacon documentation.
Create a Beacon deployment and connect it to an Organizational Unit.
Step-by-step checklist for Super Users setting up Beacon Starter.