> For the complete documentation index, see [llms.txt](https://verifund.gitbook.io/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://verifund.gitbook.io/docs/integration/guardian-ai-system.md).

# Guardian AI System

The Guardian AI System is an advanced risk analysis tool integrated into the Verifund platform. It leverages machine learning algorithms to perform rigorous evaluations on campaigns, ensuring lowered risks and informed decision-making for all stakeholders.

### System Components

The Guardian AI System is composed of several components working in tandem to deliver high-quality risk assessments:

1. **Data Ingestion Module**
   * Collects campaign metadata, donor history, and external signals
   * Handles data cleaning and normalization
2. **Feature Extraction**
   * Derives meaningful features from raw data
   * Enhances predictive accuracy by focusing on risk indicators

### Workflow

The Guardian AI System follows a structured workflow:

1. **Initial Data Collection**
   * On campaign creation, metadata is collected
   * Continuous monitoring for updated data
2. **Risk Evaluation**
   * Data is parsed and pre-processed
   * Features are extracted for model input
   * Risk score is computed and stored
3. **Review and Actionable Insights**
   * Risk score interpreted by evaluation team
   * Recommendations provided to reduce campaign risks
   * Output shared with campaigns for transparency

### Technical Details

#### Security

* **Data Encryption**: Uses SSL/TLS for data in transit
* **Access Control**: Role-based permissions ensure data confidentiality
* **Audit Trails**: Logs all access and modification events for compliance

***

Continue to User Guide for instructions on using Guardian AI System in campaigning.