> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fieloloyalty.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Security

Protecting the integrity of your loyalty program data is our highest priority. Fielo Loyalty employs a multi-layered security strategy to ensure that all data, including member and transactional data, remains secure.

### Infrastructure

As an application built on a trusted third-party cloud, we inherit enterprise-grade security controls from a platform that is **zero-trust architected** and **secure by default**.

* **Physical Security**: Data is hosted in secure, Tier-1 cloud environments with 24/7 monitoring and physical access controls.
* **Network Isolation**: We utilize Virtual Private Clouds (VPCs) and advanced firewalls to isolate program data from unauthorized external access.
* **Compliance**: Our third-party infrastructure enjoys over 20 security certifications, including **ISO 27001** and **SOC 2**.

### Data Protection

We employ industry-standard encryption protocols to safeguard sensitive information throughout its lifecycle.

* **Encryption in Transit**: All data moving between external systems of record and our API is encrypted using TLS 1.2+.
* **Encryption at Rest**: Databases storing member attributes, external IDs, and transaction histories are encrypted using AES-256.
* **Role-Based Access Control (RBAC)**: Administratively, you can manage who has the authority to view each area of the app, ensuring the principle of least privilege.

### AI

When leveraging Large Language Models (LLMs) to generate insights and recommendations, we adhere to strict security protocols:

* **No Training Policy**: We use enterprise-tier API agreements with LLM providers. Your proprietary program data and member information are *never* used to train the provider's foundation models.
* **No Data Logging**: LLM processing occurs within a secure, ephemeral environment. Data is sent for the specific task and is not stored by the LLM provider beyond the immediate session (Zero Data Retention where applicable).
* **Prompt Security**: When sending a prompt to a public LLM, content filtering applies to detect and prevent harmful content in prompts and completions.
* **Consistent Performance**: With every release, AI models are verified to ensure consistent performance and efficiency.