Skip to main content

Overview

At Biel.ai, security is a core principle of our AI chatbot-as-a-service platform, designed to meet enterprise-level requirements with robust features that ensure data protection, compliance, and transparency.

This document details how we protect sensitive information, enforce security controls, and comply with industry standards.

Security framework​

Our security approach is grounded in four key principles:

  1. Confidentiality: User and system data must remain protected from unauthorized access.
  2. Integrity: Data must be accurate and reliable, ensuring no unauthorized alterations.
  3. Availability: Security measures should not compromise the platform's accessibility or performance.
  4. Compliance & trust: Security is effective only when it aligns with regulations and builds trust with users.

These principles guide every decision, ensuring the platform remains secure and user data is safeguarded.

Key security features​

Biel.ai offers several configurable security features that allow users to customize protection based on their needs:

  1. Role-based authentication (RBAC): Restricts admin panel access to authorized users, providing a segmented permission structure.

  2. Domain restrictions: Each project has a unique Project ID to limit where the chatbot widget can be embedded. Only authorized domains can interact with the widget, ensuring controlled and secure usage.

  3. Rate limiting: To protect the system from abuse, rate limits are applied per user session and IP address. This helps prevent spam, DDoS attacks, and unauthorized interactions.

  4. PII removal: Projects can enable automatic detection and removal of sensitive information (e.g., names, emails, phone numbers) from chat conversations, enhancing user privacy.

  5. Public API security: Enforces Bearer Authentication for secure communication between clients and servers, preventing unauthorized access to API resources.

Security architecture​

Beyond configurable settings, Biel.ai’s security architecture delivers comprehensive protections at the data, network, and access layers:

  1. Network protection: Firewalls, DDoS mitigation, and continuous monitoring help identify and mitigate threats. Logs capture security events and anomalies, providing real-time alerts and audit trails for proactive management.

  2. Data protection and resilience: Sensitive data is encrypted both in transit and at rest. Regular automated backups ensure resilience, allowing for quick recovery in the event of failures or cyber incidents.

  3. Data integrity and auditability: We maintain detailed audit logs of security-sensitive operations to prevent unauthorized modifications. This provides transparency and ensures accountability by tracking system interactions.

  4. Logical account separation: Customer data is strictly isolated to ensure that each account remains separate, preventing unauthorized access and ensuring that resources are only available to the rightful owner.

  5. Opt-out for model training: Biel.ai does not share customer data with third-party training programs (e.g., OpenAI, Anthropic), ensuring that chat interactions remain private.

  6. Regulatory compliance: We adhere to GDPR to ensure user information is handled with transparency and in line with legal requirements.

By combining user-configurable security features with a robust, enterprise-grade security architecture, Biel.ai provides a multi-layered defense that ensures both flexibility and high standards of protection.

For more details on our data handling practices, see our Privacy Policy.