← Back to Blog
Security11/15/2025

Zero Trust Insurance: A Modular Deep Dive into Bob.so Security V3

Bob AI Team

Insurance AI Specialist

Zero Trust Insurance: A Modular Deep Dive into Security V3

1. Introduction: The High Stakes of Insurance PII

In the insurance industry, data isn't just an asset—it's a liability. Personally Identifiable Information (PII), from social security numbers to medical histories, requires a security posture that exceeds standard SaaS benchmarks.

2. Executive Overview: Security V3 Hardening

Bob.so Security V3 is a multi-layer defense-in-depth architecture. It replaces legacy per-server firewalls with a "Zero Trust" model where every byte of data is validated, encrypted, and partitioned at the database level.

3. Detailed Breakdown: RLS, Encryption, and Identity

Row-Level Security (RLS) reasoning

Unlike traditional applications that rely on application-level middleware for auth, Bob.so implements Supabase RLS. This moves security to the database layer:

  • Tenant Isolation: Broker A can never see Broker B's data, even if they share the same API endpoint.
  • Atomic Policies: Every SQL query is checked against the user's JWT (JSON Web Token), ensuring that sensitive policy data is only accessible to authorized individuals.

Reasoning: Modular Hardening vs. Perimeter Defense

The reasoning for this choice is simple: perimeters fail. By making the data itself "aware" of who owns it, we ensure that a breach in the frontend can never lead to a mass data leak in the backend.

4. Implementation Analysis: SOC2 and PCI Compliance

Our architecture is built to be SOC2 and PCI compliant from day one. We utilize Clerk for session-aware identity management and Stripe for encrypted payment handling, ensuring that Bob.so never stores raw financial or sensitive identity credentials on its own servers.

5. Conclusion: Compliance as a Product Feature

In the Enterprise world, security is the ultimate product feature. Bob.so Security V3 provides brokers with the peace of mind to scale their digital agency without compromising their clients' most sensitive data.