Augustine Alfred

BreachGraph AI, built by GreenLanTech, is an AI-powered cloud identity security investigation platform designed to help security teams discover excessive permissions, understand hidden attack paths and validate security fixes before applying them. In cloud environments, a single overprivileged identity can create a chain of access that allows an attacker to move from a low-privilege account to sensitive systems and data. These relationships are often difficult to identify through traditional permission reviews. BreachGraph AI analyses cloud identity and permission data, identifies risky access relationships, and visualises potential attack paths. Its deterministic Python security engine provides the underlying evidence, while Meta's AI models help explain findings and suggest least-privilege remediation. Its key differentiator is remediation simulation. Instead of simply recommending permission changes, the platform simulates proposed changes and compares the attack paths and security risk before and after remediation. This helps teams understand whether a fix reduces risk while preserving legitimate access. The project demonstrates how AI can make cloud identity security investigations more understandable, actionable and verifiable.