Lema Forensic AI Assessment is a third-party risk management platform that uses AI agents to conduct forensic analysis of vendor security documentation and artifacts. The platform automates vendor risk assessments by analyzing security documents, policies, and publicly available information to identify material risks and contradictions. The system integrates into existing workflows to capture vendor scope and validates it against known vendor profiles to detect misclassifications. It performs automated evidence collection by fetching public artifacts, conducting gap analysis against security controls, and requesting only missing documentation from vendors. The platform includes OSINT reconnaissance capabilities to uncover undisclosed information such as data breaches, organizational changes, and hidden risks that may contradict vendor-provided reports. The forensic engine cross-references thousands of pages of documentation to validate vendor claims and identify inconsistencies. Users can define custom security controls using natural language without requiring prompt engineering. The platform applies these controls to generate findings with cited evidence from source documents. The system provides continuous monitoring capabilities to detect changes in vendor risk posture over time, including new sub-processors or policy modifications. Assessment results include audit-ready evidence and actionable intelligence for risk mitigation. The platform reduces assessment time from weeks to minutes while increasing the number of vendors that can be actively monitored.