Security control plane for discovering AI agents and detecting risk across the agentic stack

Oryo was founded in 2025 in San Mateo, California, by chief executive Prashanth Ram and chief product officer Swaroop Sham, who between them worked at Okta, Wiz, and Proofpoint on identity, token-based access, and security orchestration. Its investors include Village Global, Seven Hill Ventures, and Golden Sparrow. The problem it addresses is one that identity teams recognize from the early SaaS era repeating itself: business units are switching on agents inside Salesforce, Microsoft, and cloud AI platforms faster than security can enumerate them, each agent inherits a set of permissions and connections nobody reviewed, and there is no single place to see the resulting exposure.

The platform connects to SaaS applications, cloud providers, and identity systems and pulls agent configurations, entitlements, activity logs, model and prompt details, and Model Context Protocol connections, then layers business context and third-party threat intelligence over them. From that it builds a map of every agent, what data moves through it, and which tools it can reach, and applies agent-native detections for tool misuse, missing prompt safeguards, privilege escalation, and paths that would allow data exfiltration. Findings are risk-scored and prioritized rather than dumped as a list, least-privilege enforcement acts on the over-permissioned cases, and alerts route into the SIEM, ticketing, and SOAR systems a security team already runs. It also tracks AI supply chain risk and produces compliance and governance reporting.