Onyx
Onyx builds a control plane for autonomous AI, giving enterprises visibility, governance and operational control over AI agents through five capabilities spanning observability, security, governance, orchestration and ROI.
Onyx is a security company that builds what it describes as a control plane for autonomous AI. Its platform is designed to give enterprises visibility, governance and operational control over AI agents as they are adopted and scaled across an organisation. The work sits at the intersection of several technical domains, including AI security, AI governance, AI observability, AI agent orchestration, agent risk management, data exposure prevention, permission management and runtime threat protection.
The product is organised into five core capabilities. AI Observability discovers both sanctioned and shadow AI, with visibility into prompts, responses and agent actions. AI Security enforces real-time safeguards against risks such as data exposure, excessive permissions and remote code execution attacks. AI Governance is intended to help organisations meet AI security standards and regulatory requirements. AI Orchestration reduces the setup friction for AI agents and optimises for cost, accuracy and latency, while AI ROI measures adoption and the business value delivered.
The platform integrates with five leading AI providers: Anthropic, OpenAI, Amazon, Microsoft and Google. Onyx sells primarily to enterprise customers in the cybersecurity and AI governance and risk management sectors, and its buyers are typically cross-functional, spanning security, governance and AI teams. Engineering and product work therefore touches threat detection, permissions, compliance and orchestration problems in equal measure.