The Gartner Guardian Agent category for software that keeps an AI agent's outcomes aligned with what the business intended.
Gartner defines it as startups focused on business goals and aligning agent outcomes with the organization's intent. It does not build or run the agent. It watches the gap between what the agent was built to do and what it is doing in production, across every failure mode: drift, hallucination, policy violations, bad escalation. AvonAI was named in this category.
Related: AI Agent Drift, AI Agent Hallucination, Guardian Agent, AI Agent Manager