The agent security gap: 54% of enterprises have already had an AI agent incident, and most still let agents share credentials

A survey of 107 enterprises reveals that 54% have experienced AI agent security incidents or near-misses, with most agents sharing credentials rather than having scoped identities. Only 30% isolate high-risk agents, and security controls are mostly provider-native rather than purpose-built. Despite high satisfaction with current tools, most organizations plan to change their agent security tooling within a year due to rising risks.

AI agent security gaps expose enterprises to significant risks from credential sharing and insufficient isolation controls.

Teams tracking AI Agents, Agents, product strategy, operations, and market positioning.

Organizations should adopt scoped identities and isolate high-risk AI agents to reduce security vulnerabilities.