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

Research indicates that over half of enterprises have experienced a confirmed security incident involving AI agents, highlighting a significant security gap. Many organizations lack adequate control over agent identity, with only about a third providing each agent with a scoped identity, while others share credentials among agents. The risks are escalating due to insufficient controls, exposing enterprises to potential attacks, while spending on security remains a small fraction of the overall budget.