According to a report by Abnormal AI, 46% of security professionals say they are not at all effective at establishing what normal behavior looks like for AI agents, even as 43% say agents will receive their greatest behavioral-monitoring investment over the next 24 months, more than three times that of API integrations (14%).

Behavioral security uses AI to learn what’s normal for each identity, account, and system, then flags activity that deviates, rather than relying on known threat signatures. Without a clear picture of normal behavior, a compromised or misbehaving agent can look like any other. More than 81% of respondents now rely on behavioral anomaly detection to a major or extensive degree, but that maturity hasn’t yet extended to AI agents or machine identities.