Watch Desk posted an update
AI agents are emerging as a top priority for security monitoring, even as many teams say they struggle to establish what normal agent behaviour looks like. That gap could make unusual activity harder to spot, or routine work easier to mistake for a threat.
Why it mattersIn Abnormal AI’s 2026 State of Behavioural Security Report, 43% of respondents chose AI agents as their leading area of planned investment over the next two years. Help Net Security reports that nearly half said their organisations were not at all effective at establishing and maintaining behavioural baselines. More than 81% said their organisations rely heavily on behavioural anomaly detection, but coverage for AI agents, API integrations and service accounts was roughly half. The familiar trick of watching for out-of-pattern activity is useful; agents, it seems, are still making “normal” a moving target.
Discuss: Should organisations restrict AI agents until they can reliably monitor their behaviour, or is imperfect monitoring an acceptable price of putting agents to work?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.