Palo Alto Networks’ State of Cloud Security Report 2026, published on 17 November, finds that AI agents are now embedded in development pipelines at almost every organisation surveyed. Some 99% say they use agents for autonomous or semi-autonomous actions, and 82% say AI-driven development has increased the volume of cloud updates.
Governance has not kept up. Nearly three-quarters allow agents to change code, CI/CD configuration or production, yet 40% have been unable to reconstruct an agent’s actions, rising to 46% in financial services. More than half (53%) log agent activity under a developer’s own identity or token, 43% found agents running without security approval in the past year, and revoking credentials takes 2.5 days on average. These are vendor survey results.
Why it matters for IT teams
Give every agent its own identity, scope its permissions tightly, and log its actions separately from human users. If you cannot answer “what did the agent change and when?”, incident response will be slow.

Leave a Reply Cancel reply