Cisco announced identity controls, adversarial testing tools, and SOC automation built specifically for AI agents - and the security gaps they're already creating.
As AI agents take on more autonomous work, CrowdStrike is pushing security to the endpoint, where that activity actually runs, signaling a shift in how teams monitor and govern AI across environments.
New capabilities around red teaming, runtime monitoring, and coding agent protection give MSSPs a way to manage AI risk within existing SOC workflows while building new managed service offerings.
Entro’s latest update focuses on a basic problem many security teams still can’t solve - who has access, how it’s being used, and whether it should exist at all.