Can We Control Every AI Agent Before It Becomes Our Next Privileged Insider?
Check Point, Thursday, September 24th, 2026
Check Point asks whether enterprises can govern AI agents before their autonomy turns them into privileged-insider risks.
Check Point argues that AI agents connected to systems like Microsoft 365, Salesforce, ServiceNow, and internal databases can read data, make decisions, call tools, and change systems without individual approval, creating insider-like risk when access and autonomy outpace oversight.
It contrasts traditional identity and access management, which answers what an identity is allowed to access, with the agentic question of whether an agent should perform a particular action in a given context right now - for example, read access to a customer database shouldn't automatically mean an agent can delete records. The post calls on security and business leaders to identify, govern, monitor, and control every AI agent in their environment.