AI isn’t staying in the browser and neither is enterprise AI security. Cato Networks launched Cato AI Endpoint, a new tool designed to police AI use across browsers, desktop applications and locally running agents without replacing existing security tools.
Companies need to know not just which AI tools employees use, but what information those tools touch and what their agents do. Cato argued browser extensions alone leave gaps as AI activity spreads into desktop apps and local development environments.
Cato's new tool discovers AI use, inspects prompts and responses inline and enforces policies directly on the endpoint. The company said it can uncover unapproved tools and agents, flag risky agent activity and stop sensitive data from leaving the organization.
It also links users, devices, data, prompts, responses, agent activity and policy decisions to give security teams context around AI interactions. Cato said that visibility can expose risky agents, Model Context Protocol servers and skills before they act.
For enterprises, the proposed draw is tighter AI governance without a complete security overhaul.
Read the full press release here.