We recently evaluated Palo Alto's AI Access Security for Prisma. It performed reasonably well and was more affordable than expected, but because it's deployed at the firewall, it mainly helps users connected to our corporate network. We have many remote employees who aren't on the LAN. While we could require them to use a VPN and route all traffic through it, I'm interested in how others handle AI usage security for remote users. Are agent-based or endpoint-based products worth considering?
2 Answers
Legit Security has a product called Vibeguard. It’s still fairly early and beta-level, but it may be worth evaluating depending on whether you’re trying to monitor general AI usage or a more specific workflow.
I’d be cautious about putting an AI system in charge of security decisions by itself. These tools can be useful for monitoring or enforcing narrowly defined policies, but they still need strong rules, validation, and human oversight rather than being trusted to make critical judgments independently.
The name is memorable, but I’m not sure it would go over well with executives when the product is being reviewed. The branding could make it harder to take seriously.