LE PERSPECTIVE / THE STORY

Bay has emerged from stealth with $50 million raised to secure autonomous AI agents as they work across corporate systems. The startup says some Fortune 100 customers tripled their use of autonomous agents after adopting its platform.

The risk is not always a hacked agent. Software with legitimate access can still make an unsafe decision, such as exposing a cloud resource while completing an assigned task.

The real risk is the action
An agent may have valid permissions and still do the wrong thing. Bay checks the proposed action in context before it happens, rather than relying on prompt filtering alone.

Bay was founded by CEO Yuval Zacharia, chief product officer Aner Morag and R&D chief Boris Danilovich. Its backers include Battery Ventures, YL Ventures, Striker Venture Partners, Microsoft's M12 and Micron Ventures.

The platform discovers agents, MCP servers and extensions across devices and cloud workloads. Its Lighthouse system turns prior activity into tailored policies; enforcement can allow, seek approval, steer or block individual actions.

Bay says companies can deploy through existing device-management or endpoint-security tools in under five minutes. Its policy system can also replay 30 days of agent activity to test restrictions before enforcement begins.

The threefold adoption figure is company-reported. The larger question is whether stronger controls can make wider agent autonomy safe enough for enterprise use.

Get stories like this in your inbox every week. Subscribe to The Week in Perspective.