AI agents aren’t just talking, they’re taking actions. They’re booking transactions, pulling sensitive data, and chaining tools together to get work done. As enterprises embrace these agents, protocols like the Model Context Protocol (MCP) and Agent-to-Agent (A2A) are enabling powerful new capabilities, but also creating invisible ecosystems of API-driven agent activity that traditional security tools can’t see or control.
In this session, we’ll walk through the newest Salt platform updates, including MCP Protect for complete visibility into Model Context Protocol traffic and out-of-the-box agentic AI policies enabled by default. You’ll learn how to identify blind spots, strengthen governance, and reduce risk, starting from first login.
You’ll learn how to:
- Understand today’s risks: New Salt research shows that only 37% of organizations using agentic AI have a dedicated API security solution, and 48% run 6–20 agent types, expanding the attack surface.
- Uncover hidden traffic: MCP Protect provides end-to-end visibility into agent toolchains by discovering and cataloging all MCP servers and MCP-exposed APIs, surfacing blind spots, and accelerating remediation.
- Automate governance: New Agentic AI policies deliver out-of-the-box security with seven pre-enabled checks that automatically detect the highest-risk exposures in agent-driven environments.
Don’t let hidden agent traffic put your AI initiatives at risk. See how Salt turns AI agents from a security gap into a competitive advantage.
