WitnessAI has announced extended agentic security capabilities that govern how AI agents interact with enterprise systems, tools, and Model Context Protocol (MCP) servers. With the launch of Agentic Control, enterprises have greater visibility and control over their AI agents with a single control plane to discover, monitor, govern, and restrict agent behaviors at runtime.

Enterprises are deploying AI agents across chat applications, integrated development environments (IDE), and custom workflows, but security teams lack the visibility to monitor or control their access. This shift introduces severe operational risks: compromised agents can now independently execute tools, access sensitive systems, and exfiltrate data. As most legacy security technologies were not designed to inspect MCP communications, tool invocations, or agent-to-agent workflows, organizations are left exposed.

WitnessAI Agentic Control bridges this gap by delivering deep visibility and real-time network enforcement over the tools and MCP servers AI agents can access. By establishing a single, organization-wide approved-tool policy, enterprises gain consistent runtime governance across approved agentic environments. This unified control plane secures both human and agentic AI, providing a comprehensive audit trail for safe, compliant adoption.

WitnessAI Agentic Control delivers:

• Agent discovery and visibility: The platform automatically discovers AI agents operating across enterprise environments including IDEs, applications, agent frameworks, and custom agents built in the public cloud, and identifies the specific MCP servers, tools, and downstream systems each agent reaches. WitnessAI’s new MCP Catalog scores known tools against OWASP and CVE risk classes, enabling security teams to assess risk before granting approval.
• Approved-tool and MCP server governance: WitnessAI enables organizations to define organization-wide allow lists of approved MCP servers and tools that can be consistently enforced across agentic environments.This allows for consistent enforcement across all applications, model providers, and custom-built agents.
• Runtime agentic enforcement: The platform inspects and governs conversations in agentic applications including IDEs, enabling organizations to restrict unauthorized prompts and responses and enforce approved policies at runtime. Combined with WitnessAI’s AI Firewall capabilities, organizations can also apply prompt injection, jailbreak, and response protections for AI applications and agents.

“Enterprises are moving fast to deploy AI agents that can code, access internal data, and execute complex workflows. However, security teams cannot protect what they cannot see, let alone control,” said Rick Caccia, CEO at WitnessAI. “Most AI security vendors hand the buyer a choice: govern employees, govern apps, or govern agents. WitnessAI removes that choice. By extending the platform our customers already trust to govern employee AI usage, we are providing a single control plane to protect all AI activity. A CISO can write a rule once, and it holds across every human user, IDE, chat application, and custom agent.”

The WitnessAI platform delivers a 99.3% true positive rate on employee AI guardrails, extending the same platform foundation that governs AI usage across more than 4,000 AI applications and over 100 supported model types.