DataDome secures MCP infrastructure for trusted agentic AI
DataDome announced a new capability to secure Model Context Protocol (MCP) server infrastructure. With a standard DataDome integration module, enterprises can now gain visibility into and protect MCP interactions, enabling them to build agentic customer experiences on a foundation of trust and security.
Securing MCP servers for trusted Agentic AI
Forrester reports that 36% of U.S. consumers are interested in using AI agents to transact online.
Meanwhile, businesses see the opportunity to revolutionize e-commerce experiences for consumers. In response, enterprises are racing to deploy MCP servers, which act as the switchboard that connects users’ AI agents to business applications and data.
This year alone, more than 16,000 MCP servers have been deployed across Fortune 500 companies. Yet MCP was not designed with security in mind; each deployed MCP server creates an incremental attack surface.
“Enterprises want the growth agentic AI offers, but not at the expense of unknown business risk,” said Benjamin Fabre, CEO at DataDome. “They need fast, simple protections for this new attack surface and a way to establish trust on every agentic interaction with these MCP servers.”
How DataDome protects MCP servers
DataDome provides insights into agentic AI traffic requests and continuous protection against specific threats that may disrupt or exhaust MCP server resources. Assessing the origin, intent, and behavior of AI agents acting on behalf of consumers, DataDome helps secure agentic commerce and turn buyer-initiated, agent-mediated actions into trusted business growth.
Customers can integrate DataDome into their MCP server infrastructure using a standard MCP-enabled server-side module, starting with Node.js and AWS CloudFront’s Lambda@Edge, with additional modules coming later this quarter.
This protection is delivered through three core capabilities:
- Real-time visibility into agentic traffic: Insight into the identity and intent of AI agents, with multi-layered detection that classifies and controls interactions, without blocking legitimate users.
- Autopilot protection: Malicious agentic threats are blocked at the edge before they reach MCP servers. Protection is adaptive, accurate, and continuous, ensuring operations stay secure without disruption.
- Continuous agentic AI trust relationships: By verifying identity, intent, and behavior for every interaction, human, bot, or AI agent, enterprises gain transparent decisioning, confidence, and control to enable safe, scalable agentic commerce.