MCP (Model Context Protocol) is a new buzzword in the AI community, introduced by Anthropic in November 2024, and has recently gained traction with announcements from OpenAI and Google. MCPs are experimental protocols that enable standardized interactions between AI models and external tools, systems, and data sources, allowing for more scalable integrations and democratizing access to automation. The protocol consists of three primary components: Host, Client, and Server, which work together to facilitate seamless collaboration between AI agents and tools. While security is not the primary hurdle for enterprises, governance is a growing concern as AI agents become increasingly autonomous and probabilistic. Enterprises need frameworks, observability tooling, and abstractions to trust what agents are doing without micromanaging them every step of the way. MCP has the potential to shift the default from bespoke integration to plug-in systems, enabling more flexible and scalable workflows.