What happened
Anthropic introduced the Model Context Protocol (MCP) on November 25, 2024, as an open-source standard for AI agent communication with external data sources. MCP defines how agents interact with systems, elicit user input, and utilise automated tools. Unlike APIs, designed for deterministic developer interactions, MCP abstracts functionality for non-deterministic AI agents. By February 2026, over 6,400 MCP servers registered, with OpenAI adding support to ChatGPT in March 2025 and Google following in April 2025.
Why it matters
MCP's rapid adoption shifts how platform engineers and architects design agentic workflows, moving from rigid API integrations to flexible, agent-driven tool utilisation. This standardisation reduces the complexity of connecting AI models to diverse external systems, enabling agents to perform complex, multi-step tasks by abstracting functionality rather than specific API endpoints. For security architects, the protocol's emphasis on guardrails around tools becomes critical, as increased agent autonomy necessitates robust controls to manage non-deterministic execution and prevent unintended actions.
Subscribe for Weekly Updates
Stay ahead with our weekly AI and tech briefings, delivered every Tuesday.
