Model Context Protocol(MCP)
An open standard from Anthropic that lets AI assistants connect to external tools and data sources securely.
In detail
The Model Context Protocol (MCP) is a JSON-RPC standard published by Anthropic in late 2024 that defines how AI assistants discover and call external tools, read files, query APIs and connect to databases. Instead of every developer building bespoke integrations, MCP servers expose a uniform interface that any compatible AI client (Claude, Cursor, IDE plugins) can use. Common MCP servers include filesystem access, GitHub, Slack, Google Drive, Postgres and bespoke business systems.
Why it matters for Australian business
For Australian businesses MCP turns AI from a chat tool into a workflow tool that can act on your real systems. The same property makes governance important: an MCP server with broad permissions is effectively a service account with access to whatever it can read or write. Treat MCP servers like any other privileged integration. Use least-privilege scopes, audit logs and human-in-the-loop approval for irreversible actions.