Who created MCP and is it open?

Anthropic introduced it in November 2024 as an open standard; it has since been adopted by other major AI providers, becoming a de facto standard.

Why would a marketer care about MCP?

It lets an AI agent securely connect to your specific tools and data through a standard interface, instead of being limited to hard-coded integrations.

Explainer

What is MCP (Model Context Protocol)? A plain-English guide for marketers

TL;DR

The Model Context Protocol (MCP) is an open standard introduced by Anthropic in November 2024 that standardizes how AI systems connect to external tools and data. Instead of building a custom integration per service, you expose tools, resources and prompts through one protocol — so an AI agent can securely use your ad platforms, BigQuery and internal systems.

What is the Model Context Protocol?

MCP is an open standard, introduced by Anthropic in November 2024, for connecting AI systems to external tools, systems and data through a single, secure, two-way interface. It has been adopted broadly, including by other major AI providers, making it a de facto standard.

The problem MCP solves

Without a standard, every new data source needs its own custom integration — an "N×M" problem that doesn't scale as you add models and tools. MCP replaces that with one protocol many clients and servers can speak.

How MCP works

A host runs the AI model; clients connect to servers.
The client asks a server for its capabilities; the server replies with natural-language descriptions; results flow back into the model's context.
Servers expose three primitives: Tools (executable functions), Resources (file-like data), and Prompts (templates).

MCP vs a plain API

An API is a custom integration you build per service. MCP is a standard way to expose tools and data to any AI model — so one server works with many AI clients, instead of re-integrating each time.

Why it matters for marketing and ad ops

MCP lets an AI agent securely connect to your specific tools and data — ad platforms, BigQuery, internal systems — through a standard interface, instead of being limited to whatever is hard-coded. That means your agent can use your workflows and your data.

Agent Planners supports custom MCP servers, so teams can plug in their own data sources and actions.

Frequently asked questions

Is MCP the same as an API?: Not quite. An API is a custom integration you build per service; MCP is a standard way to expose tools and data to any AI model, so one server works with many AI clients.
Who created MCP and is it open?: Anthropic introduced it in November 2024 as an open standard; it has since been adopted by other major AI providers, becoming a de facto standard.
Why would a marketer care about MCP?: It lets an AI agent securely connect to your specific tools and data through a standard interface, instead of being limited to hard-coded integrations.