Skip to content
ATDbbeta

Model Context Protocol (MCP) in advertising

An open standard for connecting AI applications to outside data and tools — now used by ad-industry protocols so AI agents can plan and buy media.

The Model Context Protocol (MCP) is an open-source standard for connecting AI applications to external systems: data sources such as files and databases, tools, and workflows. Its own site compares it to a USB-C port — one standard way to plug AI applications into other systems.

Where it came from. Anthropic open-sourced MCP on 25 November 2024, describing it as "a new standard for connecting AI assistants to the systems where data lives." The aim was to replace one-off integrations for every data source with a single protocol.

How it works. Developers either expose their data or services through an MCP server, or build AI applications — MCP clients — that connect to those servers.

The advertising angle. The Ad Context Protocol (AdCP), an open standard published by AgenticAdvertising.org, is built on MCP and the Agent-to-Agent protocol (A2A). It defines tasks and schemas that let AI agents representing brands, agencies, publishers and ad platforms discover media products, negotiate and execute media buys, manage creative, enforce brand safety and report on campaign performance. MCP is its transport for tool-calling, where an AI agent invokes AdCP tasks as tools; DSPs, SSPs, publishers and other platforms can expose their capabilities as AdCP tasks over MCP or A2A.

Standards bodies. IAB Tech Lab's site describes an MCP server reference implementation for its Deals API v1.0, donated by HyperMindZ to Tech Lab's Agentic Advertising Management Protocols (AAMP) initiative. It lets AI agents run a programmatic deal's lifecycle — create, send to provider, confirm, pause, resume and audit — through ten MCP tools.

In ATDb

Sources

Written with AI assistance from the sources below and reviewed on . Spot an error? How we check facts.