Amazon Ads has opened its Model Context Protocol (MCP) Server in beta, allowing AI agents from platforms such as Claude, ChatGPT and Gemini to connect directly to Amazon advertising accounts and manage campaigns through natural language prompts.
The Amazon Ads MCP Server acts as a translation layer between AI agents and the Amazon Ads API. Instead of building custom integrations for each tool, advertisers can use a single connection. Agents can then create, update or delete campaigns, run performance and reporting queries, manage account settings and access billing data.
The system includes pre-built workflows that reduce multi-step tasks to single prompts. One example is expanding campaigns from the U.S. and Canada to another country, or launching a full Sponsored Products campaign in one step. Paula Despins, Vice President of Ads Measurement at Amazon Ads, described the server as a way to turn complex advertising operations into simpler actions for agents.
The MCP Server builds on Amazon’s earlier Ads Agent tools and is available to Amazon Ads partners with active API credentials. It supports connections from custom-built agents as well as major AI platforms. Amazon has noted that it does not own or operate the AI models that connect through the server.
The move reflects a broader industry shift toward agentic advertising, where AI systems handle execution and optimization while humans set objectives and constraints. By standardizing how agents talk to its platform, Amazon is positioning its advertising infrastructure for more automated, agent-driven workflows.
Source: Amazon Ads
