By Karsten Weide, Chief Analyst
On February 2, 2026, Amazon Advertising officially announced the open beta test for its Model Context Protocol (MCP) server at the IAB’s Annual Leadership Meeting. With this new product, Amazon is taking an important step toward opening its advertising platform to AI agents. And because of Amazon’s size, this also energizes the efforts of the entire industry to make digital advertising simpler, faster, better and cheaper via artificial intelligence and AI agents.
Amazon Advertising’s MCP server represents a step change for its advertising business, serving as a major milestone that allows large-language models like Gemini, ChatGPT, or Claude to speak “Amazon Ads” fluently via a universal translation layer. By adopting this open standard originally developed by Anthropic, Amazon makes sure it stays among the leaders in the race for autonomous advertising software, ensuring it attracts more advertiser marketing dollars in the AI era by eliminating labor-intensive integrations, making its APIs significantly easier to use, and enhancing campaign effectiveness.
For advertisers, the biggest win is operational velocity. 70-80% reductions in the time needed to set up and run ad campaigns after adopting AI agents are not uncommon. Traditionally, launching a campaign or pulling complex reports meant navigating a maze of spreadsheets and manual console entries – a process that can take hours and days. Through natural language, you can now prompt an AI agent to “create a Sponsored Products campaign for my new ergonomic office chair with a daily budget of $5,000, automatic targeting, and an ‘up and down’ bidding strategy”, and the MCP server converts that text into executable API calls.
For bigger buyers, these lower technical barriers allow them to simplify operations and reassign ad ops personnel. But for many small-to-medium businesses with limited resources -a major target client group for Amazon – this means they can now run complex campaigns in the first place.
Second advertiser benefit after increased efficiency is improved campaign effectiveness due to AI making better informed campaign decisions than humans could. Here, the first numbers are all over the place. I’ve seen KPI improvements as low as 5-10% and as high as 250%, with the belly of the curve probably somewhere in the double-digit percentages.
For AdTech vendors, the MCP server is both a powerful tool and a competitive wake-up call that underlines the necessity to either adopt agentic AI or fall by the wayside. Vendors like PubMatic and Viant, and publishers like Yahoo show the way.
For publishers, the rise of MCP servers could herald a shift toward more direct-to-buyer sales. While the industry’s current focus is on the buy-side, the ripple effects for publishers are coming. The protocol allows for reduced friction, and as AI agents become capable of negotiating and activating deals, publishers may see more direct programmatic buys that bypass the traditional, fragmented supply chain. Already, publishers have begun to develop sales agents. Additionally, publishers may benefit indirectly from smarter demand, seeing higher-quality results and more efficient targeting as advertisers use these better-performing automation tools.
The Amazon Ads MCP server is a DSP-based buy-side tool designed to manage and optimize campaigns across the full breadth of the Amazon Ads ecosystem, which includes both Amazon’s owned-and-operated (OOP) inventory and that of its network of publishing partners.
The Amazon Ads MCP Server is available globally in open beta to all Amazon Ads partners who hold active API credentials. No date has been announced for the general availability of Amazon’s MCP server.
Leave a Reply