Amazon brings AI agents to its ad stack with new MCP server

Teaching AI agents to manage ad campaigns usually hits the same wall: legacy ad tech and a mess of custom APIs.

On Monday at the IAB Annual Leadership Meeting, Amazon announced the open beta of its Amazon Ads Model Context Protocol (MCP) Server. The tool offers advertisers and ad tech partners a single, standardized bridge to plug AI agents directly into Amazon's advertising software.

Natural language meets ad APIs

Instead of engineering custom integrations for every single workflow, advertisers can now use MCP as a shared middle layer.

The server translates natural language prompts from an agent into structured API calls that Amazon's systems can act on. Paula Despins, VP of ads measurement at Amazon Ads, noted that advertisers can hook their agents in within minutes.

Why it matters

For media buyers and brands, this cuts out weeks of custom integration work. If your team is already experimenting with MCP-compatible agents, managing campaigns or pulling metrics could soon feel as simple as sending a chat message.

We think Amazon adopting an open standard like MCP is a smart strategic play. Instead of trying to lock developers into a proprietary framework, it makes Amazon Ads instantly accessible to the broader ecosystem of autonomous agents.

That said, the stakes here are tied to real ad spend. Letting an autonomous agent execute API calls means a single hallucination could burn through a client's budget, so robust guardrails and human approval loops will still be essential before anyone hands over the keys.