It sure seems like Amazon wants AI to make more of our media decisions. I think
By Mike Feldman · October 10, 2026 · Curated by George's Blog
It sure seems like Amazon wants AI to make more of our media decisions.
I think we should let it.
Look, I get that this will make most people uncomfortable. But that is the exact tension that I found most interesting in Lauren Johnson's latest ADWEEK piece on Amazon’s increasingly agentic advertising platform.
The benefits are obvious at this point. Less time pressing buttons for teams, faster decisions for leaders and the ability to process more data and signals than a human being ever possibly could.
But the important question isn’t what stays human controlled vs. AI. It's really about who sets the objective. What data can the agent see? What is it allowed to do? And can you understand why it did it?
I’m very bullish on agents operating more of the machinery. We’re already building our commerce products to be agent-ready, with MCPs allowing agents to connect into and operate the underlying tools.
But I keep coming back to the same thing: Agentic is not a KPI.
Giving an agent more control only matters if it makes the decision better. Yes agentic buying is very important (and necessary) but the human elements are knowing what to delegate, what to control and how to learn from every decision, whether you are the one pushing the button or not.
Really enjoyed talking through this one with Lauren and appreciate the feature!
Full Adweek piece in the comments 👇