Amazon spent the past year blocking AI crawlers from its product pages and suing an AI browser for shopping on its site without permission. Last month, it started paying to advertise inside one of them.

During Prime Day, Amazon bought sponsored placements inside ChatGPT, badge and all, sitting directly below the AI's answer. Its own catalog stayed walled off the entire time. That's not a contradiction Amazon stumbled into, the timing tells you it was deliberate. Adobe found AI chatbot referrals converted 40% better than search, email, or social this Prime Day, the first time AI traffic has out-converted every other channel instead of trailing it.

Here's the part that gets missed in most of the coverage. Agentic AI still drives well under 1% of traffic to major retail sites, and Amazon's own share is the lowest of the group at roughly 0.4%, according to J.P. Morgan. So the volume isn't there yet, not close. What that gap tells us is the real story. Amazon isn't chasing the traffic ChatGPT delivers today, it's buying a seat at the table before the channel gets big enough to threaten the ad business it's built on shoppers landing on Amazon.com first.

For brands, the lesson isn't to panic about ChatGPT ad spend this quarter. It's to recognize what Amazon just told the market with its own money. Waiting for AI-driven shopping to "prove out" before you build for it is watching the wrong number. Amazon already decided the conversion data and the strategic risk were worth paying for now, while the channel is still small enough to shape rather than react to.

We'd treat this as a signal to start now on the things that are cheap today and expensive later, structured product data, clean listing content, and a presence wherever your customers are already asking AI tools for recommendations. The brands that wait for the traffic numbers to justify the investment will be optimizing for a channel that's already crowded by the time they start.