Why did Amazon and Shopify respond so differently to AI agents?

Tanay Jaipuria argues that Shopify and Amazon reacted to Meta's Muse — a personal AI agent that researches products and buys on a customer's behalf — in opposite ways because the two companies' business models expose them differently to agentic purchasing. "Shopify is making its merchants accessible through the agent. Meanwhile, Amazon has blocked it." Jaipuria writes that Amazon "has cited agent identification and credential access as reasons for blocking Muse," but adds, "it's clear they aren't the whole story." He points to two protected interests: Amazon's position as the default place to shop for routine purchases, and its advertising business, which he estimates generated "roughly $69B in advertising revenue in 2025," most of it from sponsored-product listings that require humans to actually visit the site. Shopify, by contrast, earns through "Merchant Solutions, which represents 76% of Shopify revenue, primarily driven by payments transacted on the storefronts of its Merchants" — a business that grows if merchants are easier to find and buy from, agent or not. Jaipuria cites Shopify's Harley Finkelstein saying on an earnings call that the agentic plan was "opening conversations with brands that wanted access to AI shopping but weren't ready for a full migration."

The mechanism Jaipuria gives is that it is worth separating "the companies paying for ads from the companies earning money from them." Retailers and marketplaces such as Amazon earn by selling merchants access to human shoppers; if an agent purchases on Amazon without a human ever seeing the search results or sponsored listings, Amazon can still win the transaction and still lose the margin that came from advertising around it. Shopify never depended on owning that attention in the first place — it "doesn't need to own every interaction leading to a purchase and can provide the catalog, checkout and payments layer" instead, charging merchants the same processing fees whether a human or an agent checks out.

This adds a prior, platform-level gate to the loyalty question raised in Can an AI agent serve both merchant and user interests fairly?: before an agent's own referral-fee incentives can create a conflict of interest, the host platform first decides whether to admit the agent at all, and Jaipuria's account suggests that decision turns on how much of the platform's own revenue rides on human eyeballs rather than on the transaction itself. It also gives a concrete case of the dynamic in Why do consumers delay adopting agentic search?: Amazon's blocking reads as exactly this kind of lag, a platform slowing a shift that may be useful to consumers because it is not yet revenue-superior for the platform. And it is the flip side of Will brands succeed by building their own AI shopping agents? — Shopify's strategy is to make its merchants the agent's default choice by being "the catalog, checkout and payments layer," rather than by competing for attention the old way.

The piece is a single analyst's newsletter essay, not a study, and its figures — the $69B ad-revenue estimate and the 76% Merchant Solutions share — are Jaipuria's own figures, not independently audited. The argument covers one dispute (Muse, Shopify, Amazon) and a short list of other platforms that chose to participate (Walmart, Best Buy, Sephora, Instacart, Expedia, DoorDash, PayPal) without reporting outcomes for any of them, so it does not establish how widely the admit-or-block pattern holds; Jaipuria frames his own prediction — that "advertising on shopping surfaces is particularly exposed if agents start doing more of the purchasing" — explicitly as a view ("I think"), not a finding.