Eric Seufert finds three weak spots in the agentic commerce pitch

EEric Seufert@eric_seufert

This interview could be used for ‘Agentic Commerce is a Mirage’ bingo.

✅ Asserts that agents can optimally monetize with transaction fees (fees are sub-optimal relative to ads at scale.)
✅ Highlights restaurant reservations as a core agentic use case (in-demand restaurants *benefit* from controlling scarce inventory / availability. This is also likely a tiny market, and consumers may not be willing to pay for this…

Sep 29, 2026 · View on X

The bull case got louder this week

In an interview posted by @MollySOShea, Stripe co-founder John Collison said AI agents will completely rewire internet commerce. Stripe processed $1.9T in 2025, up 34%, and is rebuilding for an internet where agents, not people, increasingly transact. His one-line case: "Computer use is the ultimate backwards compatibility layer for the real world."

Some investors see a threat to Amazon. @ChrisCamillo called agentic commerce the biggest new risk to the company, since Meta's Muse and other agents could increasingly control product discovery and purchasing. @joecarlsonshow argued Amazon will have to change direction quickly, because if Amazon blocks Meta, Meta could simply send its users to Walmart and Shopify instead.

The money model planned for Muse is a take rate. @CapexAndChill wrote that Muse is meant to stay free for most users and earn a small cut of the commerce its agent touches, not ad slots.

Seufert's objections start with the money

In late September, @eric_seufert said one interview could serve as a bingo card for the idea that agentic commerce is a mirage. He flagged three assumptions:

• Agents monetize best through transaction fees. He argues fees are sub-optimal next to ads at scale.
• Restaurant reservations are a core use case. Popular restaurants benefit from controlling scarce tables, and the market is likely tiny.
• Retail platforms must give way to all-purpose agents. Ecommerce is concentrated, retailers run their own agents, and the independent agent field is crowded.

His sharpest line closes that third point: "Agents need retail cooperation more than vice versa."

The first objection lands squarely on Muse's take-rate model.

Retailers still hold the door

The two sides overlap more than the noise suggests. Camillo himself expects the shift to play out slowly, with Amazon adapting through its own agents, including Project Moonraker, or through partnerships.

@tongxgalaxy also doubts the fancy-dinner story. The user they say they are building for at Stripe is a busy parent who uploads three back-to-school calendars and lets the agent work out what to buy.

The retailer gap has numbers attached. Billions Network (@billions_ntwk) says its own audit of the 51 largest US online retailers found 37 blocking shopping agents by accident, and none with a way for a verified agent to complete a purchase. The company says the Wall Street Journal mentioned the research. It is a self-run audit. Still, it fits Seufert's point: an agent can't buy where the door is shut.

One shopper reports moving anyway. @leifthunder says a lot of purchases that used to go to Amazon now run through Muse, straight to the brands.

Agents are a channel to watch, not a budget line yet

The core fight is unsettled: a cut of each order, or the ad auction. Among the 51 retailers in that audit, none yet lets a verified agent check out. Store owners can track agent traffic as a channel taking shape and keep funding the ads that pay for this month's orders.

Agents are knocking. Retailers still decide who gets in.