The Maze: A shopping assistant can recommend a retailer, then send the buyer to its checkout. Who pays for that introduction? Shopify's current ChatGPT route adds no selling fee for eligible sales to US customers. US beauty sellers already face percentage fees on TikTok Shop and Amazon. An illustrative 2031 investor model assumes future shopping agents could take 10%. The strategic question is whether assistants become another paid route to the customer.
Zero commission still leaves the merchant running the sale. Shopify's ChatGPT route sends shoppers to the retailer's checkout in a browser or new tab. The merchant pays ordinary payment-processing charges, but no additional ChatGPT selling fee. That is the meaning of the 0% comparison: no extra commission on this route, rather than a cost-free sale or a payment completed inside ChatGPT. The scope is eligible stores selling to US customers. Retailers can gain a new discovery channel while keeping their own checkout at the point of purchase.
Existing marketplaces show how a percentage fee works. TikTok Shop's US beauty and personal-care schedule charges a 6% referral fee. Amazon's US beauty, health and personal-care referral fees are 8% when the total sales price is $10 or less and 15% above $10, subject to a $0.30 minimum. The 8–15% range represents price tiers, not uncertainty about the rate. These fees take a share of the transaction before the seller knows its final profit. Their bases also differ: TikTok's formula accounts for discounts, tax and customer-paid shipping; Amazon includes shipping and gift wrapping. Other charges remain separate, so the percentages do not rank total channel cost.
The future 10% is an assumption about revenue, not a quoted price. Investment firm Coatue's October 2026 model uses a 10% take rate for consumer shopping agents in 2031. A take rate means the share of purchase value kept by the intermediary: at 10%, one dollar out of every ten would become agent revenue. The scenario covers global retail excluding China, whereas the current fees here are US examples. Coatue presents illustrative assumptions. No agent has thereby announced a 10% tariff, and the comparison establishes neither that rate's likelihood nor a scheduled rise from today's zero.
The commercial prize is influence over the purchase decision. If shoppers rely on an assistant to choose what to buy, retailers may value its recommendations enough to pay for resulting sales. That would give the assistant a potential commission business even when the merchant still fulfils the order. Whether that helps the retailer depends on what the agent delivers. A genuinely additional sale can justify a fee; charging for a customer who would have bought directly transfers existing margin to the intermediary. The same percentage can therefore represent useful distribution or an expensive new dependency. Checkout ownership alone does not settle who controls demand.
Why it matters: The useful comparison is between ways of reaching the buyer. Today's zero-commission ChatGPT route leaves room for retailers to test demand without an added selling fee. A future commission would change that bargain. For retailers in any market where agents gain shopping influence, the decisive question would be whether they bring additional profitable customers or collect a share of sales the retailer already knew how to win.
