Skip to content
LIVE
The Executives BriefThe Executives Brief

AI Shoppers Spend 8% More and Buy 23% More Items. Checkout Is the Wall

Three-quarters of shoppers already trust AI agents for buying advice, but the checkout gap decides whether retail's agentic upside becomes revenue or a liability line.

ByLama Al-RashidTechnology Correspondent, The Executives Brief
·3 min read
AI Shoppers Spend 8% More and Buy 23% More Items. Checkout Is the Wall
Executive summary

ZDNet reports that 74% of shoppers trust AI agents for buying advice, yet checkout remains a separate and unresolved problem, even as agent users spend 8% more than the average buyer and buy 23% more items per order. For retail and payments leaders, the trust is already banked at the research layer, so the strategic fight now moves to authentication, dispute liability, and who owns the moment money actually changes hands.

Start with the number that should be sitting on every retail CEO's desk this morning: 74% of shoppers say they trust AI agents for buying advice. Now the number that should keep them up: checkout is another story. That split comes from ZDNet, and it arrives with two figures that make the gap expensive rather than academic. Shoppers who actually use AI agents spend 8% more than the average buyer, and they buy 23% more items per order.

Read those three numbers together and the shape of the problem snaps into focus. The trust is real, and it is already showing up in basket size. What is not solved is the last mile, the moment a shopper hands an agent a payment credential and lets software commit real money on their behalf. That is where the enthusiasm thins out. It is also where the 8% and the 23% either convert into revenue or stay a rounding error in a slide deck.

For context, the shopping journey has been unbundling for years. Search took discovery away from the storefront. Marketplaces took assortment. Social took inspiration. AI agents are now taking the research layer, the part where a shopper compares specs, reads reviews, and builds a shortlist. Agents are genuinely good at that work, which explains the 74% trust figure. Checkout is a different animal. It requires identity verification, payment authorization, fraud screening, a returns policy, and a clear answer to a question nobody has fully settled: when an agent buys the wrong thing, who eats the cost?

The economics explain why this is more than a survey curiosity. An 8% lift in spend per buyer and a 23% lift in units per order do not simply add up, they compound. A shopper who delegates research tends to delegate more of the basket, and a bigger basket is a bigger order value. But bigger is not automatically better. Retailers typically run on thin margins, and every extra unit in an order carries shipping cost, packaging cost, and a probability of coming back as a return. If agent-driven baskets skew toward more items and more returns, the gross revenue lift can quietly shrink at the contribution-margin line. That is the trap hiding inside a headline number that looks like pure upside.

Then there is the question of who captures the value. In agentic commerce, the scarce asset is not the product page, it is the checkout moment. Whoever owns authentication, payment rails, and dispute resolution sits in the toll booth. Payment networks, issuers, platforms, and the agent providers themselves are all angling for that position, and the one that gets there first sets the default that everyone else has to integrate against. Historically, card networks and issuers have absorbed consumer disputes and pushed chargeback costs back down the chain to merchants. If agents start initiating purchases at scale, that same plumbing gets stress-tested by a new kind of transaction, one where the buyer never saw a button, never typed a card number, and may not remember approving the order. Merchants that assume the existing rules will simply stretch to cover agent-initiated purchases are making a bet, not a plan.

The trust split also reframes where the bottleneck actually sits. If 74% of shoppers already trust AI for advice, the industry does not have a persuasion problem. It has a plumbing and accountability problem. Consumers are not asking whether an agent can find the right product. They are asking, implicitly, what happens if it gets it wrong. That is a question about refunds, reversals, and recourse, and it is answered by policy and infrastructure, not by better model outputs. Companies that treat the checkout gap as a marketing challenge will spend money in the wrong place.

For executives, the practical stakes are concrete. Boards should want to know what share of traffic is already agent-mediated, whether product data and pricing are structured so an agent can read them accurately, and who inside the company owns the policy on agent-initiated returns and disputes. Those are unglamorous questions, and they are the ones that decide whether the 8% and 23% show up as profit or as a customer-service queue.

The 74% is the leading indicator. The 8% and the 23% are the receipts. Trust has already moved to the research layer, and the money is following it right up to the edge of the transaction. The company that solves checkout, cleanly and accountably, converts that trust into revenue. Everyone else watches the basket grow and the margin stay flat.

Executive ActionsLocked

This story's Key Insights and Take-aways are locked.

Create a free account to unlock Executive Actions for one credit.

Register to Unlock

Always free for Executives Club members. Join the Club

More in Technology