AI Shopping Agents: Shoppers Trust AI to Research, Not to Pay

What happened

Nuvei's latest research into U.S. shopper attitudes suggests that consumers will let AI help them research purchases but are far less willing to let it spend their money. Retail Customer Experience's coverage of the findings reports that 58% of consumers already find AI useful for researching products and comparing prices, and 39% say they are using it more while shopping than they were three months earlier.

Enthusiasm falls away at the point of purchase. The same research finds that 56% of respondents would never allow AI to spend money without their approval, and only 1% want the shopping process fully automated from start to finish. Reporting on the study describes shoppers as treating AI as a research assistant rather than as a wallet.

The coverage places the loss of trust at checkout, which is where the gap between using AI to browse and letting AI act becomes most pronounced.

Why it matters

The technology can now browse, compare and, in principle, complete a transaction on a shopper's behalf. This research indicates that adoption will not follow capability automatically. Organisations building agentic shopping features will need to decide where human approval sits in the journey, because the research places that decision at the centre of consumer trust rather than at its edges.

For leaders in digital transformation, the finding is a reminder that operating-model change depends on how customers experience control. Retailers and payment providers that treat the checkout step as a simple automation target may find that the shoppers most likely to use AI for discovery are the least willing to let it finish the sale.

By the numbers

• 58% of consumers find AI useful for researching products and comparing prices.
• 39% say they are using AI more while shopping than they were three months ago.
• 56% say they would never let AI spend money without their approval.
• 1% want the shopping process fully automated, start to finish.

The Renascence take

Most readers will treat the 1% figure as a ceiling to be pushed past. We read it differently: it is a design brief. The gap is less a trust deficit to be marketed away than a control problem that the product has to solve.

Shoppers are delegating judgement, not authority, and most AI shopping experiences are built the other way round. Treat the agent as an advisor by default, with explicit approval moments, spending limits and easy reversal, and autonomy becomes something customers grant over time rather than something they have to rescue. Operators who make the checkout the most transparent step in the journey, rather than the most automated, will be better placed to earn the permissions their competitors assume they already have.