If AI shopping agents take off, do people still trust them enough to actually buy — and stay loyal to brands?

This explores what has to be true about how fast people adopt AI shopping agents, and how much they trust them, for agent-commerce revenue forecasts to hold up. The corpus doesn't contain the projection itself, so this can't check its numbers. It can show which assumptions such forecasts need and how solid each one looks. Most bullish forecasts quietly assume three things. People will hand purchases to agents. They will let agents pay. And handing over control won't change what they buy or which brands they buy from. The evidence supports the first two more than you might expect and pushes back hard on the third.

Start with adoption. Surveys suggest people are willing: Accenture found 74% of 25,590 respondents would trust an AI agent with purchases Do consumers actually want to delegate shopping decisions to AI?. But in the same survey, 56% would still name the brand themselves and 37% of loyal customers wouldn't switch. So delegation is real but partial. It looks more like 'go buy my usual detergent' than 'pick the best one for me.' A projection that counts total spending passing through agents might hold up. One that assumes agents will move market share between brands needs far more delegation than people say they'll give. Uptake also isn't constant. Five studies, including real ad click-through data, found that people accept high-autonomy shopping assistants because of the benefits they see, and acceptance rises when goods are scarce What drives shoppers to accept autonomous AI assistants?. That points to adoption coming in uneven, use-case-driven bursts rather than a smooth S-curve.

Payment trust is the most interesting part. In a Visa-commissioned survey, 72% of US adults use AI assistants but only 23% trust AI to make payments. That figure rose to 61% when Visa's name was attached Does naming a payments brand increase trust in AI transactions?. So a projection doesn't need people to trust AI in general. It needs a trusted payment brand to stand behind the agent, which makes revenue depend on payment-network partnerships as much as on how good the agents are. Behavior also runs ahead of what people say. Instinct reports that 40% of its users share a credit card with its agent within three weeks How fast do users trust AI agents with payment data?. If that holds outside one company's self-reported numbers, forecasts built on survey trust figures may actually be too cautious.

The biggest assumption usually goes unstated: that the agent is a neutral pipe and the demand it handles would have happened anyway. Bilgihan and colleagues argue that agents which actually complete purchases shape what people want through repeated conversations, rather than just carrying out preferences people already had Do AI shopping agents reshape what consumers actually want?. That cuts both ways for a forecast. It could create new spending, or it could just move existing spending from one brand to another. The business model matters too. Bob Hedges warns that ad-funded agents have built-in reasons to favor paid placements Do advertising-funded shopping agents favor paid product placements?. And sellers can work out a buyer's willingness to pay almost exactly from shopping conversations alone Can a seller infer what a buyer really wants to pay?. Both are the kind of discovery that could break trust once it becomes public.

So a credible projection needs: partial delegation, not full; trust borrowed from known payment brands rather than earned by AI; adoption that arrives in bursts tied to specific use cases; and no major trust shock from biased recommendations or price extraction. The surprise is that the trust gap shrinks quickly in practice. The fragile assumption is that trust, once given, survives people learning whose interests the agent actually serves.