Shoppers Trust AI to Track Their Parcels, Not to Touch Their Payments

Customers have not rejected AI in service. Almost none of them have. What they have done is sort it, quietly and consistently, into jobs they will hand over and jobs they will not. New consumer research released ahead of the holiday peak shows that the line runs straight through the middle of most service operations: status questions on one side, anything touching money or an account on the other. Companies that route every contact through the same AI front door are working against a preference their customers have already made clear.

Confidence Falls as the Stakes Rise

Sinch surveyed 2,501 consumers across the United States, Australia, Brazil, the United Kingdom, France, Germany, Mexico and Spain. According to the company’s announcement of the findings, 81% of shoppers are confident an AI assistant can handle order tracking and shipping updates, and 74% say the same for pre-purchase questions. Confidence then steps down: 66% for account changes and 62% for payment or billing changes.

Outright refusal is close to zero. Only one of the 2,501 respondents said they would never interact with an AI chatbot. But willingness to use AI is narrower than tolerance of it. The most common intended uses are order tracking at 48% and researching products before buying at 47%, while just 36% say they would use AI for customer service. Robert Gerstmann, Sinch’s chief evangelist and co-founder, summed it up: “AI for AI’s sake doesn’t land with shoppers, but AI attached to a job they actually want done does.”

On the higher-stakes work he was just as direct, noting that for payment, billing and account changes, shoppers still want a human close by. That is not a statement about AI accuracy. It is a statement about where customers want accountability to sit.

Optimism Is Slipping, Especially in North America

The broader mood has cooled. Only 43% of shoppers think AI will make holiday shopping easier this year, down from 48% in 2025. Privacy and data use top the list of reservations, cited by 43%, and only 7% report no concerns at all. The geographic spread is wide: 78% of shoppers in Brazil expect AI to help, while in the U.S. only 26% do and 43% say it will not.

Sinch’s own write-up, syndicated by the Arizona Daily Star on October 5, adds two details operations leaders should note. Preference for human product recommendations edged up to 53% from 51% a year earlier, and the generational gap is large: 55% of Millennials and 51% of Gen Z trust AI recommendations as much as or more than a person’s, compared with 30% of Baby Boomers. A single service design will not land the same way across a customer base that splits this sharply.

Voice is where the split is most visible. According to Sinch, 48% of shoppers trust a human phone agent more than an AI voice agent, while 37% trust both equally or favor the AI. For a channel many operations had planned to automate first, nearly half of callers are telling them the person on the line still matters.

Resolution Beats the Channel Argument

The good news is that customer preference is not fixed. Verint’s State of Customer Experience 2026 research, based on 5,000 U.S. consumers, found that preference for human agents rose 5% year on year, yet 69% of those who currently prefer a human would switch to automated service if it could fully resolve their issue. Seventy-eight percent will give up their preferred channel for a faster resolution.

The same research shows what failure costs. Forty-two percent of customers say their expectations rose in 2026, up from 19% in 2024, and 51% say businesses fall short when they need help. When using automated service, 42% say access to a human agent matters more than speed alone, and 79% would switch to a competitor after a single terrible experience. Phone usage also edged back up to 30% after three years of decline. Put together, the message is that customers will accept automation that finishes the job and punish automation that traps them.

A Task-Based Service Routing Playbook

• Route by task, not by channel: Point AI first at the work customers already trust it with, such as order status and pre-purchase questions, where confidence runs at 81% and 74%. Send billing, payment and account changes to a path with a person close by.
• Make the human path visible on money tasks: With confidence falling to 62% for billing changes, show the route to a live agent before the customer has to hunt for it, not after the AI has failed.
• Invest in voice before automating it: Nearly half of shoppers trust a human phone agent more than an AI voice agent. Give phone agents full customer context when the call connects so the channel customers still prefer is also the one that resolves fastest.
• Measure full resolution, not deflection: Since 69% of human-preferring customers would switch to automation that fully resolves their issue, track whether contacts are actually closed. A deflected contact that comes back through another channel is a cost, not a saving.
• Segment the experience by customer: A 25-point trust gap between Millennials and Baby Boomers means one default flow will frustrate someone. Let customers choose, and use their choices to tune the routing.
• Lead with privacy on the AI path: Privacy is the top concern for 43% of shoppers. Say plainly what an AI assistant can see and do before it asks for account details.

The lesson for peak season is not to slow down on AI in service. It is to deploy it where customers have already said yes and to staff the moments where they have said not yet. The operations that win the holiday rush will be the ones that treat customer trust as a routing rule rather than a satisfaction score to explain afterward.