Does naming a payments brand increase trust in AI transactions?

Visa reports a split between AI familiarity and payment trust in research it commissioned from The Harris Poll. The excerpt states that "Seventy-two percent of US adults have already tried an AI assistant," yet "only 23 percent said they would trust generative AI itself to execute a payment." That hesitation changes when a known payments brand enters the picture: "Sixty-one percent of respondents named Visa as the company they would trust to handle agent-initiated transactions, the highest score among the payments, technology, and social-media names tested." Trust in Visa specifically rose to 68 percent among adults 18 to 34 and to 71 percent among frequent AI users. The underlying poll surveyed 2,065 US adults between May 26 and 28, 2026, with brand-specific trust items asked of roughly 1,030 respondents each — these are reported attitudes, not observed payment behavior.

Visa's own reading, voiced by group president Oliver Jenkyn, is that discovery and money movement are different trust problems: "people will turn to established payment experiences as they give AI more autonomy." The company frames the gap as evidence for its Visa Intelligent Commerce initiative — standards, identity checks, authentication, and consumer controls meant to let an authorized agent complete a purchase without exposing card details or removing the cardholder's ability to set limits. In Visa's account, the fix for low trust in AI-executed payment is not better AI but a familiar intermediary layered underneath it.

This sits alongside other trust surveys in the library as a case where the usual "adoption outpaces trust" pattern — seen in Why do workers feel confident with AI but get poor results? and in Why does AI trust keep falling even though it matters most? — appears again, but with a brand substitution that those surveys don't test: trust transfers to a third party rather than needing to be earned by the AI itself. That also complicates the gatekeeping concern in Could AI assistants become gatekeepers of what people see?, since here consumers are reported as actively resisting letting the AI layer be the trusted intermediary for money movement, instead wanting a separate, recognized authority in that role.

The survey was commissioned and announced by Visa itself to support the launch of its own agentic-commerce infrastructure, so the findings double as marketing for Visa Intelligent Commerce; the excerpt gives no independent verification of the trust figures, no detail on how "trust to handle agent-initiated transactions" was worded or defined, and no evidence that stated trust predicts actual delegation of a real purchase. What it does establish is that, as of this poll, US adults reported much higher comfort with a payments brand's name attached to an agent transaction than with generative AI unnamed and alone — a reported preference for intermediary branding, not a measured outcome of agentic commerce in practice.