Consumers Trust AI for Advice, Not for Purchases: Study

What happened

New research covered by Customer Experience Dive finds a clear split in how shoppers feel about AI's role in retail: they are largely comfortable with AI offering product recommendations and advice, but far more reluctant to let AI actually complete a purchase on their behalf. The findings point to a meaningful trust gap between AI as an advisor and AI as an autonomous agent acting with a shopper's money.

The research frames this as a distinction between "advisory" AI — tools that suggest, compare or guide — and "agentic" AI, which takes action such as clicking "buy." While consumers appear willing to lean on the former, adoption of the latter is being held back by hesitation over ceding control of financial decisions to a system.

Why it matters

Retailers and technology vendors have been racing to build agentic commerce features that let AI shop, compare and check out on a customer's behalf. This research suggests the technology may be arriving faster than consumer trust in it. For experience leaders, the signal is that capability and comfort are not the same thing — an AI system can be technically able to purchase for someone while still failing to earn the psychological permission to do so.

This has direct implications for how brands design and roll out agentic AI: rushing to automate the final, most consequential step of a transaction risks a backlash or simple non-adoption, even where earlier-stage AI assistance is welcomed. The advisory-versus-agentic distinction gives leaders a practical framework for sequencing AI investment around where trust already exists.

The Renascence take

The gap here isn't really about AI capability — it's about control and accountability, two of the most fundamental levers in behavioral economics. Shoppers are happy to outsource cognitive effort (comparing options, filtering choices) but not the final, irreversible act of spending, because that is where perceived risk and personal responsibility concentrate.

Most organisations will read this as a call to "build more trust in the AI" — that's the wrong lesson. The right lesson is to design the handoff, not eliminate it: let AI do the heavy lifting on discovery and recommendation, then keep a visible, low-friction human confirmation step at the point of payment. Removing that final click doesn't make the experience more advanced; it just moves the trust deficit downstream to returns, disputes and brand reputation. The operators who win with agentic commerce will be the ones who treat "who presses buy" as a deliberate design decision, not a technical inevitability.