The news: Consumers are more inclined to let AI agents make purchases on their behalf if strong guardrails are embedded within the payment process, per Synchrony’s AI in Commerce survey.

• 82% of consumers want AI shopping to keep their data secure.
• 77% want clarity on how their data is being used.
• 76% want to choose how much control they have during an AI shopping experience.

By comparison, only 58% of consumers said saving time was their highest AI commerce priority.

Why this matters: Building trust with consumers for AI and agentic commerce will be a piecemeal process aimed at shoppers’ current degree of comfort around AI.

• 79% of consumers are comfortable having AI apply their discount codes, per Synchrony.
• 74% trust AI to apply for loyalty points.

But consumers’ trust plunges once tasks get closer to the actual .

• Only 43% feel comfortable having AI agents spend up to a certain preset limit.
• 37% trust agents to buy regularly used products automatically.
• And for purchases over $5,000, 46% of consumers would not use AI in any way, and only 17% would let AI act autonomously based on their past behavior and purchases.

Recommendations for payment providers: To foster trust in these systems, payment providers need to meet customers where they’re at and address their biggest anxieties around agentic commerce.

• Embed consumer approval notifications into the AI agent shopping experience.
• Create AI/agentic protections for consumers in case of agent failure, like American Express’ Agent Purchase Protection.
• And demonstrate the utility of agents with low-risk, routine, and reversible purchases, like coffee bean orders or pet supplies.

However, a single AI agent error can seriously set back consumer sentiment. The overwhelming majority of consumers say they would limit their AI agent use if a shopping assistant bought the wrong item, shipped something to the wrong address, or used the wrong payment method—and more than 25% would stop using AI assistants entirely.