AI shopping agents are changing how consumers shop and are forcing retail sites to finally support conversational search. To enable conversational commerce, however, retailers, payment providers, and answer engines are racing to build AI shopping agents that can go beyond recommending products and actually complete purchases on behalf of consumers with increasing levels of autonomy.  

Still, there is a critical gap between what vendors claim their AI shopping agents can or will do and what consumers are willing to allow today. In April, we surveyed online adults in the US, UK, and Canada to understand their attitudes towards AI agents making payments on their behalf. Forrester clients can see the full published results here. Following is some of what we learned from these consumers: 

• Trust is the real barrier. Three-quarters of these consumers say they are uncomfortable allowing an AI agent to complete a purchase and pay autonomously, even if they could set spending limits and rules in advance. Four main concerns came up repeatedly: Loss of control, errors and liability, distrust around agentic decision-making, and data security and privacy. Trust levels are low across the board, regardless of what company the AI shopping agent comes from.  

• Convenience alone won’t drive adoption. Consumers say they need clear controls for options like spending limits and the ability to approve purchases, as well as transparency around what the agent did and why. They also want strong protections extended to them by the ecosystem as well as responsive customer support when issues arise. However, they aren’t confident today that they will be protected by anyone. Interestingly, consumers expect the answer engine, or AI platform, to be held responsible when an AI agent makes a mistake with a purchase. 

“Who exactly is going to be held responsible for the consequences of the AI’s  actions if they are negative?” 

• Consumers will allow some autonomy, just not for payments. Many consumers are willing to let AI agents recommend products and compare options, and some are even comfortable with an AI agent autonomously filling shopping carts. However, the comfort level for the vast majority of these consumers stops short of letting the agent actually complete a payment. Even for the very few who are comfortable with some level of autonomous payments, it is primarily for routine purchases, low-cost items, or transactions governed by strict preset rules and thresholds.  

In summary, AI agents may soon be able to shop and pay on consumers’ behalf. But today, most consumers aren’t willing to hand over the keys. And when it comes to payments, it’s important to pay attention to these signals of disinterest or distrust. Consider the trajectory of Apple Pay adoption in the US, Apple’s largest market. Despite clear Apple Pay value propositions like enhanced security, speed, and convenience, it took Apple more than a decade to reach the point where just one in four of the US online adults we surveyed reported using Apple Pay to make a purchase in the past three months. And that decade included a global pandemic, when fears of infection drove digital and contactless payment adoption. 

To learn more and discuss these developments in the context of your business, Forrester clients can book a guidance session or inquiry with me. 

Related Links

Related Forrester Content

Categories

Blog

Conway’s Law: Your Operating Model Matters More Than The AI Model

I love an axiom, something easy to remember, fast to say, and punchy enough to stick. With my older boys, I’ve often said, “If in doubt, don’t.” With my younger son, who is autistic, I say, “Stay close; stay safe.” These are short phrases with big truths, the kind that helps in the moment when […]

Blog

OpenAI’s Proposed IPO Opens A Trifecta Of Opportunities For It, But Don’t Lock In Just Yet

OpenAI’s IPO move, paired with its latest manifesto, highlights a race to win consumers, automate enterprises, and advance toward AGI. For business and technology leaders, the implication is clear: Accelerate AI adoption, avoid lock-in, and focus on capabilities, not vendors, as the competitive landscape shifts.