Eighty-nine percent of merchants are actively preparing for agentic commerce, yet only 3% of transactions currently involve AI agents. This discrepancy between corporate investment and actual capital flow highlights a market stalled at the point of execution. While 72% of merchants believe consumers will adopt agent-led shopping faster than the industry is prepared for, the current reality remains tethered to manual oversight.
Forty-eight percent of online shoppers now use AI to conduct research before making a purchase, but this intent rarely converts into an automated checkout. The friction is not located within the financial plumbing. According to Paymentology CTO Tim Joslyn, 99% of existing issuer processing systems could already process an agentic payment. If the infrastructure is capable of handling the volume, why does the transaction rate remain in the single digits?
Consumer behavior indicates a fundamental deficit of trust. The Product.ai Trust in AI Commerce Report found that only 14% of consumers who used AI for product research trusted those recommendations without verification. Furthermore, 86% of users verify their purchases before finalizing them, and 42% refuse to trust AI for any transaction exceeding $25. With 27% of consumers stating they trust no organization to operate an AI shopping agent, the barrier to entry is psychological and institutional rather than technical.
Retail sectors show varying levels of consumer comfort with delegation. Consumers are most willing to delegate low-risk, repetitive tasks, with 41% comfortable delegating grocery shopping and 31% for household supplies. Conversely, only 15% are willing to delegate financial services. This risk-reward calculation suggests that agentic commerce will gain its initial foothold in high-frequency, low-stakes categories where the cost of an error is minimal.
Consumers have identified three specific safeguards required to bridge the trust gap: spending caps (30%), instant revocation (29%), and easy cancellation (28%). On average, consumers would allow an AI agent to spend £177 per purchase without additional approvals, provided these guardrails are in place. These requirements define the necessary infrastructure for future commerce. Value will likely accrue not to those who build faster agents, but to those who build the trust infrastructure — the identity, authorization, and fraud-prevention layers that make these spending caps and revocation rights enforceable.
Brand loyalty is also facing a significant disruption. Checkout.com data indicates that 57% of consumers would allow an AI shopping agent to switch brands if it meant securing better value. This shift challenges the traditional merchant-consumer relationship. If an agent is optimized for value rather than brand affinity, the power dynamic moves from the brand to the agent’s underlying logic. Merchants relying on legacy loyalty programs may find their influence eroded by algorithms that prioritize price and utility over historical brand preference.
Market sentiment is shifting, albeit slowly. Riskified’s June 2026 study shows that 70% of consumers are now comfortable with agent purchases, an increase from the 55% who were uncomfortable in the first quarter of the year. While 24% of consumers maintain they will never delegate purchases to AI, the broader trend suggests a cautious warming to the concept. The challenge for the payment ecosystem is to shift focus from the rails to the trust architecture. Until consumers feel they have absolute control via caps, revocation, and verification, the 89% of merchants preparing for this shift will continue to wait.