With only four days remaining until Dreamforce 2026 and NRF Europe converge on September 15-17, the transition to the agentic enterprise has become the primary focus for platform providers like Salesforce, who are positioning merchant readiness as the next frontier of digital retail. Yet, as the hype cycle accelerates, a quiet disconnect persists between the infrastructure being built and the actual behavior of the software agents tasked with spending consumer capital.
The agentic commerce landscape is currently defined by a paradox. According to a Checkout.com report, 42% of merchants are testing agent-mediated systems, and 89% are actively preparing for them. Despite this, only 3% of current transactions involve AI agents. The industry is investing heavily in a channel that, by its very design, threatens to erode the margins of the brands funding the development.
The friction lies in the logic of the agents themselves. Unlike human shoppers, who may be swayed by brand narrative, aesthetic, or emotional resonance, agents prioritize structured data, price, and availability. When an agent shops, it treats a premium brand as a set of variables. If a product is priced 30% higher than a competitor, the agent requires machine-readable proof—such as data showing 50% greater durability—to justify the purchase. Without this, the agent simply reroutes the traffic to the cheapest SKU. In commodity categories, this logic challenges the maintenance of brand equity.
Consumer trust remains a significant hurdle. Data from the Worldpay/Paypers Global Ecommerce Report 2026 indicates that while roughly 50% of global consumers trust AI agents for digital goods under £50, that confidence drops to 30-34% for retail in the same price bracket, and falls further to 21-24% for items between £101 and £500. A Product.ai report from April 2026 reinforces this skepticism: only 14% of US online shoppers trust AI recommendations without verifying them through another source. When an AI makes a mistake, the brand suffers; 81% of consumers will not return to a brand after a single bad experience facilitated by an agent.
For merchants, the path to differentiation is narrow. Some are turning to financial engineering to maintain relevance. Stripe’s integration of Buy Now, Pay Later (BNPL) services via Shared Payment Tokens allows agents to initiate purchases using a buyer’s preferred method without exposing card credentials. Merchants utilizing these tools have reported up to a 14% revenue lift. It is a rare example of a technical integration that provides a tangible edge in an environment otherwise optimized for price-matching.
Market projections include McKinsey’s estimate of a $3-5 trillion global market by 2030, and MarketsandMarkets’ forecast for the agentic AI sector to reach $205.88 billion by 2033. However, the merchant readiness being touted at upcoming conferences presents a challenge regarding the commoditization of the brand. As dual-protocol merchants capture 40% more agentic traffic than their single-protocol counterparts, the cost of entry is rising. For many, the question is no longer whether they can afford to integrate, but whether they can afford to compete in a marketplace where their premium is the first thing to be optimized away.