The Readiness Gap
Only 11% of SMBs and 19% of large merchants currently qualify as agent-ready, according to the PYMNTS/Visa GDSI Merchant Edition. This stark statistic serves as the primary bottleneck for agentic commerce, separating the theoretical potential of autonomous shopping from the reality of digital retail. While the industry fixates on the reasoning capabilities of large language models, the actual constraint is the state of the merchant infrastructure those agents are expected to navigate.
Expectations Versus Infrastructure
There is a profound disconnect between merchant ambition and operational capability. Roughly 68% of merchants anticipate that AI agents will drive at least 5% of their digital sales within two years, with 38% expecting that figure to exceed 15%. Yet, only 15% of merchants possess the structured product data necessary for agents to access and process information effectively. Furthermore, only 23% of merchants can distinguish AI-driven traffic from human users. Without this basic visibility, merchants are essentially operating in the dark, unable to optimize for the very traffic they claim to expect.
The Platform-Native Divide
A bifurcated landscape is emerging in the retail sector. Platform-native merchants, such as those using Shopify, are gaining automated readiness by default. Shopify CEO Tobi Lütke recently confirmed that Shop Pay is live for the Muse agent, effectively making every store on the platform agent-ready without requiring individual merchant intervention. In contrast, independent merchants face significant connectivity friction. They are forced to navigate a fragmented ecosystem where their existing systems are often incompatible with the requirements of autonomous agents, leaving them at a distinct disadvantage in an increasingly automated market.
Friction Over Reasoning
The limitations of current agentic commerce are often misattributed to AI failure, but the evidence suggests otherwise. When PYMNTS tested the Muse agent on September 9, all three attempted errands failed. Crucially, these failures were not due to a lack of AI reasoning, but rather to credential sharing friction and app connection issues. The technology is ready to act, but the digital storefronts are not configured to receive it. This is further supported by data from Checkout.com, which shows that while 42% of merchants are testing agentic commerce, only 3% of actual transactions currently involve agents.
The Liability Wall
Financial risk remains a primary deterrent for merchant adoption. The industry is currently at an impasse regarding who should be held accountable for the actions of autonomous agents. A significant 93% of merchants believe that the AI provider should bear the financial loss for incorrect purchases. This lack of consensus on liability is directly limiting the product range merchants are willing to expose to agents, with only 28% currently willing to offer their full inventory. Until this liability framework is resolved, merchants will likely continue to restrict agent access to their catalogs.
The Plumbing Problem
The timeline for widespread adoption depends on resolving core issues like liability, identity, and data standardization. The industry is attempting to respond with initiatives like the cross-network Know Your Agent framework, currently being developed by Visa, Mastercard, and Ant International, which aims to provide a foundation for secure agent interactions. Additionally, tools like Stripe WebMCP are attempting to standardize agent checkout UIs. Whether these efforts will be sufficient to overcome the deep-seated structural issues remains an open question. As we head into the next holiday season, the focus will likely remain on the mundane, yet critical, work of fixing the plumbing of digital commerce, rather than the hype of agent capabilities.