Marketplace Mandates: Why Merchants Must Navigate the Agentic Policy Purge

As we move through mid-2026, the narrative surrounding Agentic Commerce has bifurcated. On one side, financial institutions and tech giants have raced to build open rails—protocols like the Universal Commerce Protocol (UCP) and Machine Payments Protocol (MPP)—to facilitate seamless machine-to-machine transactions. On the other side, the world's largest e-commerce platforms have initiated a systematic crackdown on autonomous agents.

This divergence creates a critical strategic challenge for merchants: how to capitalize on the projected $5 trillion global agentic commerce market without losing visibility on the dominant retail platforms that are actively banning these very actors.

The Shift from Gatekeeping to Enforcement

In early 2026, the industry viewed AI-powered "shopping bots" largely as a novelty or a tool for savvy consumers. That perception shifted rapidly when major marketplaces determined that unregulated agents were degrading the quality of the shopping experience.

On January 20, 2026, eBay updated its User Agreement to explicitly ban third-party "buy-for-me" AI agents and LLM-driven bots from interacting with its platform [1]. Effective February 20, 2026, the platform began blocking autonomous end-to-end purchasing flows, signaling to merchants that relying on generic bots for discovery or acquisition would no longer yield returns on eBay [3]. Similarly, Amazon issued new Automated Seller Rules effective March 4, 2026, tightening enforcement against unauthorized automated tools [4].

"eBay has an obvious reason to dislike AI shoppers: It charges sellers a variable final value fee, which sees it earn more money on higher-priced items. High-frequency bot trading threatens the integrity of its auction and buy-it-now models." — The Register, analyzing the strategic motivation behind eBay's policy shift [2].

Economic Drivers Behind the Purge

The crackdown is driven by distinct economic imperatives rather than technical limitations alone. Marketplaces argue that autonomous agents prioritize speed over intent, leading to cart clogging and inventory hoarding by bots that rarely convert to human-like engagement patterns. These behaviors strain server resources and distort demand signals, ultimately degrading user experience for actual consumers.

For merchants, this means that simply plugging into an Open Web protocol like UCP does not guarantee access to all inventory channels. A product may be discoverable by Google Search agents operating on open protocols but invisible to agents attempting to scrape or purchase on restricted platforms. The disconnect between open discovery layers and closed transaction environments forces retailers to confront the reality that agentic reach is now bounded by platform permission structures.

The Walled Garden Defense: Apple's Approach

A similar philosophy is evident in the iOS ecosystem. At WWDC 2026 in June, Apple unveiled its Siri AI overhaul powered by Google Gemini. Unlike the aggressive open-commerce pushes seen in the Android space, Apple positioned its Siri integration as a personal concierge within a secure walled garden [5].

Reports suggest Apple intends to vet app integrations strictly, limiting the types of autonomous commercial interactions allowed outside the approved App Store framework [6]. This approach reinforces the broader trend where major tech incumbents are choosing to channel agentic activity through controlled interfaces rather than enabling unrestricted agent proliferation across the open web.

Strategic Implications for Merchants

For merchants operating in this dual environment, the "plug-and-play" mentality of early agentic adoption is ending. Two key adaptations are required:

• Official Integration over Scraping: Merchants are finding that negotiating direct API partnerships with platforms for authorized agent access is becoming more valuable than optimizing for public scraping. Recent analysis of browser dominance highlights that raw scraping workflows are increasingly fragile against evolving countermeasures, pushing merchants toward stable, authorized connections [7]. This ensures compliance with Terms of Service while retaining revenue streams from legitimate AI buyers.
• Data Segmentation: With different protocols governing different environments, retailers must prepare for increased complexity in their digital catalogs. Product feeds must satisfy the structural requirements of agentic discovery engines while simultaneously adhering to the restrictive taxonomies and access controls enforced by closed-loop marketplaces. Failure to segment data effectively could result in an agent being capable of finding a product via open web search but unable to resolve the purchase path due to platform-specific API authentication failures.

The rise of Agentic Commerce is undeniable, but its trajectory is now defined by the tension between open infrastructure and platform control. As we progress into late 2026, the winners will be those who master both the technology of autonomous buying and the art of platform diplomacy. Merchants that secure explicit permissions within these curated ecosystems will preserve their ability to transact with machines, while those relying on unregulated scraping risk complete exclusion from the most lucrative marketplaces.