June 22, 2026

Shopify Drops the Gatekeeper, Adyen Becomes the Translator, Europe Names the Execution Gap

June 16-22, 2026. Last week the card networks shipped the rails. This week the commerce platforms shipped the tooling that sits on top of them. Shopify made its agentic stack self-serve, Adyen turned the multi-protocol mess into one integration, and Europe’s biggest retail gathering put a name to what is actually holding adoption back.

Last week we covered Visa wiring itself into ChatGPT, Mastercard launching machine-to-machine payments, and the consumer trust gap that none of the plumbing has closed yet. That was the payments layer. This week the platforms answered with the merchant layer. On June 17, Shopify’s Spring ‘26 Edition removed the approval gate on its agentic commerce stack and shipped a Catalog API built for AI agents. On June 16, Adyen launched Adyen Agentic to spare enterprise merchants from rebuilding for every new AI surface. And at Shoptalk Europe in Barcelona, the room stopped debating whether agentic commerce is real and started arguing about why so few merchants can actually execute on it.

Here’s what happened.

Shopify makes agentic commerce self-serve

Shopify’s Spring ‘26 Edition, shipped June 17, packed more than 150 updates, but the headline is one removed sentence. Building on Shopify’s agentic commerce layer used to require approval. That requirement is gone. Developers no longer apply for access to the Universal Commerce Protocol, the standard Shopify co-developed with Google for how AI agents discover, cart, and check out across merchants. They register an agent profile in the Developer Dashboard, call the public MCP endpoint, and build the full flow from product search to checkout. Agentic commerce on Shopify is now self-serve for the entire developer ecosystem.

The merchant-facing pieces matter more than the developer plumbing. The new Catalog API turns a store’s product data into queryable infrastructure for agents, with image search, multi-modal search that mixes text and image, product URL resolution, and richer metadata like size, color, and delivery estimates. Shopify’s number on it is the one to remember: AI searches powered by Shopify Catalog convert at twice the rate of those running on scraped data. Shop Pay opened to non-Shopify brands, putting one-click checkout and a base of 250 million-plus shoppers behind merchants who never ran on Shopify. Shop Campaigns, where merchants only pay when a customer converts, expanded to ChatGPT and other surfaces. And a new Agentic Storefronts dashboard pulls management of all these AI channels into one place. Shopify pointed at two merchants for proof: Omnilux said AI channels drove 3.2% of total revenue in March, and Cozy Earth reported AI channel revenue up 20x year over year.

Why it matters for merchants: This is the moment the structured-data advantage stopped being a thesis and became a published conversion number. Two times the conversion rate on clean catalog data versus scraped data is the entire discoverability argument in one statistic: an agent that has to scrape and guess at your product converts half as well as one handed complete, structured attributes. That gap is your opportunity if your catalog is rich and your problem if it is thin. The removal of the approval gate is the other signal. The friction to plug agents into commerce is being engineered toward zero, which means the moat is no longer access, it is data quality. When anyone can build an agent against any catalog with no gatekeeper, the merchants who win are the ones whose product data answers the agent’s questions before it has to ask. (See our coverage of the 99% catalog-readiness gap for how far most feeds are from passing, and our piece on catalogs becoming ad inventory for where Shop Campaigns is heading.)

Adyen ships a universal translator for AI surfaces

On June 16, Adyen launched Adyen Agentic, pitched as a “universal translator” for the next era of commerce. The problem it targets is the one every multi-protocol merchant has been complaining about: each AI surface runs on different protocols, demands different product data formats, and has its own cart and checkout requirements, so selling on a new one means rebuilding from scratch. Adyen Agentic is a set of modular APIs in three layers, Agentic Feed, Agentic Cart, and Agentic Payments, that map a merchant’s existing commerce systems onto whatever protocol the surface speaks, from discovery through to payment. Karan Katyal, Adyen’s global head of agentic commerce, put the bet plainly: “Every new agentic surface asks merchants to rebuild from scratch. We believe the future of agentic commerce should be open.” Early participants span the networks and platforms (American Express, Mastercard, Salesforce, Visa) and the retail side (Sézane, SharkNinja, Scheels, ESW), per PYMNTS’ reporting. For now the offering is limited to US-based enterprise merchants.

Why it matters for merchants: This is the same lesson the Shopify news teaches, aimed at the merchants who are not on Shopify. The fragmentation tax is real: ChatGPT, Google’s UCP surfaces, Copilot, and Meta each want your catalog and your checkout in their own shape, and a merchant who hand-builds for each one is signing up for permanent integration debt. Adyen’s pitch is to integrate once and inherit the surfaces as they switch on, which is the protocol-layer logic we have argued for since the start, packaged for the enterprise. The caveat is the eligibility line: US enterprise only, for now. The strategic read for everyone else is that the integration problem is being commoditized fast, by Shopify on its side and by acquirers like Adyen on theirs, so the lasting differentiator is not which integration you pick. It is whether the catalog flowing through it is good enough to win the agent’s choice. (See our UCP vs ACP breakdown for the protocol map this translates between.)

Europe puts a name to the real bottleneck: execution

At Shoptalk Europe in Barcelona earlier this month, agentic commerce was, by the event’s own coverage, the defining phrase of the show, used for everything from assisted shopping to fully delegated buying. The more useful output was the diagnosis. Three themes ran through the sessions: an AI execution gap, where adoption is racing ahead of measurable business impact; a consumer trust deficit, the same hesitation the checkout.com study quantified last week; and the shift from search-first to discovery-led commerce. Zalando’s co-CEO David Schröder gave the standout adoption data point, reporting that the retailer’s AI assistant grew from 6 million users in 2025 to 10 million in the first quarter of 2026.

Why it matters for merchants: Europe naming the execution gap is the corrective to a week of “look how easy we made it” platform announcements. Shopify and Adyen both shipped tools that lower the barrier to plugging in. Shoptalk Europe’s point is that plugging in is not the same as performing, and most retailers can show AI activity but not AI impact. The two halves fit together: the platforms removed the integration excuse, so the unglamorous work, clean catalog, accurate inventory, measurable attribution, is now the whole game. For European merchants specifically, the signal is that this is no longer a US-watching exercise. The continent’s flagship retail event treated agentic commerce as the present tense, and a European retailer the size of Zalando is posting agent-adoption numbers in the millions. The European window we have flagged since April is open, and the work that fills it is execution, not waiting. (See our coverage of Europe’s first-party channel for the strategic context.)

Meanwhile, the case that decides whether agents can visit your site went to appeal

Running underneath the tooling week is the legal question that determines whether any of it is allowed. On June 11, the Ninth Circuit Court of Appeals heard oral arguments in Amazon v. Perplexity, the first federal appellate test of whether an AI agent acting on a user’s explicit permission is an authorized visitor to a logged-in website. The backstory: Amazon sued in November 2025, a district judge granted a preliminary injunction in March 2026 finding Amazon likely to win its Computer Fraud and Abuse Act claim against Perplexity’s Comet browser, and the appeals court paused that injunction pending appeal. Perplexity’s argument to the panel was that it built a browser letting an agent act for a user, not a system that hacks Amazon, and that Amazon is stretching anti-hacking statutes because it wants customers to use its site Amazon’s preferred way. No ruling came; the case was submitted, and CFAA appeals typically take months.

Why it matters for merchants: This is the question every platform announcement this week assumes is already settled. Shopify, Adyen, Google, and Visa are all building agent access on the premise that an agent carrying a user’s permission may transact on a merchant’s behalf. Amazon is litigating the opposite: that permission from the user is not authorization from the site, and a retailer can wall agents out. Whichever way the Ninth Circuit eventually rules sets the default for everyone. A win for Amazon hands every merchant a legal lever to block or gate agents, which would slow the open-access model Shopify just embraced. A win for Perplexity cements the agent’s right to act for the shopper, which makes the discoverability work non-optional because you will not be able to opt out of being shopped. Either way, the strategic posture is the same: assume agents are coming to your catalog and get it ready, because the upside case makes readiness mandatory and the downside case still leaves your own first-party agent channel wide open. (See our original coverage of the Amazon vs Perplexity battle for how this started.)

Where the AI commerce stack stands this week

Layer	Update this week
Platform tooling	Shopify Spring ‘26: UCP access goes self-serve, no approval, public MCP endpoint
Catalog readiness	Shopify Catalog API: AI searches convert at 2x the rate of scraped data
Checkout reach	Shop Pay opens to non-Shopify brands, 250M+ shoppers, one-click
Enterprise integration	Adyen Agentic ships “universal translator,” Feed/Cart/Payments layers, US enterprise only
Europe	Shoptalk Europe names the execution gap; Zalando’s AI assistant hits 10M users in Q1
Legal	Amazon v. Perplexity Ninth Circuit oral arguments held June 11, ruling pending

What merchants should do this week

1. Treat the 2x conversion number as your business case. Shopify just published that AI searches on structured catalog data convert at twice the rate of scraped data. That is the discoverability argument made measurable. Use it to justify the unglamorous work: complete attributes, accurate pricing, real-time inventory, multi-modal-ready product imagery. The agent that scrapes you converts half as well as the agent you feed properly.

2. Integrate once, at the protocol layer, not per surface. Both Shopify and Adyen shipped tooling this week to spare you from rebuilding for every AI channel. If you are on Shopify, the agentic stack is now self-serve, so there is no access excuse left. If you are an enterprise elsewhere, the acquirer layer (Adyen, and the others that will follow) is the integration-once path. Pick the layer, then spend your real effort on catalog quality, because that is the part no integration fixes for you.

3. Measure impact, not activity. Shoptalk Europe’s verdict was that most retailers can show AI is happening but not that it is working. Stand up attribution for your AI channels now: which agent surfaces drive sessions, carts, and revenue. Omnilux could say AI was 3.2% of revenue because it was measuring. If you cannot put a number on your agent channel, you cannot manage it.

4. Build as if agents have the right to shop you, because the safe bet is they will. The Ninth Circuit will eventually rule on whether merchants can wall agents out, but the strategic move does not depend on the outcome. If agents win the right to act for shoppers, catalog readiness is mandatory. If merchants win the right to gate them, your own first-party agent channel still needs that same readiness. Either path ends at the same to-do list, so start it now.