Shopify Is "Winning" Agentic Commerce. Nobody's Asking What the Prize Actually Is.

Shopify's had a quarter. The kind of quarter where the financial press starts using words like "vindication" and "paradigm" in the same sentence (always a reliable sign that someone's about to get humbled, but let's enjoy it while it lasts).

The headline numbers: traffic reaching Shopify merchants from AI tools rose roughly eightfold year on year in Q1 2026. Orders generated through AI-powered searches? Up close to thirteenfold. For a company that spent January being eulogised as a probable casualty of AI shopping, these numbers landed like a plot twist in a film nobody was paying attention to.

Then Q2 arrived and the same two measures grew by roughly threefold year on year. Still impressive by any normal standard. But threefold is a long way from thirteenfold, and the step-down came exactly one quarter later.

Before anyone starts writing obituaries in the other direction — a number that triples off a larger base can represent more absolute activity than one that grew thirteenfold off almost nothing. Quarterly comparisons distort easily when the starting base is "basically zero." The point isn't that agentic commerce has peaked. The point is that everyone's so busy declaring Shopify the winner that nobody's stopped to ask what, precisely, the trophy is made of.

What Shopify Actually Built (And Why Everyone's Clapping)

Quick primer for anyone who's been mercifully ignoring the phrase "agentic commerce" (enjoy your remaining innocence while you can): it's shopping where an AI assistant does the work a human shopper used to do — searching, comparing, shortlisting, sometimes buying — on someone's behalf. You describe what you want to ChatGPT or Gemini, and the assistant returns products from across the web. You still decide what you need. The shopping itself is increasingly done by software.

The problem this creates for merchants is immediate. An AI assistant doesn't respond to your beautiful product photography or your brand's origin story about a goat farm in Vermont. It parses structured data — machine-readable fields describing price, availability, materials, dimensions. A listing built to charm a human can be functionally invisible to an agent.

Shopify's answer is Catalog. It takes messy, inconsistent product data and renders it into clean, standardised, machine-readable structure, built on the Universal Commerce Protocol (co-developed with Google and a coalition of retailers). Merchants' products appear across multiple AI surfaces from a single setup. Purchases route back through Shopify's checkout.

The bullish case writes itself: Shopify helped author the protocol, structured over a billion products into it, and can syndicate merchants into major AI shopping surfaces from one dashboard. Clean Catalog data converts at roughly twice the rate of scraped web data. Revenue grew by around a third year on year in H1. GMV cleared $100 billion a quarter.

Coherent story. Also the kind of story that sounds bulletproof right up until someone asks which part of the bullet is actually doing the proofing.

The Layer Everyone's Praising Is the Layer Most Easily Replaced

Structured product data feels like a moat. It's difficult, unglamorous work. Most merchants do it badly. The payoff — being legible to the systems that now mediate discovery — is real.

But difficulty for an individual merchant is not the same thing as durability for the platform that organises it. Google's Shopping Graph holds on the order of sixty billion listings and traces its lineage back over two decades. That's the same raw asset Catalog is assembling, at a scale that dwarfs it, owned by one of the surfaces Shopify is syndicating into (which is the kind of competitive dynamic that should make your eye twitch slightly).

The frontier AI labs have already concluded that scraping product pages was never going to work and are building their own commerce plumbing to ingest merchant data directly. If the structured-data layer can be reconstructed by the very platforms that control discovery, then being the best current supplier of that data is a convenience, not a fortification. Convenience is valuable. Convenience is also the kind of thing that gets competed away by companies with more engineers than some countries have citizens.

The parts of Shopify's position that actually resist erosion sit further from the data and closer to the money: the checkout shoppers trust, the payment processing handling roughly two-thirds of transaction volume, and merchant-of-record status. When one major AI company attempted native in-chat checkout, fewer than thirty of Shopify's several million merchants went live before the feature was quietly withdrawn. The payments problem proved far harder than the discovery problem. The applause has been directed at the layer that's easiest to strip out. The hard, defensible work sat at the transaction all along.

"You Own the Customer" Is Becoming True and Hollow at the Same Time

Shopify's founding promise has always been customer ownership. In the agentic model, this is preserved by design: discovery happens inside an AI assistant, but the shopper is handed to the merchant's checkout. Ownership intact. On paper.

The difficulty is that ownership of the transaction is not the same as ownership of the relationship. And the two are quietly separating.

Google has begun letting users grant its assistant access to their Gmail, Photos and Drive, so that recommendations can be shaped by what the model already knows about them. The surface that holds a shopper's accumulated context — their taste graph, their standing model of preference — is the surface that decides which products clear the bar into consideration. A merchant can own the receipt and own none of the upstream signal that determined whether the sale had a chance of happening.

You own the checkout. You don't own the reason the checkout fired.

The Concession Might Be the Strategy

Follow that diagnosis to its conclusion and something counterintuitive emerges. If discovery is the contested, commoditisable layer and the durable strength sits at checkout and payments, then the rational move isn't to fight to own discovery. It's to concede discovery on purpose and drive all the way into being the layer every agent must call to complete a sale — the point at which a recommendation becomes a paid order, wherever that recommendation originated.

The early shape is already visible. Catalog has been opened to brands that don't even run stores on Shopify. The checkout is being engineered to be invoked from surfaces Shopify doesn't own. Payments already clear the majority of the platform's volume. Each move loosens the assumption that Shopify's business is the store and tightens a different one: that its business is the settlement — the clearing of the money.

Taken to its end, Shopify stops trying to be where shopping starts and positions to be where shopping settles. Every agent that surfaces a product still needs something trustworthy to run the payment, honour the return, and carry the liability. That's the expensive, unglamorous competence Shopify already has. What looks like the company's exposure — the loss of the discovery relationship — is precisely what pushes it toward a larger prize: infrastructure for all of agentic commerce rather than the storefront for a slice of it.

What This Does to the D2C Playbook

Here's where it gets uncomfortable for the rest of us.

The D2C playbook has had one durable shape for fifteen years: buy attention on Meta and Google, drive it to your store, convert it there, capture the data, retarget, build lifetime value. Every part of that machine points at a single destination — the store is where discovery, comparison and conversion all happen, and the tracking pixel watching on-site behaviour is what makes the flywheel turn.

Agentic discovery detaches the middle of that funnel from the store. When a shopper describes a need to an assistant and gets handed a shortlist, the browsing, comparing and narrowing happen inside the model — on a surface you cannot instrument. The sale may still arrive, but you never saw the consideration that produced it. The retargeting pixel has nothing to retarget, because the part of the journey it was built to watch happened somewhere it cannot see.

The irony is worth stating plainly. Brands adopted the buy-traffic-and-own-the-store model precisely to escape dependence on Amazon's marketplace — to avoid renting access to their own customers from an intermediary. Agentic commerce reinserts an intermediary between brand and shopper. And the platform that sold independence from marketplaces is, on the reading above, optimising to become a neutral rail that by definition does not care whether your brand or your competitor's clears through it. The tool built to help brands own the customer may be heading toward a model that profits from throughput and is indifferent to whose (which is, if you think about it, exactly the thing everyone was trying to escape by leaving Amazon in the first place).

So the target of acquisition spend has to move. It shifts toward two harder things: salience — making shoppers ask for your brand by name inside the assistant, because a named request narrows an agent's consideration set to you specifically — and legibility good enough to win, not merely to qualify. Feed quality, accuracy and verification complete enough that when the request is generic, the agent picks you over a near-identical rival.

Only around a quarter of shoppers currently say they're comfortable buying inside an AI chat. The timing question is genuinely open. There's no risk-free answer — anyone claiming certainty is guessing with conviction, which is a polite way of saying "bluffing."

The Bottom Line

Shopify has earned a lot of the applause. It moved early, helped write the standard, and is strong in the parts of the stack that are hard to replicate. But its most defensible future might be to become the settlement rail beneath every agent — winning by processing volume rather than protecting any individual merchant's customer ownership. That route, if taken, sits in quiet tension with the promise the whole company was built to keep.

The harder question isn't whether Shopify survives the agentic shift. It's what the merchant becomes in the world Shopify's own logic points toward: the customer it serves, or the throughput it clears? And how long does "the merchant owns the customer" hold as a claim once the customer's taste lives inside a model no merchant controls, and the platform that made the promise is busy building the toll booth?

P.S. If you're a Shopify merchant who just spent three months perfecting your Catalog data, congratulations — you've bought yourself a seat at a table whose seating chart is being redrawn quarterly. Welcome to ecommerce in 2026.

• Shared Claude conversations turned up in Google search results, some exposing crypto keys and personal details. Anthropic fixed the issue; similar leaks have hit ChatGPT and Grok, a recurring industry problem.

• New study finds citations spread across categories more readily than brand mentions in ChatGPT. Depth in a few categories beats breadth, and named mentions cluster near a brand's core expertise.

• Ex-Lululemon CIO warns of AI wishing and AI washing, noting 95% of enterprise GenAI pilots fail and 75% of executives call their AI strategy more for show. The hard work is fixing messy data.

• Researchers find Meta approved and ran over 50 paid ads containing AI-generated child abuse imagery across its platforms, many linking to nudify apps. Meta removed them after WIRED inquired.

• An agency reports influencing major LLMs to reflect its ideal customer profile within a week by aligning brand messaging across titles, site copy, and LinkedIn. Smaller brands prove easier to steer.

• Kroger launches an AI shopping assistant with ads from day one, showing sponsored product listings alongside organic results. Existing Precision Marketing campaigns appear automatically, no extra setup.

📹 Outset - An AI tool that interviews users/candidates with video.

💬 HeyGen GPT - Transform text into lifelike and customized videos instantly

📊 Papermark AI - Revolutionize your document sharing and collaboration. Share pitch decks, sales proposals and other docs securely with real-time analytics and white-labeling options.

📈 Optimove - Data consolidation platform offering a Customer Data Platform (CDP) for unified, actionable customer insights across various sources.

📹 Creatify AI - Turn a product URL into a video ad instantly

😍It was great, all I needed

🙂Good, not great

🙄It sucked