Shopify Goes Agentic: What It Actually Changes for Your Store

In brief

‍

At its Winter '26 Edition, Shopify shipped more than 150 features built around agentic commerce. Agentic Storefronts make your products discoverable across AI channels (ChatGPT, Microsoft Copilot, Google AI Mode) through the Shopify Catalog, with no integration to build. Two details the marketing coverage skips: in ChatGPT, discovery now sends the buyer back to your own checkout, and eligibility requires selling to US customers. The real challenge for merchants isn't technical. It's knowing where to automate a task and, above all, where not to delegate judgment to an AI.

• Agentic Storefronts syndicate your products to AI channels through the Shopify Catalog, on by default since March 2026
• For ChatGPT, the purchase no longer closes in the conversation: the agent discovers, then redirects to your checkout, and you must sell to US customers to appear there
• Sidekick moves from assistant to operator: the risk shifts from technical to governing delegated decisions

‍

Shopify shipped a major strategic shift in late 2025 that many merchants are still trying to decode. At its Winter '26 Edition, nicknamed "Renaissance," the company announced more than 150 features, nearly all oriented toward what it calls agentic commerce. The term sounds like marketing jargon until you look at what it describes: AI agents that discover products, compare options, and in some channels complete purchases on behalf of a user, inside chat interfaces like ChatGPT or Perplexity.

‍

Here's the concrete version. A customer asks ChatGPT to find running shoes under $150. ChatGPT queries the Shopify Catalog, surfaces relevant products with pricing and images, then sends the customer back to your checkout to close. The nuance matters: since March 2026, a purchase in ChatGPT no longer finishes inside the chat window. OpenAI walked back its in-chat checkout model, and the buyer now completes on your store, in an in-app browser on mobile or a new tab on desktop.

‍

This isn't speculative. According to Shopify's own data, AI-referred traffic to stores on the platform has grown sevenfold and AI-attributed orders elevenfold since January 2025. For agencies advising Shopify clients, the real question isn't whether this matters. It's how to position it without overselling capabilities that are still in motion, and without letting a client confuse "turning on a feature" with "handing a decision to a machine."

Why is Shopify accelerating so hard on AI right now?

‍

Shopify's agentic push is as much defensive as it is innovative. The company faces a disintermediation threat: if shoppers discover and buy products directly inside ChatGPT, Perplexity, or Google AI Mode, the merchant's storefront risks becoming a mere payment step. Shopify's bet is to become the infrastructure layer connecting merchants to all of these interfaces, rather than letting them build direct integrations with each AI platform.

‍

The numbers explain the urgency. McKinsey estimates the global agentic commerce opportunity at three to five trillion dollars by 2030. ChatGPT alone exceeds 880 million monthly active users. And according to Shopify's 2025 holiday report, a majority of consumers already expected to use AI for at least part of their shopping. Tobi Lütke, Shopify's CEO, summed up the ambition by saying he wants to make every store "agent-ready by default." The phrase is marketing, but the mechanics are real: activation is now automatic rather than opt-in.

‍

There's also an infrastructure and standards angle. Shopify co-developed the Universal Commerce Protocol (UCP) with Google, an open standard for AI agents to transact with any merchant, already backed by more than twenty retailers and platforms including Amazon, Microsoft, Stripe, and Target. In parallel, OpenAI and Stripe are pushing the Agentic Commerce Protocol (ACP). The good news for merchants: you touch neither directly. Agentic Storefronts abstract both protocols. You simply choose which channels to enable in your admin.

‍

What does "agentic selling" actually mean in practice?

‍

Agentic selling refers to transactions where an AI agent acts as the purchasing intermediary. The user delegates research, comparison, and sometimes the transaction to the agent. The agent queries the catalog, evaluates options against the stated criteria, and presents a selection. What happens next depends on the channel, and this is where most summaries get it wrong.

‍

For ChatGPT, the agent handles discovery and redirects to your checkout. For Microsoft Copilot and Google AI Mode, an embedded checkout lets the buyer close without leaving the interface, if you enable it. Concretely: a user asks Copilot for a stainless steel French press under $60, the agent queries the Shopify Catalog, filters by price, material, and delivery, and offers a few options with reasoning. The channel determines whether the purchase closes in the conversation or on your store.

‍

For a Quebec or Canadian merchant, one nuance is decisive: discoverability in ChatGPT requires selling to US customers, even if your store is based in Canada. A merchant that only sells domestically won't appear in ChatGPT shopping for now. The Google channels (AI Mode and Gemini) run through the Google & YouTube sales channel and follow their own eligibility rules. Before promising anything to a client, check which market they actually sell into.

‍

The merchant's role shifts from driving traffic to the storefront to exposing the right signals. Product attributes must be machine-readable: material, dimensions, compatibility, real-time stock. The factors that surface a product are no longer backlinks or domain authority, but catalog completeness, description clarity, and how well price matches intent. A beautiful storefront whose specs live only in unstructured HTML is hard for an agent to interpret.

‍

Task automation vs. delegated judgment: why does this distinction matter?

‍

The Sidekick feature illustrates the real dividing line. Task automation executes a defined action, like adding a keyword to every product description, with no contextual decision. Delegated judgment lets the tool decide what to do based on context, like adjusting descriptions according to inventory velocity. At Winter '26, Sidekick gained much deeper in-admin execution, reusable skills, and even app generation, with its most advanced capabilities reserved for Advanced and Plus plans.

‍

This distinction changes the trust model. Task automation is low-risk: you review the output before it goes live. Delegated judgment has more leverage but introduces real risk if the agent's decisions drift from your brand strategy. A merchant selling premium artisan goods doesn't want an agent generating discount codes to move inventory faster, because it undermines positioning. A commodity electronics merchant might welcome that exact behavior.

‍

This is where the real trap in Shopify's announcements sits. The danger lies in automating the wrong things, the kind that gets paid back in customer experience. For agencies, that creates genuine consulting value: helping the client draw the line between what an AI can decide on its own and what must stay under human control. That decision is strategic before it is technical. Merchants with clear operational guardrails and high tolerance for machine-driven optimization are best placed to delegate; those whose brand rests on nuance should move more carefully.

‍

Which Shopify AI features are worth it now, and which should you wait on?

‍

First reflex to correct: for many stores, agentic discovery is no longer a feature to turn on, it's a default state to verify. Since March 24, 2026, Agentic Storefronts have been on automatically for eligible US merchants, in opt-out mode. So your first move is to go to Sales channels then Agentic in your admin, see which channels are active, and make sure your policies and product data are clean. ChatGPT discovery adds no fee beyond your usual payment processing.

‍

Prioritize catalog hygiene. Structured attributes, clean metafields, precise descriptions, current stock. That's the lever that decides whether an agent recommends your product or a competitor's. It's also low-risk: it improves both agent discoverability and traditional SEO.

‍

Test with guardrails: Sidekick's autonomous content generation. On a high-SKU, low-differentiation catalog, letting Sidekick write attribute-based descriptions can improve discoverability without brand risk. Keep brand-voice copy under human control and monitor the output for the first few weeks.

‍

Defer agent-to-agent price negotiation. It gets talked about as imminent, but for now it's an analyst projection from the likes of McKinsey, not a confirmed Shopify feature. Most merchants don't have margin structures that support real-time negotiation without eroding profitability. Wait for the market and margin protections to mature. Note one separate technical deadline too: on June 30, 2026, Shopify Scripts give way to Functions. If a client still runs on Scripts, that's a project to handle independently of the agentic question.

‍

How should an agency present these changes to cautious clients?

‍

Frame the conversation around clarity, not fear of missing out. The useful message isn't "AI agents will replace your storefront." It's "AI discovery is already active on your store by default, so let's make sure your products are well presented and that you keep control of the decisions that matter."

‍

Emphasize the incremental nature. Agentic Storefronts didn't change a merchant's existing site or checkout flow. It's a discovery channel running in parallel. Cautious clients respond better to "we're optimizing a channel that's already there" than to "we're rebuilding how you sell."

‍

Manage expectations by the client's market. For a merchant selling into the US, ChatGPT discoverability is a concrete channel to optimize. For a purely Canadian merchant, be honest about current eligibility limits and steer the effort toward accessible channels and catalog hygiene, which will pay off when eligibility widens.

‍

Avoid overselling autonomy. Most clients aren't ready to let an AI decide pricing or inventory without oversight. Position Sidekick as a tool that surfaces recommendations, with the merchant keeping final approval. That reduces anxiety and builds trust before moving to more autonomous configurations. For agencies managing several Shopify accounts, the conversation is best held proactively: a short document explaining what's already active, what's real, and what's still marketing, followed by an assessment call.

‍

Which types of merchants are best positioned to benefit from agentic commerce?

‍

High-SKU catalogs with clear specifications win immediately. Electronics, supplements, fashion basics, industrial supplies, and replacement parts fit agent strengths: filtering by attributes, comparing specs, checking stock. These categories map cleanly to structured queries like "USB-C cables under $20 with lots of positive reviews," which agents handle well.

‍

Merchants with mature infrastructure see faster returns. If your system already syncs real-time stock and exposes attributes in a structured format, integration is light. Stores on outdated systems face an upgrade before value appears. In those cases, the real question is: do we invest in catalog readiness now, or defer until volume justifies it?

‍

Conversely, highly experiential brands see less immediate value. Luxury goods, artisan crafts, design-forward furniture, and custom apparel rely on visual storytelling and emotional discovery, which agents can't replicate in a chat interface. A user looking for a handcrafted ceramic vase will get functional results, but the agent won't convey the maker's story or the texture of the glaze. For these merchants, the storefront remains the primary conversion environment and agentic channels serve as secondary discovery.

‍

One expectation to correct on B2B: even though the structured buying of procurement teams looks like a natural agent use case, B2B-only products aren't supported by Agentic Storefronts for now. It's a likely direction, not a current reality. For agencies, client fit assessment should happen early: an attribute-rich catalog with conversion that doesn't lean on visual immersion means aggressive adoption; an editorial brand built on lifestyle imagery means agentic channels as a safety net, with resources focused on the storefront.

‍

Where is Shopify pushing commerce in the next 12 months?

‍

The direction is clear: make Shopify the transactional layer for every AI interface. UCP is expanding, with a recent update adding multi-item carts, live catalog queries, and loyalty program integration. At the Spring '26 Edition, access to the agentic infrastructure became self-serve for all developers, which will multiply the shopping experiences built on the Shopify Catalog.

‍

Several bets remain uncertain, and it matters not to present them as settled. Agent-to-agent negotiation is projected by analysts, not confirmed as a product. Voice commerce, with Siri or Alexa routing through the same API layer, is plausible but not deployed at scale. And the monetization model itself is moving: OpenAI's removal of its 4% in-chat checkout fee, under merchant pressure, shows the economic rules aren't fixed. Any forecast about paid access tiers is therefore a hypothesis.

‍

What's certain is that the channel already exists, that it's on by default for a large share of merchants, and that adoption is accelerating. The work to do in 2026 is operational and comes down to four moves: clean up product data, verify which AI channels are active on the store, define clearly where judgment can be delegated and where it must stay human, and manage client expectations by their actual market. Search Engine Journal details the protocol specifications small merchants should prioritize, which aligns with this hygiene logic. For an agency, that's exactly where the value sits: translating a real shift into concrete moves, without overselling it.

‍

Sources: Shopify, Winter '26 Edition · Shopify Help Center, ChatGPT storefront · Shopify, agentic commerce guide · McKinsey, The agentic commerce opportunity · Search Engine Journal