Agentic Commerce for Shopify: Protocols, Platforms, and What to Prioritize in 2026

A customer asks ChatGPT: "What's the best lightweight rain jacket under €200 that packs into its own pocket?" Within seconds, ChatGPT returns five options with pricing, reviews, and product images. The customer taps "buy" and completes the purchase without ever visiting a single brand's website.

That scenario is already happening at scale. Adobe Digital Insights tracked a 4,700% year-over-year increase in AI-driven traffic to retail sites by mid-2025. Shopify reports that AI-attributed orders on its platform grew 11x between January 2025 and January 2026. And as of late March 2026, Shopify is activating Agentic Storefronts by default for every store on the platform.

Most e-commerce leaders at large brands are hearing about agentic commerce constantly but finding mostly hype. This is the practical version: what's actually live, what's coming, and where to focus your team's time.

Agentic commerce means AI agents that browse, compare, recommend, and in some cases purchase products on behalf of consumers. It's a new sales channel, like marketplaces or social commerce before it.

Don't confuse this with merchant-side AI tools like Shopify's Sidekick or AI-powered customer service chatbots. Those help you run your store. Agentic shopping agents work for the consumer, acting as their personal shopping assistant across the entire web.

The implication for product discovery is significant. Today, most discovery flows through Google search, paid ads, social media, or direct site visits. With agentic commerce, an AI agent pre-filters the options and presents a curated shortlist. If your products aren't structured in a way these agents can read and evaluate, you don't make the cut.

Shopify has made what's arguably its biggest strategic bet since launching Shopify Plus. Rather than simply adding AI features, the company is trying to become the commerce infrastructure through which all AI agents transact, regardless of which platform the consumer uses.

The cornerstone is Agentic Storefronts, announced in the Winter '26 Edition and activating by default for all Shopify stores by late March 2026. Products are automatically syndicated through the Shopify Catalog, a global product data index that uses specialized AI models to categorize, enrich, and standardize product data across billions of items. The Catalog infers categories, extracts attributes, consolidates variants, and keeps pricing and inventory current across all AI channels.

Separately, the Agentic Plan (announced at NRF, January 2026) targets brands that aren't on Shopify at all. By letting competitors' merchants join Shopify's Catalog, Shopify is using agentic commerce as a wedge to pull brands into its ecosystem. There's no public pricing yet, and it remains on a waitlist.

On the merchant-side AI front, Sidekick has evolved from a basic chatbot into a proactive AI coworker with background analysis (Sidekick Pulse), workflow automation, and a skills system. Shopify Magic now covers text generation, media editing, store building, and theme generation.

Our take: Shopify has a meaningful head start here. Co-developing both UCP (with Google) and having deep integration with ACP (via Stripe), plus the Catalog intelligence layer, gives Shopify merchants an agentic commerce advantage that's hard to replicate on other platforms. This is one area where being on Shopify pays dividends quickly.

The conversion data looks good. Adobe's holiday 2025 data showed AI-referred shoppers converted 31% higher than other traffic, with revenue per visit up 254% year-over-year. A Visibility Labs study of 94 e-commerce stores found ChatGPT referrals converted at 1.81% versus 1.39% for non-branded organic search. Adyen's 2026 Retail Report revealed that 35% of US shoppers now use AI assistants for shopping, up from just 12% a year earlier. And 51% said they'd be willing to let AI handle the entire purchase process.

The volume doesn't. AI-driven traffic still represents under 1% of total retail traffic for most merchants. Shopify reports AI traffic is up 7x since January 2025, but it's growing from a very small base. A larger academic study of 973 sites covering $20 billion in revenue found ChatGPT underperformed most traditional channels on last-click attribution. The likely explanation is that many consumers discover products through AI agents but then search for the brand on Google to complete the purchase, making the AI contribution invisible in standard analytics.

Our take: prepare for this channel seriously, but don't reallocate significant budget from proven channels yet. The brands that invest now in agent-readiness will compound a visibility advantage as the channel scales. Bain projects agentic AI could account for 15-25% of US e-commerce by 2030. Even the conservative end of that range represents a massive channel.

Six actions, roughly in order of impact.

Verify your Agentic Storefronts settings. These are activating by default, but review which AI channels are enabled in your Shopify Admin. Ensure your products are visible and your store policies are accurate, since AI agents will use them to answer customer questions.

Audit your structured data. AI agents don't browse your product pages like humans. They extract Schema.org markup and compare factual specs. Shopify's default themes generate basic Product JSON-LD, but they miss critical fields: GTIN/EAN identifiers, aggregate ratings and reviews, return policies, shipping details, and FAQ schema. Apps like JSON-LD for SEO by Ilana or Schema Ninja can close these gaps. The key principle: "100% GOTS certified organic cotton, 200 GSM" outperforms "luxuriously soft premium cotton" in every agent ranking algorithm.

Optimize your Google Merchant Center feed. Google's Shopping Graph feeds multiple AI agents, not just Google's own. Updated guidelines from NRF 2026 recommend titles of 30+ characters, descriptions of 500+ characters, GTIN always populated, and a minimum of 3 additional product images. Google is also piloting new conversational attributes like answers to common product questions and compatible accessories.

Configure your robots.txt for AI crawlers. Allow retrieval crawlers that power real-time search (OAI-SearchBot, ChatGPT-User, PerplexityBot, Google-Extended) while optionally blocking training crawlers (GPTBot, anthropic-ai). The distinction matters: OAI-SearchBot drives ChatGPT search results; GPTBot crawls for model training. Shopify recently added default language restricting automated scraping, so check your current configuration at yourstore.com/robots.txt to make sure AI retrieval crawlers aren't accidentally blocked.

Populate the Knowledge Base App. This lets you define policies, FAQs, and brand voice so AI agents give accurate, on-brand responses when customers ask questions about your products.

Rewrite product descriptions for AI readability. Agents compare factual specifications, not marketing copy. Include materials, dimensions, weights, compatibility information, and use cases. Your pages don't need to become boring spec sheets for human visitors. But the technical detail needs to be present in the structured data and description body alongside your brand storytelling.

Agentic commerce isn't equally relevant across all categories.

Brands with highly customized or configurable products (custom furniture, bespoke tailoring) lose too much of the purchase experience when an AI agent tries to simplify the journey. The same goes for brands where tactile experience or visual storytelling drives the purchase decision. An AI agent describing a luxury fragrance or an artisan textile can't replicate the sensory discovery that closes those sales.

Very small catalogs (under 50 SKUs) don't benefit much from AI-powered discovery because the browsing advantage is minimal. And brands in categories where Amazon dominates may find that AI agents route around their stores entirely when the same product is available with Prime shipping.

The bigger strategic concern is disintermediation. When transactions happen inside AI platforms, brands lose direct traffic, first-party data, and control over the purchase experience. OpenAI's retreat from Instant Checkout suggests this model has limitations, but Google's UCP is building toward exactly this. Attribution is also unsolved. If a customer discovers your product through ChatGPT but buys through Google search, most analytics tools will credit organic search and miss the AI contribution entirely.

Our take: the disintermediation risk is real but manageable. The brands best positioned are those investing in structured data quality and protocol participation now, while also building direct customer relationships that make them the preferred merchant even when an AI agent is mediating the discovery.

Consider what happened in just the last eight weeks: OpenAI abandoned its checkout model, Google expanded UCP with three major capabilities, Stripe launched machine-to-machine payments, and Amazon's $50 billion OpenAI investment reshuffled the power dynamics.

The open questions are significant. Will ACP and UCP merge, or will merchants need both indefinitely? How does the Amazon-OpenAI partnership affect independent brand visibility? And will AI agents start showing paid ads alongside organic results? (Google is already testing Shopping Ads in AI Mode, so the answer there is almost certainly yes.)

The structured data, feed quality, and crawler configuration work you do today compounds over time as AI agents learn which merchants provide reliable, complete product information.

Brands that wait for the channel to mature before investing will find themselves playing catch-up against competitors who started earlier.