When Shopify released their Spring 2026 Edition last month, the individual updates told the story of Agentic Commerce. Viewed together they described something broader: Shopify is building AI into every layer of commerce, from the admin dashboard where Side kick segments customers and applied discounts to the protocol layer that standardizes and syndicates product data across agent models.
Here we’ve compiled the AI-specific updates into one digest, organized around the three pillars of AI commerce: Discoverability, Commerce Innovation, and Operational Efficiency.
AI Discoverability: how customers and AI systems find products
Shopify's Search & Discovery tools already give merchants control over recommendations, synonyms, filters, and merchandising rules. Now, the platform is moving search away from exact keyword matching to semantic matching: so a search for "athletic footwear" also surfaces sneakers and trainers. This is where we'd point most enterprise teams first, since it's one of the few AI capabilities that's fully available today and squarely within a brand's control.
Shopify's Catalog API extends that same discovery logic beyond Shopify's own storefront: it syndicates a merchant's product, variant, media, and availability data from a single catalog into AI assistants such as ChatGPT, Google AI Mode and Gemini, Microsoft Copilot, and Perplexity, with no separate integration required for each one. AI referral traffic to Shopify stores is already up more than 220% year over year, which makes this one of the fastest-growing discovery channels merchants have.
In the case of both of these updates, whether it’s onsite search or external systems, the tools only work if the underlying product data, taxonomy, metafields, attributes, and variants, is current, complete and structured. We cover the groundwork this requires in more detail in our work on AI search optimization.
Commerce innovation: the brand still owns the experience
We think brands should keep ownership of the customer experience even as they experiment aggressively with AI on the storefront, and Shopify's latest updates are built around that same balance.
Shopify and Google co-developed the Universal Commerce Protocol, a standard meant to let AI agents find a product, build a cart, and complete checkout and payment on a shopper's behalf. Shopify has designed its side of the protocol to preserve merchant business rules and existing systems, and it's positioning Shop Pay and its checkout stack as the transaction layer underneath these AI-driven experiences. Its theme layer got a matching update: themes can now emit standardized events and expose actions, so an AI agent interacting with a storefront does so predictably rather than improvising its way through the page.
Getting that right comes down to familiar fundamentals: accurate inventory, pricing, fulfillment, and regional rules, since an agent can only complete a transaction as reliably as the data behind it allows. This is the territory our agentic commerce work addresses.
Operational efficiency: AI workflows, human accountability
Our own delivery model pairs AI-accelerated workflows with a person who stays accountable for what ships, and Shopify is building the same pairing into its own admin tools.
Shopify Sidekick now lets merchants describe what they need in plain language and have Shopify handle the mechanics: segmenting customers, updating a catalog, creating a discount, or adjusting store settings. Any change Sidekick proposes still requires merchant review before it takes effect, which keeps a person in the loop on anything that touches a live store.
Shopify Magic generates first-draft copy: product descriptions, email copy, blog posts, headings. It also handles routine image work, including background removal, background generation, and canvas extension, and it can turn a plain-language prompt into a basic theme block or snippet of Liquid code.
Shopify's new Rollouts feature adds the same discipline to bigger changes: merchants can publish a new theme, checkout configuration, or customer accounts setup at a scheduled time, either to everyone or as a test to a subset of shoppers first. Shopify's SimGym pairs with it, generating AI shoppers that test a store and report back on what worked, though those findings still need a strategy or UX person to check them before anyone acts on them.
For enterprise teams, this raises a practical question: who has permission to approve what Sidekick, Magic, and Rollouts propose, to ensure product accuracy, brand voice, accessibility, and code quality? And how do those approvals get tracked?
AI readiness still depends on enterprise foundations
Despite being exciting advancements, it’s worth noting that using any of this well depends on things Shopify doesn't provide: data governance, clear user permissions, and a way to measure what's actually working. Complex analytics will often need a dedicated business intelligence layer, and complex service or inventory workflows will often need specialist applications designed for that purpose. AI-generated content and AI-initiated actions still need a second look, for a simple reason: the output can be generic, incomplete, or simply wrong.
Next steps for commerce leaders
Use now
- Shopify Magic for first-draft content and routine image edits
- Sidekick for administrative tasks, analytics creation, or bulk updates
- Rollouts and SimGym for safe, reversible A/B testing with no third party required
- Search merchandising and synonym management
- Catalog data cleanup
- Fraud analysis and risk indicators
Prepare next
- Structured product attributes and metafields
- Catalog API readiness
- Semantic and intent-based search
- Connections between Shopify and ERP, PIM, OMS, CRM, and service platforms
- Governance for AI-assisted workflows
- Measurement for AI-driven discovery and assisted transactions
Monitor
- Universal Commerce Protocol adoption
- AI shopping channel availability
- Agent-led checkout and post-purchase functions
- New Sidekick integrations
- AI action logging and enterprise controls
- Reporting for AI-driven discovery and transactions
Teams working toward the "prepare next" and "monitor" categories are exactly who our agentic commerce accelerator serves.
Domaine's perspective
We track every one of Shopify's product, platform, and developer updates to provide insights to our clients, and we sort them using the same three lenses we bring to any platform's AI roadmap: discovery, experience, and operations. That's the follow-through we spend our time on: what a brand's architecture, catalog, customer experience, operating model, and commercial priorities need before a given feature is worth adopting. This Edition adds more of the technical groundwork for agentic commerce on top of what already existed, and how much of that work pays off for any one brand comes down to data quality, architecture, governance, and the effort put into implementation.
