We’re all looking for more details on Shopify’s new AI release: what’s available now and what’s coming. Our team took a look at all the announcements we’re seeing and summarized it to give you a glimpse of what’s on the horizon and how your brand can prepare for the new AI landscape ahead.
Shopify is introducing client-facing AI technology in several phases, the first of which launched in early 2025. Let’s take a look at Shopify’s evolution of AI to date:
1. Phase I: Agentic Chatbot, Shopify Magic & Shopify Sidekick
Shopify’s earliest AI features were modest but meaningful. The platform introduced an agentic chatbot; Shopify Magic, a suite of AI powered features; and Shopify Sidekick,
- Shopify’s agentic chatbot lets users ask questions conversationally, get product suggestions, and navigate the store in a more natural way. Still, it could not access customer-specific data, meaning it couldn’t access customer orders, customer account history, or private information. That limited its usefulness for post-sale support or account-specific queries, especially when other third parties had chatbots that could answer questions on orders, etc.
- Shopify Magic is Shopify’s suite of AI-powered features designed to help merchants save time and run their businesses more efficiently. Built directly into the Shopify admin, it can generate product descriptions, blog posts, and email content, assist with store setup, provide smart search and recommendations, and even automate workflows. Essentially, it acts like a built-in commerce-focused AI assistant—helping merchants create, optimize, and scale without needing external tools or technical expertise.
- Shopify Sidekick is Shopify’s AI assistant, built on Shopify Magic, that acts like a co-pilot inside your store. Using natural language prompts, Sidekick helps merchants with everyday tasks, from generating product descriptions or marketing content, setting up discounts or customer segments, analyzing sales trends, to guiding you through Shopify settings. Sidekick can also propose theme adjustments, create visuals, and surface data insights.
Of all 3 of these early AI features, only the agentic chatbot were available to shoppers. While these early features turned Shopify stores into more conversational, smarter storefronts, there was no cross-system data access, no in-chat checkout, no real “assistant” function that could act on your behalf. It was also unclear how their onsite agentic chatbot would play nicely with other 3rd parties, like a Gorgias chat bot for instance.
2. The Middle Phase: Product Discovery + Intent-Based Suggestions
Over the past months, Shopify (in collaboration with OpenAI) has pushed deeper into conversational commerce. The bot is becoming more than a FAQ-assistant; it’s evolving into an embedded shopping interface.
Some highlights:
- Embedded product preview within chat allows users to get product cards with images, pricing, reviews, and “Buy Now” style cues directly in chat, rather than just links. (Source: CX Today)
- Code leaks suggest “shopify_checkout_url” and “buy_now” strings inside ChatGPT’s assets, implying deeper integration not just to recommend but to transact. (Source: Omni Talk)
- Shopify’s new Storefront AI Agent API / Model Context Protocol (MCP) enables chat agents to query product metadata, manage cart state, and hand off to the checkout flow. (Source: Shopify)
- Acquisition of Vantage Discovery (an AI search company) signals Shopify’s commitment to more powerful AI-driven search, likely used to surface better results in chat. (Source: Business Insider)
- Shopify is also building out a universal cart capability, allowing users to consolidate products from multiple merchants (and possibly conversations) before checking out. (Source: PYMNTS.com)
Instead of navigating through search pages; the user chats, sees relevant picks, and moves toward purchase without changing interfaces.
3. What’s Coming: In-Chat Checkout & Agentic Commerce
This week, Shopify publicly flagged that its next AI phase is coming soon: chat + commerce. What’s expected (or at least anticipated):
- Full in-chat checkout: The user may never leave the chat. The entire payment flow, order confirmation, and merchant backend processing happen seamlessly.
- Agentic behavior: The AI can autonomously complete tasks for the user, within guardrails: e.g. “Reorder my last toner cartridges,” “Find a gift under $50 and checkout with my card.”
- Deeper customer-data access (with permissions): Allowing the AI to view order history, account preferences, and more to make better personalized decisions.
- Revenue sharing / commission model: Reports suggest OpenAI will take a cut of sales processed through ChatGPT-integrated commerce. (Source: Financial Times)
- Open sourcing of the “agentic commerce protocol” developed alongside Stripe, enabling wider adoption by third parties. (Source: Reuters)
- Strict boundaries / trust guardrails: Shopify is already inserting language in merchants’ robots.txt (and site scripts) to control what autonomous AI agents are allowed to do on their storefronts. (Source: Modern Retail)
If you are interested in getting notified when this next stage is released, you can sign up on Shopify’s website.
4. What’s Included vs. What’s Not for Shoppers — Right Now
So let’s summarize what Shopify AI functionality is ready now to put on your front-end experience, vs. what is coming:
| Capability | Status / Availability | Notes & Limitations |
| Conversational chat / FAQ / product lookup | Live / broadly available | The agent can answer product questions, suggest items, interpret user intent. |
| Customer-specific data access (orders, accounts) | Not yet | The bot today lacks insight into your actual account or transaction history. |
| In-chat checkout / payment | Not yet generally available | Some leaked code and early tests suggest this is coming soon. |
| Autonomous agentic actions | Not yet | The AI cannot yet act fully on your behalf (e.g. auto-checkout). |
| Universal cart across merchants | In development / preview | Shopify is building toward it, but it hasn’t fully shipped. |
| Guardrails & opt-in permissions | Partially live / being developed | Shopify is already putting limits in storefront code to manage agent access. |
5. What Merchants Should Be Doing Now to Prepare
To get ahead of the curve, here are some recommendations:
- Optimize your catalog metadata
- Ensure product titles, descriptions, attributes, and images are clean, consistent, and AI-friendly. Expect that chat ranking / “AIO” (AI optimization) will matter as much as classic SEO.
- Use structured data and rich attributes (e.g. in Shopify’s metafields) so the AI can reason about variants, sizes, availability.
- Enable Shopify’s APIs, permissions, and checkout readiness
- Adopt Shopify’s Storefront API, cart APIs, and ensure your checkout is flexible (e.g. supports app layers, custom scripts).
- Be ready to accept permissioned data sharing (e.g. let customers allow the AI agent to view their past orders).
- Map trust & consent flows
- Design your UX so that customers clearly grant consent before the AI accesses personal data. Transparency will be key for trust.
- Have fallback flows for when the AI can’t satisfy a request, such as escalating to a human agent.
- Monitor guardrail policies and Shopify’s agent rules
- Shopify is already inserting directives to block certain autonomous agent behaviors. Watch the developer docs and policy updates. (Source: Modern Retail)
- Make sure your storefront scripts, robots.txt, and APIs align with permitted use.
- Experiment early with AI search & chat tools
- Integrate existing chatbot / search tools, and try out what Shopify has available already
- Run pilot flows to test how your inventory and variants fare under conversational queries.
What began as a conversational overlay is rapidly evolving into conversational commerce: AI agents not just talking, but doing. These are exciting times ahead, but privacy/consent, hallucination in commerce context, scalability, cost (if OpenAI takes commission and token costs), and channel cannibalization will definitely be topics we’ll be seeing more of in the months to come.
To learn more, please view our AI Services or contact us.
