Agentic Commerce Explained: What Shopify Merchants Need to Know
For 25 years, online commerce worked the same way: a customer searches Google, clicks a result, visits your store, and buys. That model is ending. AI agents are now browsing, comparing, and purchasing products on behalf of shoppers — and the merchants who aren't ready are already invisible
How We Got Here: Three Eras of Online Commerce
To understand agentic commerce, it helps to see how dramatically the buying journey has changed — and how fast.
Era 1: Traditional Search Commerce (2000-2023)
This is the model every Shopify merchant knows:
- Customer searches Google: “best running shoes for flat feet”
- Google returns a mix of organic results and paid ads
- Customer clicks through to your store (or your competitor’s)
- Customer browses your site, reads reviews, checks sizing
- Customer adds to cart and buys
Your job was clear: rank on Google through SEO, run Google Ads to capture search intent, and optimize your site for conversions. The customer did all the work — researching, comparing, deciding, and purchasing. You just had to be findable and trustworthy.
This model drove trillions in commerce. It’s also completely dependent on one thing: the customer visiting your website.
Era 2: AI-Assisted Search (2024-2025)
Then AI chatbots started answering shopping questions:
- Customer asks ChatGPT or Perplexity: “What’s the best running shoe for flat feet?”
- AI synthesizes information from across the web and returns a curated recommendation
- Customer reads the recommendation and clicks through to buy on their own
The AI replaced Google’s “ten blue links” with a direct answer. But the customer still completed the purchase themselves — visiting your store, adding to cart, entering payment info. The AI was a research assistant, not a buyer.
This shift was already significant. On Black Friday 2025, AI-driven traffic to U.S. retail sites surged 805% year-over-year. 73% of shoppers are using AI somewhere in their buying journey. For merchants, this era introduced a new optimization challenge: Generative Engine Optimization (GEO) — making your products visible to AI engines, not just Google.
Era 3: Agentic Commerce (2026)
Now the AI doesn’t just recommend. It buys:
- Customer tells an AI agent: “Buy me a pair of running shoes for flat feet, under $150, delivered by Friday”
- AI agent searches across merchants, compares prices, checks inventory and shipping speed
- AI agent applies a coupon, uses your stored payment info, and places the order
- The merchant ships directly to you
The customer never visits a website. Never compares tabs. Never enters a credit card number. The AI agent handles the entire transaction — from discovery to checkout.
That’s agentic commerce. The word “agentic” comes from agency — the AI has the authority to act on your behalf. It’s the difference between a concierge who makes suggestions and a personal assistant who handles the purchase.
On Shopify alone, orders from AI searches are up 15x since January 2025. McKinsey projects that agentic AI will drive $3-5 trillion in global commerce by 2030. Bain estimates $300-500 billion in U.S. agentic commerce alone — 15-25% of total online retail. Gartner predicts AI agents will intermediate $15 trillion in B2B purchases by 2028. PwC, IBM, Salesforce, and Mastercard have all published frameworks for it. This isn’t a niche concept — every major player in commerce is building for it.
Why This Matters for Your Store
Here’s the uncomfortable shift: in Era 1 and Era 2, you could win by being findable. Good SEO, good ads, good content — and the customer showed up. In Era 3, the customer might never show up. The AI agent evaluates your product data, compares it against competitors, and decides — all without a human ever seeing your homepage.
If your product data is incomplete, your descriptions are thin, or your structured data is missing, you’re not just invisible to Google anymore. You’re invisible to the AI agents that are increasingly making purchase decisions.
What’s Actually Live Right Now
Let’s cut through the hype. Here’s what’s real in March 2026 — not announced, not planned, but live and processing transactions:
ChatGPT Shopping
OpenAI launched shopping research in November 2025 and briefly tested Instant Checkout — a way to buy directly inside the chat. With 900 million weekly users, ChatGPT is the largest AI shopping surface by far.
However, on March 6, 2026, OpenAI pulled back Instant Checkout. The reasons were telling: users research products in ChatGPT but prefer to complete purchases on merchant sites; real-time inventory sync and fraud detection proved too complex at scale; and only a small number of merchants were actively using native checkout. OpenAI is now pivoting to an “Apps” model where purchases happen inside connected services.
What’s still live and working: - You ask ChatGPT for a product recommendation - It returns product cards with images, prices, reviews, and specs - Product results are organic and unsponsored, ranked purely on relevance - Purchases now redirect to the merchant’s site or connected payment apps
What matters for merchants: ChatGPT pulls from Google Shopping feeds and structured product data. If your products are in Google Merchant Center with complete data, they can appear in ChatGPT shopping results. The Instant Checkout pullback actually favors merchants — customers land on your site where you control the experience.
Perplexity Shopping + Buy with Pro
Perplexity launched shopping features for Pro users in late 2024 and has been expanding since. In May 2025, they partnered with PayPal for instant checkout.
How it works: - Ask Perplexity a shopping question - It returns ranked product recommendations with prices, pros/cons, and stock status - Pro users can buy directly with “Buy with Pro” — free shipping included - Perplexity doesn’t take a cut from merchants
The Perplexity Merchant Program is free to join. Benefits include higher visibility in recommendations, checkout integration, free API access, and a dashboard with search trend insights.
Google AI Mode + Universal Commerce Protocol
At NRF 2026 in January, Google CEO Sundar Pichai announced the Universal Commerce Protocol (UCP) — an open standard co-developed with Shopify that lets AI agents browse catalogs, check inventory, and complete purchases through a standardized API.
Google AI Mode now shows product results with a Buy button that lets shoppers purchase without leaving the AI conversation. The merchant stays the merchant of record throughout — you get the order, handle fulfillment, and own the customer relationship.
For a deeper dive on UCP, see our complete guide to the Universal Commerce Protocol.
Amazon Buy for Me
Amazon launched “Buy for Me” in April 2025 — an AI agent that purchases products from third-party brand websites on the shopper’s behalf, without leaving the Amazon app.
How it works: - If Amazon doesn’t sell what you’re looking for, Buy for Me shows products from other sites - The AI agent navigates to the brand’s website and completes checkout using encrypted customer payment info - Powered by Amazon Nova and Anthropic’s Claude models - Over 500,000 items were included by end of 2025
Important caveat: Buy for Me has faced backlash from small merchants who found products listed without their consent. Unlike UCP (which is an open, opt-in standard), Amazon’s approach scrapes third-party sites. This distinction matters.
Shopify Agentic Storefronts
At their Winter ‘26 Edition, Shopify launched Agentic Storefronts — a system that surfaces your products in AI conversations across ChatGPT, Perplexity, Microsoft Copilot, and Google AI Mode.
Key components: - Shopify Catalog — structures your product data so AI agents can understand it (categories, attributes, variants) - Knowledge Base app — lets you upload brand guidelines, FAQs, and policies so AI agents represent your brand accurately - Agentic Storefronts channel — available in Settings > Sales Channels, lets you toggle which AI platforms can sell your products
Orders from AI agents appear in your Shopify admin like any other order. You handle fulfillment. You own the customer data.
The Technology Stack: How It All Fits Together
There are four layers that make agentic commerce work:
1. AI Agents (the “shoppers”)
These are the AI systems acting on behalf of consumers: ChatGPT, Perplexity, Google AI Mode, Amazon Alexa+, Microsoft Copilot. They understand natural language requests, compare products, and execute purchases.
2. Commerce Protocols (the “language”)
Two open standards are competing and coexisting:
- The Universal Commerce Protocol (UCP) — co-developed by Google and Shopify, handles product discovery, cart, checkout, and returns. Backed by 20+ partners including Etsy, Wayfair, Target, and Walmart
- The Agentic Commerce Protocol (ACP) — co-developed by OpenAI and Stripe, connects AI agents to merchants through a standardized interaction model
Additionally, the Model Context Protocol (MCP) lets agents interact with merchant tools and data sources.
Without protocols, every AI platform would need custom integrations with every merchant — not scalable. These open standards mean a single implementation works across multiple agents.
3. Payment Infrastructure (the “trust”)
AI agents can’t use your credit card number directly. Instead, credential providers (Google Pay, Shop Pay, PayPal) issue payment tokens — encrypted, single-use representations of your payment info that agents can pass to merchants securely.
This is where PayPal, Visa, Mastercard, and Stripe come in: - PayPal partnered with Perplexity and others for instant checkout inside AI conversations - Visa launched the Trusted Agent Protocol — an open framework that helps merchants distinguish legitimate AI agents from bots. Backed by Microsoft, Stripe, and Shopify. Visa predicts millions of consumers will use AI agents for purchases by the 2026 holiday season - Mastercard introduced Agent Pay to register and authenticate AI agents, plus a new open-source Verifiable Intent standard (March 2026) creating tamper-resistant proof of user authorization - Stripe released the Agentic Commerce Suite with Shared Payment Tokens (SPTs), fraud protection via Stripe Radar, and dashboard controls for selecting which AI agents can sell your products. Already used by URBN, Etsy, and Coach
4. Merchant Platform (the “store”)
This is you — the Shopify merchant. Your product data, inventory, pricing, shipping rules, and fulfillment capabilities. The quality of this data determines whether agents recommend your products or skip them.
What This Means for Shopify Merchants
Let’s get practical. Here’s what changes and what doesn’t.
What Changes
Your customer might be an AI agent. When an agent purchases from your store via UCP or Agentic Storefronts, the order arrives in your Shopify admin like a normal order. You won’t necessarily know it came from an AI conversation — and you don’t need to. You fulfill it the same way.
Product data quality becomes critical. AI agents can only recommend what they can understand. If your product titles are vague, descriptions are thin, and images lack alt text, agents will skip you for competitors with complete data. This isn’t hypothetical — it’s how ranking already works in ChatGPT Shopping and Perplexity.
You’re selling through new channels without building new channels. AI conversations are a new surface — but unlike building a marketplace listing or social storefront, you don’t need to create separate product feeds or manage new accounts. UCP and Shopify Catalog handle the syndication automatically.
Comparison shopping is instant. AI agents can compare your product against every competitor in seconds. Price, specs, reviews, shipping speed, return policy — everything is weighed at once. Your differentiation needs to be clear and machine-readable.
What Doesn’t Change
You’re still the merchant of record. Unlike marketplace models where the platform owns the customer, UCP-based agentic commerce keeps you as the merchant of record. You handle fulfillment, customer service, and returns. The customer relationship is yours.
Fulfillment is the same. Whether an order comes from your website, an AI conversation, or a social commerce channel, you ship it the same way. Shopify’s fulfillment network doesn’t change based on the order source.
Your brand still matters. AI agents cite brands they trust. Brand search volume is the strongest predictor of appearing in AI recommendations. Building a strong brand isn’t less important in agentic commerce — it’s more important.
The Honest Concerns
Agentic commerce isn’t all upside. Here are the real risks merchants should think about:
Disintermediation
When an AI agent handles the purchase, you lose direct interaction with the customer. They might never visit your site, see your brand experience, or join your email list. You get the sale, but you may lose the relationship.
Mitigation: Focus on post-purchase experience. Great packaging, follow-up communication, and exceptional products drive repeat purchases — whether the first order came from your site or an AI agent.
Fraud Risk
A Riskified study of 5,000 consumers found that while 70% of shoppers are comfortable with AI agents making purchases, only 13% have actually done it. As adoption grows, so will fraud attempts. AI agents add a layer between the buyer and merchant, potentially obscuring suspicious behavior.
The study recommends that platforms share behavioral data (IP, device info, behavioral signals) with merchants to enable fraud detection. Without this visibility, merchants are operating blind.
It’s Early — and Messy
OpenAI’s Instant Checkout pullback in March 2026 is a reminder that “buy inside the chat” is harder than it sounds. Forrester called it a signal that the transaction layer needs more time. The near-term reality: AI platforms are becoming the new discovery layer, while checkout and fulfillment remain anchored in existing merchant infrastructure. That’s actually good news for merchants — you keep control of the buying experience.
Price Pressure
When AI agents can compare prices across every merchant instantly, the temptation is to compete on price alone. This is a race to the bottom for commodity products.
Mitigation: Differentiate on things agents can communicate — unique materials, superior reviews, faster shipping, better return policies. If your only advantage is price, you’re vulnerable.
Amazon’s Approach vs. Open Standards
There’s a meaningful difference between Amazon’s “Buy for Me” and UCP-based approaches. Amazon’s agent scrapes third-party websites without merchant consent. UCP is an open standard where merchants opt in and maintain control.
For Shopify merchants, the UCP/Agentic Storefronts approach is preferable — you control which AI platforms can sell your products, and you retain the merchant of record status.
The Data: How Fast Is This Growing?
The numbers tell the story:
| Metric | Value | Source |
|---|---|---|
| AI shopping traffic (Black Friday 2025, YoY) | +805% | Adobe Analytics |
| AI traffic (full 2025 holiday season, YoY) | +693% | Adobe Analytics |
| Shopify orders from AI searches (since Jan 2025) | Up 15x | Shopify |
| Shoppers using AI in buying journey | 73% | Riskified |
| Shoppers who’ve completed AI-referred purchase | 13% | Riskified |
| Shoppers comfortable with AI buying for them | 70% | Riskified |
| AI users’ purchase completion rate vs. non-AI | 38% higher | Adobe Analytics |
| Projected agentic commerce market by 2030 (global) | $3-5 trillion | McKinsey |
| Projected U.S. agentic commerce by 2030 | $300-500 billion | Bain |
| Consumers who trust AI for full shopping process | Only 7% | Bain |
| ChatGPT weekly active users | 900 million | OpenAI |
The gap between “73% using AI” and “13% completing purchases” is where the opportunity lives. Adoption is ahead of transaction. As checkout friction drops (via UCP, tokenized payments, and Instant Checkout), that 13% will climb fast.
What You Should Do This Week
You don’t need to overhaul your store. Start with the high-impact basics:
1. Check your Agentic Storefronts status
Go to Settings > Sales Channels in your Shopify admin. If “Agentic Storefronts” appears, make sure all AI channels are toggled on. If it’s not there yet, Shopify is rolling it out — you’ll get access automatically.
2. Audit your product data
This is the single most important thing you can do. AI agents can only recommend what they can understand:
- Titles: Specific and descriptive. “Men’s Merino Wool Running Socks - Moisture-Wicking, Size 10-12” not “Running Socks”
- Descriptions: 150+ words of unique content per product. Include materials, dimensions, use cases
- Images: High-quality photos with descriptive alt text
- Variants: Properly labeled (color/size, not “Option 1/Option 2”)
- Metafields: Structured attributes like material, care instructions, and certifications
For a complete audit framework, use our Shopify SEO Checklist — every item that improves traditional SEO also improves your agentic commerce readiness.
3. Set up your Knowledge Base
Install Shopify’s Knowledge Base app and upload your FAQs, return policies, and brand guidelines. This ensures AI agents answer customer questions about your brand accurately — in your voice, with your policies.
The quality of this app is not great and if you like, create a similar page for your self in your shopify theme. But the idea remains that answer storewide policy questions for AI to read and understand.
4. Create your llms.txt file
An llms.txt file gives AI crawlers a machine-readable summary of your store. It’s like a cover letter for AI — here’s who we are, what we sell, and why it matters.
5. Ensure your Google Merchant Center is connected
ChatGPT Shopping pulls from Google Shopping feeds. If your products aren’t in Google Merchant Center, they can’t appear in ChatGPT results. Install Shopify’s free Google & YouTube channel to sync automatically.
6. Monitor AI channels
Search for your products in ChatGPT, Perplexity, and Google AI Mode. Are you appearing? Are competitors showing up instead? This reconnaissance tells you where to focus improvement. Understanding how AI engines discover and recommend products is the overlap between GEO and traditional SEO.
FAQ
What is agentic commerce in simple terms?
Agentic commerce is when AI assistants don’t just recommend products — they actually buy them for you. Instead of showing you search results and letting you handle the purchase, the AI agent compares options, checks inventory, applies your payment info, and places the order. The merchant ships directly to you.
Do I need to install anything on my Shopify store for agentic commerce?
No. Shopify implements UCP at the platform level — there’s no app to install. Agentic Storefronts is being rolled out automatically to eligible stores. Your job is to ensure your product data, shipping, and payment settings are properly configured, and optionally set up the Knowledge Base app.
Will AI agents replace my website?
No. Your website remains your primary sales channel. Agentic commerce is an additional channel — like adding Instagram Shopping or a marketplace listing. Most purchases still happen on merchant websites. AI agents are adding a new discovery and checkout path, not replacing existing ones.
How do returns work when an AI agent makes the purchase?
Returns follow your standard return policy. The customer contacts you (not the AI platform) and you handle the return as you normally would. The AI platform is a facilitator, not a marketplace — you’re the merchant of record.
Is Amazon’s “Buy for Me” the same as Shopify’s Agentic Storefronts?
No. Amazon’s Buy for Me scrapes third-party websites and completes purchases without merchant consent. Shopify’s Agentic Storefronts use the open Universal Commerce Protocol, where merchants opt in and maintain control. With UCP, you choose which AI platforms can sell your products and you retain the customer relationship.
What percentage of my sales will come from AI agents?
It’s early. The 805% growth in AI traffic sounds dramatic but starts from a small base. For most Shopify merchants today, AI-driven orders are under 1% of total sales. But that number is growing fast — Shopify’s 15x order growth from AI shows the trajectory. The merchants who prepare their product data now will capture this traffic as it scales.
Want to see how ready your store is for agentic commerce? Run a free AI readiness scan — it audits your products across 12 data quality checks and shows you exactly what AI agents see when they evaluate your catalog.