Shopify's President Says AI Agents Are Ecommerce's New Front Door. Here's What That Actually Means.
By Steve Merrill | March 22, 2026
Harley Finkelstein, Shopify's president, stood at the Upfront Summit in Los Angeles this week and said something that every Shopify merchant should be paying attention to: "We're going to begin to use these agentic applications as these kinds of personal shoppers." He wasn't speculating about the future. He was describing what Shopify is actively building toward right now.
What Did Harley Finkelstein Actually Say About AI Agents?
Speaking at the Upfront Summit on March 16, Finkelstein argued that AI-powered shopping agents will act as a new entry point for ecommerce merchants. He described AI agents as "personal shoppers" that guide buyers from intent to purchase, replacing or supplementing the traditional path of Google search, product page browse, and checkout.
As reported by TechCrunch and PYMNTS, Finkelstein acknowledged the initial rollout will take time. But the direction is clear: Shopify is betting that the next major shift in ecommerce distribution runs through AI agents, not through direct Google traffic or social media discovery.
This isn't a vague exec vision statement. Shopify has already built the Shopify Catalog (the product data layer that feeds ChatGPT), is actively integrating with major AI platforms, and is building its own AI shopping infrastructure. The strategy is real. The question is whether your store is ready for it.
What Is the "New Front Door" Model and How Does It Work?
The old front door was Google. A buyer searches "best hiking boots for wide feet," Google surfaces your page, the buyer clicks, browses, and maybe buys. You owned the experience from click to checkout.
The new front door works differently. A buyer tells an AI agent what they need: "I'm looking for hiking boots for wide feet, under $180, that ship in 3 days, and come from a brand that's not owned by a conglomerate." The agent searches across data sources, evaluates options against all those criteria, and surfaces 2-3 recommendations.
Your store doesn't get the traffic unless the AI agent can find you, trust your data, and match you to the buyer's criteria. The buyer never even sees the stores that don't get recommended.
That's a fundamentally different competitive dynamic. In the old model, you could show up for any search query with the right SEO. In the agent model, you either match the criteria or you're invisible.
Why Is the Initial Rollout "Taking Time" If This Is Already Happening?
Finkelstein's comment about the rollout taking time is important. The infrastructure is being built in stages.
The Shopify Catalog is live and feeding ChatGPT. But the full agentic experience, where an AI agent can browse your store, evaluate product fit against a buyer's detailed criteria, handle questions about inventory, manage returns, and complete purchase autonomously, that's still in progress.
The current state: AI agents are good at discovery and recommendation. They're still clunky at complex transactional tasks. That will change faster than most merchants expect.
The window to prepare is open right now. Once agentic commerce is fully live and buyers are routinely using AI agents for shopping, the stores with strong data and agent-readable structure will have a compounding advantage that's very hard to close.
What Does "Agent-Ready" Actually Mean for a Shopify Store?
Agent-ready isn't a single feature. It's a combination of data quality, structure, and accessibility that lets AI systems confidently recommend your products.
Here's what it breaks down to in practice:
Product data that answers buyer questions. An AI agent serves a buyer with specific needs. Your product descriptions need to answer the questions that buyer would ask. Not "high quality leather wallet", "full-grain vegetable-tanned leather wallet, 4 card slots, 1 cash slot, fits in front pocket, 3.5 x 4 inches, made in Portugal."
Schema markup that AI can parse. Product schema tells AI systems your price, availability, brand, category, and review data. If this is missing or incomplete, an AI agent can't confidently recommend you because it doesn't have the data it needs to match you to a buyer's criteria.
An llms.txt file that guides AI agents. This is a relatively new standard (similar to robots.txt but for language models) that tells AI agents which content on your site is authoritative, which products are your best sellers, and what makes your store unique. Very few Shopify stores have this yet.
Fast, clean product pages. Some AI agents browse your actual product pages to get data. Slow pages, broken links, and poor mobile experience all reduce an agent's confidence in recommending your store.
Is This Change Coming Fast Enough That Merchants Need to Act Now?
Yes. Here's why.
The stores building agent-ready infrastructure today will have months of data quality, catalog visibility, and schema consistency before the full agentic commerce wave hits. AI systems learn patterns. If they've been successfully recommending your products for months and buyers convert, that's a signal that reinforces future recommendations.
Waiting until agentic commerce is the dominant channel to start preparing is the same mistake I made with Facebook ads in 2016. I had a working business. I didn't think I needed to change. By the time I did, my competitors had two years of compounding advantage I couldn't close. I watched them go from $2.5M to $80-100M while I ground from $2.5M to $10M.
The early adopters in agentic commerce are building that same kind of advantage right now.
What Should Shopify Merchants Actually Do This Week?
Three things. No more, no less.
First, go to your Shopify admin and check your Shopify Catalog connection. Verify your products are in the feed and that your product data is complete. This is the most direct path to ChatGPT visibility today.
Second, pick your 20 best-selling products and rewrite their descriptions for buyer-language specificity. Add who it's for, when to use it, specific dimensions and materials, and what problem it solves. This helps on ChatGPT, Perplexity, Google AI Mode, and Google Shopping.
Third, audit your Product schema on at least your top 10 products. Check that price, availability, brand, and review aggregate fields are all present and accurate. Fix anything that's missing.
That's enough to move the needle. Do it this week.
FAQ: Shopify AI Agents and Agentic Commerce
What is the Shopify Catalog and how does it connect to AI shopping agents?
The Shopify Catalog is a centralized product data layer that Shopify built to feed AI platforms like ChatGPT with merchant product information. When a buyer asks ChatGPT a shopping question, ChatGPT queries the Shopify Catalog to find matching products. Merchants whose stores connect to the Catalog are eligible to appear in those recommendations.
Do I need to do anything to connect my Shopify store to AI agents?
The Shopify Catalog connection is being rolled out through Shopify's platform. Check your Shopify admin for any AI commerce or Catalog connection settings. Clean product data and proper schema markup improve your visibility regardless of the specific connection mechanism.
Will AI agents replace Google as a traffic source for Shopify stores?
Likely not a full replacement, but a significant shift in distribution. Early data shows AI-referred traffic converting at meaningfully higher rates than organic search traffic. The volume is still smaller, but the intent is stronger. Both channels will matter for the foreseeable future.
What is an llms.txt file and do I need one for my Shopify store?
An llms.txt file is placed at your domain root (yourstore.com/llms.txt) and tells AI language models what your site is about, what content is most useful, and how to focus on different sections. It's emerging as a standard for helping AI systems understand your site better. Worth adding.
How long will it take for my store to start appearing in AI shopping recommendations?
With a clean Shopify Catalog connection and updated product data, merchants often see visibility improve within 1-2 weeks. Schema updates can take 2-4 weeks to fully propagate. Building a track record of AI recommendations takes longer, 1-3 months of consistent data quality.
Finkelstein's comment at Upfront wasn't an announcement. It was a confirmation. Shopify has been building toward agentic commerce for months. The infrastructure is in motion. The question is whether your store is positioned to benefit from it when the channel matures.
The best time to prepare was six months ago. Second best time is today.
Find out if your Shopify store is ready for AI agent traffic →
