Harley Finkelstein Says AI Will Be Your Customers' Personal Shopper. Is Your Shopify Store Ready?
By Steve Merrill | March 23, 2026
Last week at the Upfront Summit in Los Angeles, Shopify President Harley Finkelstein said something that deserves more attention than it got: "We're going to begin to use these agentic applications as these kinds of personal shoppers." He wasn't talking about a future roadmap item. He was talking about now.
Most Shopify merchants nodded along and kept doing what they were doing. That's a mistake.
What Did Harley Finkelstein Actually Say About AI Personal Shoppers?
Finkelstein's framing at Upfront Summit was specific: AI agents are becoming a new entry point for ecommerce merchants. The customer won't open your store. They'll ask ChatGPT. They'll talk to a Perplexity shopping agent. They'll use whatever AI assistant is on their phone. That agent will decide what to recommend, and the customer will buy based on that recommendation.
That's a purchasing funnel with your store nowhere in it — unless you've prepared your data, your catalog, and your product descriptions for how AI agents actually evaluate products.
Here's the thing. Finkelstein knows Shopify's position in this. They've already built the infrastructure to connect merchants to AI conversations. The Shopify Catalog feeds product data directly to ChatGPT. But the catalog only works if your data is good. And most stores' data isn't.
I've run audits on hundreds of Shopify stores. The same problems show up every time: vague product titles, missing specifications, descriptions written for humans but not parseable by AI, no structured data to confirm category, use case, or key attributes. An AI agent hitting that store has almost nothing to work with.
Why Does the "Personal Shopper" Model Change the Rules for Product Data?
A personal shopper doesn't browse your catalog. They ask questions and expect answers.
"What's a good gift for a runner who just finished their first half-marathon?" A great personal shopper matches that to a product with the right attributes: endurance-focused, celebratory, appropriate price range, something a runner would actually want. An AI agent does the same thing, pulling from the data available in your product feed.
If your product says "Performance Running Shoes" with a 200-word generic description, the agent can't help. If it says "Performance Road Running Shoe, neutral cushioning, 8.5oz, ideal for half-marathon training and race day, heel-to-toe drop 8mm," the agent can match it to that query. One product gets recommended. The other doesn't show up at all.
This is the gap Finkelstein is talking about. The infrastructure is there. The question is whether your product data is ready to travel through it.
According to a study from the Baymard Institute, 67% of ecommerce product pages are missing the attribute data that shoppers use to make decisions. That number gets worse when you filter for AI-readable structured data. We're talking about a small fraction of Shopify stores that are genuinely prepared for AI personal shopping agents right now.
What Is Shopify Actually Doing to Connect Merchants to AI Conversations?
Shopify's approach is platform-level. They've built the connection layer so that any merchant's products can surface in AI conversations through the Shopify Catalog. That's meaningful. It means you don't have to build a custom API integration with every AI company yourself.
But "connected" and "recommended" are not the same thing.
Being connected means your products are in the pool. Being recommended means the AI agent chose your product over the other 40 million products in that pool. The selection criteria come down to data quality, relevance, trust signals, and review data. Shopify provides the pipe. You fill it with data that actually works.
Finkelstein's timeline signal matters here too. He said the "initial rollout will take time," which means the window for early adoption is still open. Stores that get their product data right in the next 60-90 days will have a real advantage over stores that wait until agentic shopping is mainstream to start thinking about it. That's the exact pattern I watched play out with Facebook ads in 2014. The early movers compounded away from everyone else.
How Should Shopify Merchants Respond to This Shift Right Now?
Start with your product data. Not your homepage. Not your blog. Your product titles, descriptions, and metafields.
Every product should have: a specific, accurate title with key attributes in it, a description that answers "who is this for and what does it do," clear size/weight/material specs where relevant, and honest review data that gives agents social proof signals to work with.
Add structured data markup if you haven't already. Product schema with Product, Offer, and AggregateRating types is the minimum. If your theme doesn't generate it automatically, use an app or add it manually to your product template.
Check whether you have an llms.txt file at your domain root. This file tells AI crawlers what your store is about and which pages to prioritize. It's small effort, high signal.
And watch what Shopify announces over the next 60 days. Their agentic commerce infrastructure is still expanding. Merchants who understand the direction before it fully rolls out will be positioned to take advantage of it. The ones who wait for the announcement email will be playing catch-up again.
FAQ: Harley Finkelstein, Shopify, and AI Personal Shoppers
What exactly did Harley Finkelstein say about AI personal shoppers?
Speaking at the Upfront Summit in Los Angeles on March 16, 2026, Finkelstein said Shopify expects AI agents to serve as personal shoppers, acting as a new entry point for merchants. He called it the beginning of a fundamental shift in how customers discover and buy products.
Does every Shopify store automatically get access to AI shopping agent distribution?
Shopify has built the Shopify Catalog infrastructure to connect merchants to AI conversations. But your products only get recommended if your data is accurate, specific, and structured in a way AI agents can use. Connection to the network doesn't guarantee recommendations.
What product data matters most for AI personal shoppers?
Specific attributes (size, material, use case, compatibility), clear titles that answer what the product is and who it's for, honest review data, and structured schema markup. AI agents match products to queries based on these signals.
How soon will AI personal shoppers affect Shopify sales?
It's already happening. ChatGPT Shopping, Perplexity's shopping features, and Google AI Overviews are all surfacing products now. Finkelstein's comments suggest Shopify expects the volume to increase significantly as agentic applications mature. Stores that wait will be late to a compounding advantage.
What's the fastest thing I can do to prepare my Shopify store for AI shopping agents?
Audit your top 20 products for attribute completeness. Fix vague titles and descriptions. Add or verify Product schema markup. These three steps take a few hours and immediately improve your AI visibility.
The window for early adoption in AI commerce is still open. Stores that act now will have a compounding advantage over the ones that wait. See how WRKNG Digital helps Shopify stores get ready for AI personal shoppers.
