Why AI Agents Recommend Your Competitors Instead of You (And the Specific Data That Fixes It)

By Steve Merrill, Founder of WRKNG Digital — June 5, 2026

Ask ChatGPT to recommend a product in your category. There's a good chance your competitors come up before you do.

That's a data problem. AI shopping agents don't crawl your website the way a human would. They pull from structured data sources: product feeds, schema markup, and third-party product databases. When those sources are thin or missing fields, your products don't show up in the results. Your competitors' products do.

We ran this analysis across 47 Shopify stores over the past six months. Only 14% had product data complete enough to appear consistently in AI recommendations. The other 86% had at least one critical field missing. Most had three or four.

What's Actually Happening When AI Recommends a Product?

AI shopping agents like ChatGPT Shopping, Perplexity, and Google's AI Overviews work differently than traditional search engines. They're not ranking your site by domain authority or backlinks. They're matching product attributes against buyer intent, pulling from structured data sources in real time.

The main inputs are Google Merchant Center product feeds, Schema.org markup embedded in your product pages, and aggregated product databases that compile attributes at scale. When someone asks Perplexity "what's the best [product] under $100," Perplexity queries those sources, finds products that match the intent, and surfaces results.

Your competitor gets recommended because their product data hits the signals AI agents look for. Yours gets skipped because it doesn't. That's the whole story.

What's the Specific Data Gap That Gets Stores Skipped?

The data doesn't lie.

Four fields show up missing most often in our audits, and they're the same four fields AI agents weight most heavily when generating recommendations.

GTINs (Global Trade Item Numbers). The barcode on your product. This is how AI shopping agents match your product to a known item across multiple data sources. According to Google's structured data documentation for products, GTIN is one of the most important identifiers for product matching and recommendation eligibility. In our audits, 61% of Shopify stores had at least one top-selling product with no GTIN set.

Brand attribute. AI agents use brand to match products to brand-specific queries and recommendation contexts. A product without a brand attribute is essentially anonymous in the data layer. Shopify lets you set this in the Vendor field under Product Organization. Most merchants leave it blank because they don't know it connects to AI visibility. Shopify's product setup documentation covers where this field lives, but the link to AI shopping recommendations isn't something the docs explain.

Short product descriptions. AI agents need enough text to understand what a product is, who it's for, and how it differs from alternatives. A 40-word description gives them almost nothing to work with. The Schema.org Product type specification includes a description field for exactly this reason. Most stores write product descriptions for a shopper who's already on the page and mostly sold. AI agents start from zero and need real context.

Thin or missing review data. Review count and average rating are recommendation signals. Products with zero reviews get deprioritized. Stores with no review schema markup lose the signal they've already earned from customers who did the work of leaving feedback.

How Do You Know If Your Store Has the Gap?

Most stores have no idea. That's the real problem.

I've run AI audits on stores where the owner was certain their SEO was solid. Google rankings were good. Traffic was steady. But when we pulled the product feed and checked what ChatGPT Shopping actually ingests, the problems were obvious. Missing GTINs on 12 top products. Vendor field blank on 80% of the catalog. Descriptions averaging 47 words. The store was completely invisible to AI shopping agents, and the owner had no idea.

Here's a quick check you can run right now without any paid tool:

1. Go to your Shopify admin and export your product catalog as a CSV.
2. Check the "Variant Barcode" column. Count the blanks. For products you manufacture, you'll need a GTIN from GS1. For resale products, the manufacturer's barcode is your GTIN.
3. Check the "Vendor" column. That's your brand field in Shopify. Blanks mean anonymous products in the AI data layer.
4. Open your top 5 products and count the words in each description. Under 100 words means thin.

Most stores find at least two problems in 15 minutes. Some find all four.

How Do You Fix the Data Gap in Your Shopify Store?

The path is straightforward. The work takes time, but none of it is complicated.

Start with your top 20 products. Don't try to fix everything at once. Your best sellers are the products AI agents are most likely to see. Get those right first, then work down the catalog.

Add GTINs to every audited product. For products you resell, the manufacturer's barcode is your GTIN. For private-label products, register with GS1 (gs1us.org). For most small catalogs, that's under $250 per year. Enter the GTIN in the Variant Barcode field in Shopify.

Set the Vendor field on every product. In Shopify, open the product, scroll to Product Organization, and fill in the Vendor field with your full brand name. Don't abbreviate. Don't leave it blank.

Rewrite thin descriptions. Write for a machine that knows nothing about you. What is this product? Who is it for? What materials or specs matter? What makes it different from the three competitors selling something similar? Aim for 150 words minimum, written in natural language that answers the questions a shopper would actually ask.

Connect to Google Merchant Center and submit a complete feed. This is probably the highest-impact single action you can take. Google Merchant Center feeds are a direct input for ChatGPT Shopping, Google AI Overviews, and Perplexity shopping results. An absent or thin feed keeps you out of the conversation entirely. A complete, current feed gets you in.

Why Does the Timing on This Matter?

The shift is accelerating.

ChatGPT added shopping features in early 2025. Perplexity launched its shopping product. Google's AI Overviews now surface products directly in search results. AI-assisted product discovery is growing every quarter. It's still a minority of overall shopping searches, but it's the fastest-growing discovery channel most Shopify stores are completely ignoring right now.

I've seen this pattern before. I grew an ecommerce business to $10 million a year, and when Facebook changed its algorithm in 2013 and the ad platform opened up, I refused to adapt. I watched competitors hit $80-100 million while I was still figuring out what I'd missed. Same product category. Same market. Two-year head start.

The stores that build complete product data and feed hygiene now will compound that advantage as AI shopping grows. The ones waiting for proof it matters will face the same catch-up cost I paid in 2014. Some of them won't recover from it either.

The window is open. Won't stay open forever.