Here's a pattern playing out in Shopify stores right now: a shopper asks ChatGPT for a product recommendation, gets a shortlist, clicks through (or types the store name into their browser), and buys. The merchant gets the sale — and has no idea AI had anything to do with it.
Ask most store owners how much revenue AI drives for them and they'll say "none, as far as I can tell." The honest answer is usually "some — but it's hiding in your reports."
Why AI-driven sales are invisible
Analytics attributes a sale to a channel by reading where the visitor came from — the referrer, or UTM tags on the link. AI-driven shopping breaks that in several ways:
Shoppers often don't click a link at all. They read ChatGPT's recommendation, then open a new tab and type your store name or Google it. That session has no referrer from the AI. Your analytics logs it as Direct or Organic Search — and the AI's role vanishes.
When they do click, the trail is fragile. Some AI surfaces pass a referrer, many don't, and links copied out of an AI chat and pasted into a browser arrive with nothing attached. Every one of those lands in the Direct bucket.
The buying journey is split. A shopper might get the recommendation on their phone during a commute and buy on their laptop that evening. Even good analytics struggles to stitch that together; the AI touchpoint that started everything is on a different device entirely.
The result: "Direct" quietly becomes a junk drawer that includes a growing slice of AI-referred buyers — and the channel driving them gets zero credit. You can't invest in a channel you can't see.
Why this matters more every month
If AI discovery were a rounding error, mismeasuring it wouldn't matter. It isn't. Shoppers increasingly start with "what should I buy?" in an AI assistant before they ever open Google — and the brands that get named win the sale before a search ever happens. (We've written about how to test whether AI recommends your store and why the same competitors keep winning those answers.)
That creates an uncomfortable loop: merchants don't see AI revenue in their reports → they conclude AI doesn't matter for them → they ignore their AI visibility → competitors who are measuring it invest, get recommended more, and compound their lead.
The merchants who figure out measurement first get to make decisions with data while everyone else guesses.
What you can actually do about it
Tag what you can control. Anywhere you can influence the link AI might surface — your llms.txt, structured data, profiles — make sure clean URLs with sensible UTM parameters exist. It won't catch everything, but it converts some invisible visits into visible ones.
Watch Direct for unexplained growth. If Direct traffic is climbing while your other channels are flat, some of that growth is likely misattributed discovery — AI included. It's a blunt signal, but it's a signal.
Match orders to AI signals, not just clicks. This is the real fix. Referrers and UTMs from AI sources do show up on a meaningful share of orders — if something is actually reading order data and looking for them. Shopify's own reports weren't built to hunt for AI fingerprints across referrer URLs, UTM sources, and sales-channel data. Purpose-built attribution is.
That last one is exactly what Vanto's Confirmed AI Revenue does: it reads your real Shopify orders and attributes revenue to the AI channel that drove each one — ChatGPT, Perplexity, Gemini, Copilot — using six different signals, net of refunds. Not estimates, not modeled projections: actual orders, matched to actual AI channels. Combined with per-product visibility tracking, you finally see both halves — whether AI recommends you, and what that's actually worth in dollars.
The shift to AI-driven discovery is happening whether your analytics can see it or not. The only question is whether you'll be measuring it before your competitors do.