How do I actually know if traffic from ChatGPT and AI search is turning into sales?
You keep hearing that ChatGPT and other AI tools send shoppers to stores, but Shopify's reports don't hand you a clean answer. Here's how to actually check it with what you already have.
What's going on
You keep hearing that ChatGPT, Perplexity, and other AI assistants are sending shoppers to stores, and you want to know if that's true for yours, and more importantly, whether those visitors actually buy anything. When you open Shopify's Analytics, there's no single AI traffic toggle. Referral sources show up scattered across direct, search, and various referrer domains, and it isn't obvious how to isolate AI platforms from everything else or judge whether they're worth paying attention to.
This matters because the instinct many merchants have is to jump straight to optimizing for AI search, chasing citations, tweaking content, installing tools, before establishing whether there's any traffic or revenue there to optimize for in their specific case. Without a baseline, you can't tell later whether a change you make actually moved anything.
The complication is that referrer attribution itself is imperfect. AI assistants, especially their mobile apps and in-app browsers, don't always pass referrer information the way a normal browser link click does, so some AI-originated visits get miscounted as direct traffic. That means even a careful look at your reports gives you a reasonable estimate, not a perfect count.
Why it happens
Shopify's analytics were built around traditional acquisition categories, search, social, email, direct, referral, and AI assistants get bucketed into referral traffic like any other external site that links to you, rather than being called out separately. You have to filter for the specific AI domains yourself to see them.
On top of that, referrer information is inconsistently passed by AI apps and in-app browsers, so a portion of genuinely AI-referred visits land in your direct bucket instead, understating the real number regardless of which AI platform sent them.
5 ways to fix it
Pull a referrer breakdown and check conversion rate, not just traffic volume
Shopify's Analytics reports let you break sessions down by referrer, so you can isolate domains like chatgpt.com, perplexity.ai, or gemini.google.com from your general search, social, and direct totals. Once you've isolated those sessions, compare their conversion rate and average order value against your store average using the same report rather than judging by raw visit counts alone. A referrer that sends a lot of traffic but converts far below your baseline isn't actually worth chasing yet, and this comparison is free and available in every Shopify plan.
Don't take the raw AI-referral number at face value, bot fetches inflate it
AI systems also crawl and re-fetch your catalog pages automatically so they have current data to draw on later, and those automated requests can land in the same traffic logs as real human clicks from an AI answer. If a referrer's session count looks unusually high but its conversion rate is near zero, some portion of that traffic is likely non-human fetching rather than shoppers. Treat any AI-referral number as a rough estimate that needs a conversion-rate sanity check, not a precise headcount of interested visitors.
Expect real AI-referred visits to undercount as Direct, don't dismiss shifts in that bucket
Many AI assistants, especially their mobile apps and in-app browsers, don't reliably pass referrer information the way a normal browser link click does. A real click-through from an AI answer can end up logged as Direct traffic instead of being attributed to the AI platform. This is a general limitation of referrer-based attribution, not something specific to Shopify. If your Direct traffic or its conversion rate has shifted noticeably around the same time AI-referral activity picked up, that's worth investigating rather than dismissing. A simple how did you hear about us prompt on your post-purchase or order confirmation flow can catch some of what referrer data misses.
Fix your underlying product data before optimizing for AI visibility at all
Whether AI-referred traffic converts well depends heavily on whether AI systems can extract accurate facts about your products in the first place, things like price, availability, variant options, and identifiers such as barcodes or GTINs. Shopify includes basic structured product data automatically on modern themes, but it's only as complete as the information you've entered. Filling in variant barcodes and writing complete, accurate product descriptions in your product admin is free, improves conventional search visibility too, and is a prerequisite for AI systems representing your products correctly, not an optional extra step.
For ongoing tracking, this is a reporting habit to build, not a problem that needs an app
There's no built-in dashboard that automatically labels a session as AI-referred and tracks its long-term value for you, you're assembling that picture manually from the reports above. If you want it monitored on an ongoing basis rather than checked occasionally, a marketing analytics app or help from a Shopify partner can set up a standing referrer-segmentation view for you. For most stores, though, periodically re-running the manual comparison above is enough to know whether this is worth paying more attention to before spending on new tooling.
Bottom line
Start with what Shopify's Analytics already gives you for free, a referrer breakdown filtered to specific AI domains and compared against your store's normal conversion rate, before assuming you need new tooling. If AI-referred visitors are converting reasonably well, the next priority isn't an app that chases AI traffic, it's making sure your product data (structured data, barcodes, accurate descriptions) is complete enough for AI systems to represent correctly, which is a data-cleanup task, not a purchase. If you later want this tracked continuously without manually re-running reports, that's a case for analytics setup help rather than a single-purpose app, since no app can fill in incomplete product data for you.
Browse the rest of the problem library, run a free storefront scan to catch issues like this automatically, or email us and we'll work out a custom solution for it.